Add files using upload-large-folder tool

output/preprocess/Allergies/code/TCGA.py

@@ -0,0 +1,71 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Allergies"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z1/preprocess/Allergies/TCGA.csv"
+out_gene_data_file = "./output/z1/preprocess/Allergies/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z1/preprocess/Allergies/clinical_data/TCGA.csv"
+json_path = "./output/z1/preprocess/Allergies/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import pandas as pd
+
+# Step 1: Identify the most appropriate TCGA subdirectory for the trait "Allergies"
+# Since TCGA cohorts are cancer types and none relate to allergies, we attempt a keyword search.
+keywords = [
+    "allerg", "hypersens", "atopy", "atopic", "asthma", "urticaria", "rhinitis", "eczema", "hayfever", "hay_fever"
+]
+
+# List available TCGA subdirectories
+available_subdirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+# Find candidates whose names contain any allergy-related keyword
+candidates = [d for d in available_subdirs if any(k in d.lower() for k in keywords)]
+
+selected_dir = None
+if len(candidates) > 0:
+    # If multiple matches, choose the one with the longest keyword overlap (more specific)
+    def score_dir(name: str) -> int:
+        lname = name.lower()
+        return sum(lname.count(k) for k in keywords)
+
+    candidates.sort(key=score_dir, reverse=True)
+    selected_dir = candidates[0]
+
+# If no suitable directory found, mark as skipped and stop here
+if selected_dir is None:
+    print("No TCGA cohort directory matches the target trait 'Allergies'. Skipping this trait.")
+    # Record as unavailable for this trait
+    validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+    clinical_df = None
+    genetic_df = None
+else:
+    print(f"Selected TCGA cohort directory: {selected_dir}")
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+
+    # Step 2: Identify clinical and genetic file paths
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+    print(f"Clinical file: {clinical_file_path}")
+    print(f"Genetic file:  {genetic_file_path}")
+
+    # Step 3: Load both files
+    clinical_df = pd.read_csv(clinical_file_path, sep="\t", index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_file_path, sep="\t", index_col=0, low_memory=False)
+
+    # Step 4: Print column names of the clinical data
+    print("Clinical data columns:")
+    print(list(clinical_df.columns))

output/preprocess/Alopecia/code/GSE148346.py

@@ -0,0 +1,189 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Alopecia"
+cohort = "GSE148346"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Alopecia"
+in_cohort_dir = "../DATA/GEO/Alopecia/GSE148346"
+
+# Output paths
+out_data_file = "./output/z1/preprocess/Alopecia/GSE148346.csv"
+out_gene_data_file = "./output/z1/preprocess/Alopecia/gene_data/GSE148346.csv"
+out_clinical_data_file = "./output/z1/preprocess/Alopecia/clinical_data/GSE148346.csv"
+json_path = "./output/z1/preprocess/Alopecia/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Step 1: Determine data availability
+is_gene_available = True  # Based on study design (biopsy transcriptomic biomarkers), gene expression data is likely available.
+
+# Step 2: Variable availability and conversion functions
+trait_row = 3  # 'tissue disease state: LS/NL' -> can be used as binary disease state (lesional vs non-lesional)
+age_row = None
+gender_row = None
+
+def _after_colon(val):
+    if val is None:
+        return None
+    if not isinstance(val, str):
+        val = str(val)
+    parts = val.split(":", 1)
+    return parts[1].strip() if len(parts) == 2 else val.strip()
+
+def convert_trait(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v_low = v.strip().lower()
+    # Map lesional vs non-lesional to 1/0
+    if v_low in {"ls", "lesional", "lesion", "lesional scalp", "lesional skin"}:
+        return 1
+    if v_low in {"nl", "non-lesional", "nonlesional", "non lesional", "nonlesion", "non-lesion", "non-lesional scalp"}:
+        return 0
+    return None
+
+def convert_age(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    # Remove common units/labels
+    for token in ["years", "year", "yrs", "yr", "y", "age", " "]:
+        v_low = v_low.replace(token, "")
+    v_clean = v_low.strip().replace("+", "")
+    try:
+        return float(v_clean)
+    except Exception:
+        # Try to extract any leading number
+        import re
+        m = re.search(r"(\d+(\.\d+)?)", v_low)
+        return float(m.group(1)) if m else None
+
+def convert_gender(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v_low = v.strip().lower()
+    if v_low in {"female", "f", "woman", "girl"}:
+        return 0
+    if v_low in {"male", "m", "man", "boy"}:
+        return 1
+    return None
+
+# Step 3: Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Step 4: Clinical feature extraction (only if clinical data available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age if age_row is not None else None,
+        gender_row=gender_row,
+        convert_gender=convert_gender if gender_row is not None else None
+    )
+    preview_output = preview_df(selected_clinical_df, n=5)
+    print(preview_output)
+    # Save selected clinical features
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+# Affymetrix probe set IDs (e.g., 1007_s_at, 1405_i_at) are not human gene symbols and require mapping
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# 1. Decide the appropriate columns for mapping:
+#    - Probe identifiers in gene expression data match the 'ID' column in annotation.
+#    - Gene symbols are in the 'Gene Symbol' column.
+
+# 2. Extract mapping dataframe from gene_annotation
+mapping_df = get_gene_mapping(annotation=gene_annotation, prob_col='ID', gene_col='Gene Symbol')
+
+# 3. Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Bias check and remove biased demographic features
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Final validation and save cohort info
+is_gene_available_final = (normalized_gene_data.shape[0] > 0) and (normalized_gene_data.shape[1] > 0)
+is_trait_available_final = (trait in linked_data.columns)
+
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available_final,
+    is_trait_available=is_trait_available_final,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=""
+)
+
+# 6. Save linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Alopecia/cohort_info.json

@@ -1,52 +1 @@
-{
-  "GSE80342": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": true,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 31
-  },
-  "GSE66664": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 140
-  },
-  "GSE18876": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE148346": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 129
-  },
-  "TCGA": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  }
-}
+{"GSE81071": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "INFO: Trait not available; skipped linking and QC. No SYMBOL column available in annotation; kept Entrez-indexed matrix."}, "GSE80342": {"is_usable": false, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": true, "has_age": true, "has_gender": true, "sample_size": 31, "note": "INFO: Samples=31, Genes=19845, Age_included=True, Gender_included=True."}, "GSE66664": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 140, "note": "INFO: DP cell line dataset; trait derived from 'cell line: BAB'(1) vs 'BAN'(0). Male-only; no age available. DHT dose/time present but not included as covariates."}, "GSE18876": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE148346": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 129, "note": ""}, "TCGA": {"is_usable": false, "is_gene_available": false, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}}

output/preprocess/Type_2_Diabetes/code/GSE271700.py

@@ -0,0 +1,447 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Type_2_Diabetes"
+cohort = "GSE271700"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Type_2_Diabetes"
+in_cohort_dir = "../DATA/GEO/Type_2_Diabetes/GSE271700"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Type_2_Diabetes/GSE271700.csv"
+out_gene_data_file = "./output/z6/preprocess/Type_2_Diabetes/gene_data/GSE271700.csv"
+out_clinical_data_file = "./output/z6/preprocess/Type_2_Diabetes/clinical_data/GSE271700.csv"
+json_path = "./output/z6/preprocess/Type_2_Diabetes/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Gene expression data availability
+is_gene_available = True  # Whole-genome microarray indicates gene expression data.
+
+# Variable availability (Responder/Non-Responder used as proxy for T2D remission status)
+trait_row = 3
+age_row = 1
+gender_row = 0
+
+def _after_colon(x):
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        val = parts[1] if len(parts) > 1 else parts[0]
+        val = val.strip()
+        return val if val != "" else None
+    return None
+
+def convert_trait(x):
+    # Map Responder -> 0 (no T2D, remission), Non-Responder -> 1 (has T2D, persistent)
+    val = _after_colon(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if "non" in v and "responder" in v:
+        return 1
+    if "responder" in v:
+        return 0
+    if "remission" in v:
+        return 0
+    if "persistent" in v:
+        return 1
+    if "no diabetes" in v or "without diabetes" in v:
+        return 0
+    if "diabetes" in v:
+        return 1
+    return None
+
+def convert_age(x):
+    val = _after_colon(x)
+    if val is None:
+        return None
+    try:
+        return float(val)
+    except Exception:
+        import re
+        nums = re.findall(r"[\d.]+", val)
+        return float(nums[0]) if nums else None
+
+def convert_gender(x):
+    val = _after_colon(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if v in {"female", "f"}:
+        return 0
+    if v in {"male", "m"}:
+        return 1
+    return None
+
+# Trait availability determined by trait_row
+is_trait_available = trait_row is not None
+
+# Initial filtering and save metadata
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Clinical feature extraction
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df)
+
+    import os
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+import os
+import io
+import gzip
+import re
+import pandas as pd
+
+# Preserve original probe-level data
+probe_level_data = gene_data.copy()
+
+def _pick_probe_id_col(df: pd.DataFrame) -> str:
+    candidates = ['ID', 'ID_REF', 'PROBE_ID', 'Probe ID', 'SPOT_ID', 'Probe Set ID', 'Probe.Set.ID']
+    for c in candidates:
+        if c in df.columns:
+            return c
+    # Heuristic for Affy-like IDs
+    def affy_like_ratio(series: pd.Series, n: int = 2000) -> float:
+        s = series.astype(str).head(n)
+        return s.str.contains(r'(?:_at|_s_at|_x_at)$', regex=True).mean()
+    ratios = {c: affy_like_ratio(df[c]) for c in df.columns}
+    return max(ratios, key=ratios.get)
+
+def _symbol_candidates(df: pd.DataFrame):
+    # Prioritized list of columns that may contain gene symbols
+    priority_names = [
+        r'^Gene[\s._-]*Symbol$', r'^GENE[\s._-]*SYMBOL$', r'^Symbol$', r'^SYMBOL$',
+        r'Gene\.Symbol', r'gene[\s._-]*symbol', r'symbol'
+    ]
+    # Expand to include plausible text columns for fallback
+    fallback_patterns = [
+        r'Gene[\s._-]*Title', r'GENE[\s._-]*TITLE', r'Gene[\s._-]*Name', r'GENE[\s._-]*NAME',
+        r'Gene[\s._-]*Assignment', r'GENE[\s._-]*ASSIGN', r'DESCRIPTION', r'Definition', r'Product', r'title', r'description'
+    ]
+
+    cols = list(df.columns)
+    # Priority matches
+    pri = []
+    for pat in priority_names:
+        pri.extend([c for c in cols if re.search(pat, str(c), flags=re.I)])
+    # Unique while preserving order
+    seen = set()
+    pri = [c for c in pri if not (c in seen or seen.add(c))]
+
+    # Fallbacks
+    fb = []
+    for pat in fallback_patterns:
+        fb.extend([c for c in cols if re.search(pat, str(c), flags=re.I)])
+    seen = set()
+    fb = [c for c in fb if (c not in pri) and not (c in seen or seen.add(c))]
+
+    return pri + fb
+
+def _extractability_score(series: pd.Series, n: int = 2000) -> float:
+    s = series.astype(str).head(n)
+    extracted = s.map(extract_human_gene_symbols)
+    return extracted.map(lambda x: len(x) > 0).mean()
+
+def _build_mapping_from_annotation(df: pd.DataFrame) -> pd.DataFrame:
+    if df is None or df.empty:
+        return pd.DataFrame(columns=['ID', 'Gene'])
+    probe_col = _pick_probe_id_col(df)
+    symbol_cols = _symbol_candidates(df)
+    best_col = None
+    best_score = -1.0
+    for c in symbol_cols:
+        try:
+            score = _extractability_score(df[c])
+        except Exception:
+            continue
+        if score > best_score:
+            best_score = score
+            best_col = c
+        # Early stop if a perfect match is found
+        if score >= 0.9:
+            break
+    if best_col is None or best_score <= 0:
+        return pd.DataFrame(columns=['ID', 'Gene'])
+    try:
+        mapping = get_gene_mapping(df, prob_col=probe_col, gene_col=best_col)
+    except Exception:
+        mapping = pd.DataFrame(columns=['ID', 'Gene'])
+    return mapping
+
+def _read_gpl_soft_platform_table(path: str) -> pd.DataFrame:
+    # Extract lines between !platform_table_begin and !platform_table_end
+    def _read_lines(open_fn):
+        lines = []
+        with open_fn(path, 'rt') as fh:
+            in_table = False
+            for line in fh:
+                line = line.rstrip('\n')
+                if line.startswith('!platform_table_begin'):
+                    in_table = True
+                    continue
+                if line.startswith('!platform_table_end'):
+                    break
+                if in_table:
+                    lines.append(line)
+        return lines
+
+    lines = []
+    try:
+        lines = _read_lines(gzip.open)
+    except Exception:
+        try:
+            lines = _read_lines(open)
+        except Exception:
+            lines = []
+
+    if not lines:
+        return pd.DataFrame()
+    table_str = '\n'.join(lines)
+    try:
+        return pd.read_csv(io.StringIO(table_str), sep='\t', low_memory=False, on_bad_lines='skip')
+    except Exception:
+        return pd.DataFrame()
+
+def _read_gpl_annot_file(path: str) -> pd.DataFrame:
+    try:
+        return pd.read_csv(path, sep='\t', compression='gzip', low_memory=False, on_bad_lines='skip')
+    except Exception:
+        try:
+            return pd.read_csv(path, sep='\t', low_memory=False, on_bad_lines='skip')
+        except Exception:
+            return pd.DataFrame()
+
+# Try multiple sources to obtain mapping
+mapping_df = pd.DataFrame(columns=['ID', 'Gene'])
+chosen_probe_col = None
+chosen_symbol_col = None
+mapping_source = None
+
+# 1) Try platform table embedded in the provided SOFT file
+platform_df = _read_gpl_soft_platform_table(soft_file)
+if not platform_df.empty:
+    tmp_map = _build_mapping_from_annotation(platform_df)
+    if not tmp_map.empty:
+        mapping_df = tmp_map
+        mapping_source = 'SERIES_SOFT_PLATFORM'
+        # Infer chosen columns for logging
+        chosen_probe_col = _pick_probe_id_col(platform_df)
+        chosen_symbol_col = [c for c in _symbol_candidates(platform_df) if c in platform_df.columns][0] if _symbol_candidates(platform_df) else None
+
+# 2) Search for GPL files in the cohort directory if needed
+if mapping_df.empty:
+    files = os.listdir(in_cohort_dir)
+    gpl_annot_files = [os.path.join(in_cohort_dir, f) for f in files if ('gpl' in f.lower() and 'annot' in f.lower())]
+    gpl_soft_files = [os.path.join(in_cohort_dir, f) for f in files if ('gpl' in f.lower() and 'soft' in f.lower())]
+
+    candidate_platform_dfs = []
+    for p in gpl_annot_files:
+        dfp = _read_gpl_annot_file(p)
+        if not dfp.empty:
+            candidate_platform_dfs.append(('GPL_ANNOT', dfp, p))
+    for p in gpl_soft_files:
+        dfp = _read_gpl_soft_platform_table(p)
+        if not dfp.empty:
+            candidate_platform_dfs.append(('GPL_SOFT', dfp, p))
+
+    for src, dfp, pth in candidate_platform_dfs:
+        tmp_map = _build_mapping_from_annotation(dfp)
+        if not tmp_map.empty:
+            mapping_df = tmp_map
+            mapping_source = src
+            chosen_probe_col = _pick_probe_id_col(dfp)
+            # Best symbol col (for logging) by score
+            syms = _symbol_candidates(dfp)
+            if syms:
+                scores = {c: _extractability_score(dfp[c]) for c in syms}
+                chosen_symbol_col = max(scores, key=scores.get)
+            break
+
+# 3) Recursive search in the trait directory if still not found
+if mapping_df.empty:
+    candidate_platform_dfs = []
+    for root, dirs, files in os.walk(in_trait_dir):
+        for f in files:
+            fl = f.lower()
+            if ('gpl' in fl) and (('annot' in fl) or ('soft' in fl)):
+                full = os.path.join(root, f)
+                if 'annot' in fl:
+                    dfp = _read_gpl_annot_file(full)
+                    if not dfp.empty:
+                        candidate_platform_dfs.append(('GPL_ANNOT', dfp, full))
+                elif 'soft' in fl:
+                    dfp = _read_gpl_soft_platform_table(full)
+                    if not dfp.empty:
+                        candidate_platform_dfs.append(('GPL_SOFT', dfp, full))
+    # Prefer annot over soft
+    candidate_platform_dfs.sort(key=lambda x: 0 if x[0]=='GPL_ANNOT' else 1)
+    for src, dfp, pth in candidate_platform_dfs:
+        tmp_map = _build_mapping_from_annotation(dfp)
+        if not tmp_map.empty:
+            mapping_df = tmp_map
+            mapping_source = src
+            chosen_probe_col = _pick_probe_id_col(dfp)
+            syms = _symbol_candidates(dfp)
+            if syms:
+                scores = {c: _extractability_score(dfp[c]) for c in syms}
+                chosen_symbol_col = max(scores, key=scores.get)
+            break
+
+print(f"Mapping source: {mapping_source if mapping_source else 'None'}")
+print(f"Selected probe ID column: {chosen_probe_col if chosen_probe_col else 'None'}")
+print(f"Selected gene symbol column: {chosen_symbol_col if chosen_symbol_col else 'None'}")
+print(f"Mapping size: {len(mapping_df) if not mapping_df.empty else 0}")
+
+# Apply mapping or keep probe-level data as fallback
+mapped_gene_data = None
+if not mapping_df.empty:
+    try:
+        mapped_gene_data = apply_gene_mapping(expression_df=probe_level_data, mapping_df=mapping_df)
+    except Exception:
+        mapped_gene_data = None
+
+if mapped_gene_data is not None and not mapped_gene_data.empty:
+    gene_data = mapped_gene_data
+    print(f"Gene mapping completed: {probe_level_data.shape[0]} probes -> {gene_data.shape[0]} genes.")
+else:
+    gene_data = probe_level_data
+    print("WARNING: Gene mapping failed or produced empty result. Falling back to probe-level data.")
+
+# Step 7: Data Normalization and Linking
+import os
+import re
+import pandas as pd
+
+# Detect Affymetrix-like probe IDs using a robust ratio computation
+def _is_probe_level_index(index, n_check: int = 2000) -> bool:
+    pat = re.compile(r'.*_(?:at|s_at|x_at)$', flags=re.IGNORECASE)
+    sample = [str(x) for x in list(index)[:min(len(index), n_check)]]
+    if not sample:
+        return False
+    hits = sum(1 for x in sample if pat.match(x))
+    ratio = hits / len(sample)
+    return ratio >= 0.5
+
+# Ensure we can access the clinical dataframe (fallback to file if needed)
+try:
+    selected_clinical_df  # noqa: F401
+except NameError:
+    if os.path.exists(out_clinical_data_file):
+        selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)
+    else:
+        raise
+
+note_msgs = []
+
+# 1. Normalize gene data only if index already looks like gene symbols
+gene_df_to_use = gene_data
+attempted_normalization = False
+if not _is_probe_level_index(gene_df_to_use.index):
+    attempted_normalization = True
+    try:
+        tmp_norm = normalize_gene_symbols_in_index(gene_df_to_use.copy())
+    except Exception as e:
+        tmp_norm = pd.DataFrame()
+        note_msgs.append(f"ERROR: Normalization failed with error: {e}")
+
+    if tmp_norm is not None and not tmp_norm.empty:
+        normalized_gene_data = tmp_norm
+    else:
+        normalized_gene_data = gene_df_to_use
+        note_msgs.append("WARNING: Normalization yielded empty result; using original gene matrix.")
+else:
+    normalized_gene_data = gene_df_to_use
+    note_msgs.append("INFO: Probe-level identifiers detected (e.g., '_at' suffix). Mapping to symbols failed earlier; skipping gene symbol normalization and keeping probe-level data.")
+
+# Ensure output directory exists and save gene data
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link the clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values in the linked data
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Determine whether the trait and some demographic features are severely biased, and remove biased features.
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Conduct quality check and save the cohort information.
+is_gene_available_flag = (normalized_gene_data.shape[0] > 0 and normalized_gene_data.shape[1] > 0)
+is_trait_available_flag = trait in selected_clinical_df.index
+
+note = " ".join(note_msgs) if note_msgs else "INFO: No special notes."
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available_flag,
+    is_trait_available=is_trait_available_flag,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note
+)
+
+# 6. If the linked data is usable, save it
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Type_2_Diabetes/code/GSE281144.py

@@ -0,0 +1,382 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Type_2_Diabetes"
+cohort = "GSE281144"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Type_2_Diabetes"
+in_cohort_dir = "../DATA/GEO/Type_2_Diabetes/GSE281144"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Type_2_Diabetes/GSE281144.csv"
+out_gene_data_file = "./output/z6/preprocess/Type_2_Diabetes/gene_data/GSE281144.csv"
+out_clinical_data_file = "./output/z6/preprocess/Type_2_Diabetes/clinical_data/GSE281144.csv"
+json_path = "./output/z6/preprocess/Type_2_Diabetes/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+import pandas as pd
+
+# 1) Gene expression availability
+is_gene_available = True  # Microarray gene expression mentioned in background info
+
+# 2) Variable availability
+trait_row = 1  # 'diabetes status: Control' vs 'Diabetic'
+age_row = None  # No age information in the sample characteristics
+gender_row = 0  # 'Sex: Female' / 'Sex: Male'
+
+# 2.2) Converters
+def _after_colon_lower(x):
+    if x is None:
+        return None
+    s = str(x)
+    if ':' in s:
+        s = s.split(':', 1)[1]
+    s = s.strip().lower()
+    return s if s else None
+
+def convert_trait(x):
+    s = _after_colon_lower(x)
+    if s is None:
+        return None
+    # Positive diabetes indicators
+    if any(k in s for k in ['diabetic', 'type 2', 't2d', 'dm2', 'type ii']):
+        return 1
+    # Negative diabetes indicators
+    if any(k in s for k in ['control', 'non-diabetic', 'nondiabetic', 'without diabetes', 'no diabetes', 'normoglycemic']):
+        return 0
+    # Generic yes/no if context only has diabetes status
+    if s in ['yes', 'y', 'true', 'positive', '+']:
+        return 1
+    if s in ['no', 'n', 'false', 'negative', '-']:
+        return 0
+    # If the original field includes 'diabetes' keyword try a final heuristic
+    orig = str(x).lower()
+    if 'diab' in orig:
+        if 'control' in orig or 'non' in orig:
+            return 0
+        return 1
+    return None
+
+def convert_age(x):
+    s = _after_colon_lower(x)
+    if s is None:
+        return None
+    m = re.search(r'(\d+(\.\d+)?)', s)
+    if not m:
+        return None
+    try:
+        return float(m.group(1))
+    except Exception:
+        return None
+
+def convert_gender(x):
+    s = _after_colon_lower(x)
+    if s is None:
+        return None
+    # Map female -> 0, male -> 1
+    if s in ['female', 'f', 'woman', 'girl']:
+        return 0
+    if s in ['male', 'm', 'man', 'boy']:
+        return 1
+    # Heuristics
+    if 'female' in s:
+        return 0
+    if 'male' in s:
+        return 1
+    return None
+
+# 3) Initial filtering and save metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (only if clinical data available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age if age_row is not None else None,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    # Preview and save
+    preview = preview_df(selected_clinical_df, n=5)
+    print(preview)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+import os
+import re
+import pandas as pd
+
+# Keep original expression data
+expr_df = gene_data.copy()
+
+# Prepare expression ID sets (exact and stripped suffix like '_st')
+expr_ids = expr_df.index.astype(str)
+expr_ids_set = set(expr_ids)
+expr_ids_stripped = expr_ids.str.replace(r'_[^_]+$', '', regex=True)
+expr_ids_stripped_set = set(expr_ids_stripped)
+
+def compute_overlap(series: pd.Series) -> tuple:
+    vals = series.astype(str).str.strip()
+    exact = int(vals.isin(expr_ids_set).sum())
+    stripped_vals = vals.str.replace(r'_[^_]+$', '', regex=True)
+    stripped = int(stripped_vals.isin(expr_ids_stripped_set).sum())
+    return exact, stripped
+
+def collect_probe_col_candidates(df: pd.DataFrame) -> list:
+    preferred = [
+        'probeset_id', 'Probe Set ID', 'probesetid', 'probeset',
+        'ID', 'ID_REF', 'ProbeSet ID', 'probeset identifier',
+        'transcript_cluster_id', 'Transcript Cluster ID'
+    ]
+    cands = []
+    for c in df.columns:
+        cl = c.lower()
+        if (('probe' in cl or 'id' in cl or 'transcript' in cl) and df[c].dtype != 'float') or (c in preferred):
+            cands.append(c)
+    pref_set = [c for c in preferred if c in df.columns]
+    others = [c for c in cands if c not in pref_set]
+    return pref_set + others
+
+def collect_gene_symbol_candidates(df: pd.DataFrame) -> list:
+    known = [
+        'gene_symbol', 'Gene Symbol', 'Symbol', 'GENE_SYMBOL',
+        'gene_assignment', 'mrna_assignment', 'gene symbols'
+    ]
+    cands = [c for c in known if c in df.columns]
+    cands += [c for c in df.columns if ('gene' in c.lower()) and (c not in cands)]
+    return cands
+
+# Prefer platform-specific annotation if available; otherwise fall back to family SOFT
+annotation_sources = []
+annotation_sources.append((gene_annotation, os.path.basename(soft_file)))
+for f in os.listdir(in_cohort_dir):
+    path = os.path.join(in_cohort_dir, f)
+    fl = f.lower()
+    if f == os.path.basename(soft_file):
+        continue
+    try:
+        if ('gpl' in fl and fl.endswith('.soft.gz')) or ('annot' in fl and fl.endswith('.gz')):
+            try:
+                df = pd.read_csv(path, compression='gzip', sep='\t', low_memory=False, on_bad_lines='skip')
+            except Exception:
+                df = pd.read_csv(path, compression='gzip', sep=',', low_memory=False, on_bad_lines='skip')
+            annotation_sources.append((df, f))
+        elif (fl.endswith('.txt.gz') or fl.endswith('.tsv.gz')) and fl.endswith('.gz'):
+            try:
+                df = pd.read_csv(path, compression='gzip', sep='\t', low_memory=False, on_bad_lines='skip')
+            except Exception:
+                df = pd.read_csv(path, compression='gzip', sep=',', low_memory=False, on_bad_lines='skip')
+            annotation_sources.append((df, f))
+        elif fl.endswith('.csv.gz'):
+            df = pd.read_csv(path, compression='gzip', low_memory=False, on_bad_lines='skip')
+            annotation_sources.append((df, f))
+    except Exception as e:
+        print(f"Skipping {f} due to error while reading: {e}")
+
+best = {
+    'overlap_exact': -1,
+    'overlap_stripped': -1,
+    'probe_col': None,
+    'gene_col': None,
+    'df': None,
+    'source': None
+}
+
+for ann_df, source in annotation_sources:
+    if not isinstance(ann_df, pd.DataFrame) or ann_df.shape[1] == 0:
+        continue
+    probe_cols = collect_probe_col_candidates(ann_df)
+    gene_cols = collect_gene_symbol_candidates(ann_df)
+    if not probe_cols or not gene_cols:
+        continue
+    for pcol in probe_cols:
+        try:
+            exact, stripped = compute_overlap(ann_df[pcol])
+        except Exception:
+            continue
+        # Choose gene column (prefer ones with 'symbol' or 'assignment')
+        gcol = None
+        for cand in gene_cols:
+            if cand in ann_df.columns:
+                gcol = cand
+                break
+        if exact > best['overlap_exact'] or (exact == best['overlap_exact'] and stripped > best['overlap_stripped']):
+            best.update({
+                'overlap_exact': exact,
+                'overlap_stripped': stripped,
+                'probe_col': pcol,
+                'gene_col': gcol,
+                'df': ann_df,
+                'source': source
+            })
+
+print(f"Selected annotation source: {best['source']}")
+print(f"Probe column candidate: {best['probe_col']}")
+print(f"Gene column candidate: {best['gene_col']}")
+print(f"Overlap (exact): {best['overlap_exact']}, Overlap (stripped): {best['overlap_stripped']}")
+
+if (best['overlap_exact'] <= 0) and (best['overlap_stripped'] <= 0):
+    tried_sources = [name for _, name in annotation_sources]
+    raise ValueError(
+        f"No overlap between expression probe IDs and any annotation columns. "
+        f"Tried sources: {tried_sources}. "
+        f"Expression IDs example: {list(expr_ids[:5])}"
+    )
+
+# Show a small sample of intersecting IDs for verification
+ann_ids_series = best['df'][best['probe_col']].astype(str).str.strip()
+ann_ids_set = set(ann_ids_series)
+intersection_exact = list(expr_ids_set & ann_ids_set)
+if not intersection_exact:
+    # Try stripped
+    ann_ids_stripped_set = set(ann_ids_series.str.replace(r'_[^_]+$', '', regex=True))
+    intersection_stripped = list(expr_ids_stripped_set & ann_ids_stripped_set)
+    print("Sample intersecting IDs (stripped):", intersection_stripped[:10])
+else:
+    print("Sample intersecting IDs (exact):", intersection_exact[:10])
+
+# Build mapping dataframe (ID + raw gene text)
+raw_mapping_df = get_gene_mapping(best['df'], prob_col=best['probe_col'], gene_col=best['gene_col'])
+
+# Deterministic parsing of Affymetrix-style "gene_assignment" field to symbols
+def symbols_from_assignment(text: str) -> list:
+    if not isinstance(text, str) or not text.strip():
+        return []
+    symbols = []
+    # Split entries by ' /// ' and then fields by ' // '
+    for entry in text.split('///'):
+        parts = [p.strip() for p in entry.split('//')]
+        if len(parts) >= 2:
+            sym = parts[1].strip().strip('"')
+            if not sym or sym in {'---', 'NULL'}:
+                continue
+            u = sym.upper()
+            # Exclusions: transcript IDs, non-gene placeholders, miRNA/LOC/LINC etc.
+            if re.match(r'^(NR_|XR_|XM_|NM_|ENST|ENSG|OTTHUMT|OTTHUMG|UC)', u):
+                continue
+            if re.match(r'^(LOC\d+|LINC\d+)', u):
+                continue
+            if u.startswith('MIR') or u.startswith('SCARNA') or u.startswith('SNORD') or u.startswith('RNU'):
+                continue
+            # Keep plausible symbols: uppercase letters/digits/hyphens or C#ORF#
+            if re.match(r'^(?:[A-Z0-9-]{2,20}|C\d+ORF\d+)$', u):
+                symbols.append(u)
+    # Deduplicate in order
+    return list(dict.fromkeys(symbols))
+
+# Clean mapping: produce a gene-only string that apply_gene_mapping will parse cleanly
+clean_mapping = raw_mapping_df.copy()
+clean_mapping['GeneList'] = clean_mapping['Gene'].apply(symbols_from_assignment)
+clean_mapping = clean_mapping[clean_mapping['GeneList'].map(len) > 0].copy()
+clean_mapping['Gene'] = clean_mapping['GeneList'].apply(lambda lst: ' '.join(lst))
+clean_mapping = clean_mapping[['ID', 'Gene']]
+
+# Decide whether to use exact or stripped mapping for probe IDs
+use_stripped = best['overlap_stripped'] > best['overlap_exact']
+
+# Prepare expression dataframe aligned to the mapping ID format
+if use_stripped:
+    expr_tmp = expr_df.copy()
+    expr_tmp.index = expr_tmp.index.to_series().str.replace(r'_[^_]+$', '', regex=True)
+    if expr_tmp.index.duplicated().any():
+        expr_tmp = expr_tmp.groupby(expr_tmp.index).mean()
+    expression_for_mapping = expr_tmp
+else:
+    expression_for_mapping = expr_df
+
+print("Preview of cleaned mapping_df:")
+print(clean_mapping.head(5))
+
+# Apply mapping to convert probe-level to gene-level
+gene_data = apply_gene_mapping(expression_df=expression_for_mapping, mapping_df=clean_mapping)
+
+# Quick sanity check
+print("Gene-level data shape:", gene_data.shape)
+print("Gene-level preview (first 5 genes):")
+print(gene_data.head(5))
+
+# Step 7: Data Normalization and Linking
+import os
+import pandas as pd
+
+# 1. Normalize gene symbols and save
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# Ensure clinical features are available in the current context
+if 'selected_clinical_df' not in globals():
+    selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)
+
+# 2. Link clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Assess bias and remove biased demographic features
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Final validation and save cohort info
+note = "INFO: Age unavailable; Gender available; Affymetrix-style mapping applied."
+is_usable = validate_and_save_cohort_info(
+    True, cohort, json_path, True, True, is_trait_biased, unbiased_linked_data, note=note
+)
+
+# 6. Save linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Type_2_Diabetes/code/GSE98887.py

@@ -0,0 +1,216 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Type_2_Diabetes"
+cohort = "GSE98887"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Type_2_Diabetes"
+in_cohort_dir = "../DATA/GEO/Type_2_Diabetes/GSE98887"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Type_2_Diabetes/GSE98887.csv"
+out_gene_data_file = "./output/z6/preprocess/Type_2_Diabetes/gene_data/GSE98887.csv"
+out_clinical_data_file = "./output/z6/preprocess/Type_2_Diabetes/clinical_data/GSE98887.csv"
+json_path = "./output/z6/preprocess/Type_2_Diabetes/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+import os
+import pandas as pd
+
+# 1) Gene expression availability (scRNA-Seq -> gene expression data present)
+is_gene_available = True
+
+# 2) Variable availability from the provided Sample Characteristics Dictionary
+# Only 'tissue: inlet cells' is available and constant; trait/age/gender not available.
+trait_row = None
+age_row = None
+gender_row = None
+
+# 2.2 Conversion functions
+def _after_colon(x):
+    if x is None:
+        return None
+    s = str(x)
+    parts = s.split(":", 1)
+    return parts[1].strip() if len(parts) == 2 else s.strip()
+
+def convert_trait(x):
+    v = _after_colon(x)
+    if v is None or v == "":
+        return None
+    s = v.lower().strip()
+    s = s.replace("–", "-").replace("—", "-")
+
+    # Exclude Type 1 diabetes explicitly (not our trait label)
+    if re.search(r"\b(type\s*1|type[-\s]?i|t1d)\b", s):
+        return None
+
+    # Negative/control indications first to avoid substring collisions (e.g., "non-diabetic")
+    neg_patterns = [
+        r"\bnon[-\s]?diabet\w*\b",
+        r"\bcontrol(s)?\b",
+        r"\bhealthy\b",
+        r"\bnormo?glyc\w*\b",
+        r"\bno diabetes\b",
+        r"\bnormal\b",
+        r"\bhc\b"
+    ]
+    if any(re.search(pat, s) for pat in neg_patterns):
+        return 0
+
+    # Positive (T2D) indications
+    pos_patterns = [
+        r"\btype\s*2\b.*diabet\w*",
+        r"\btype[-\s]?ii\b.*diabet\w*",
+        r"\bt2d(m)?\b",
+        r"\bdm2\b",
+        r"\btype2\b.*diabet\w*",
+        r"\bdiabetic\b"  # falls back to dataset context; kept but neg handled above
+    ]
+    if any(re.search(pat, s) for pat in pos_patterns):
+        return 1
+
+    # Case/Control shorthand
+    if s == "case":
+        return 1
+    if s == "control":
+        return 0
+
+    # Prediabetes/IGT/ambiguous -> None
+    ambig_patterns = [r"\bpre[-\s]?diabet\w*\b", r"\bigt\b", r"impaired glucose"]
+    if any(re.search(pat, s) for pat in ambig_patterns):
+        return None
+
+    return None
+
+def convert_age(x):
+    v = _after_colon(x)
+    if v is None or v == "":
+        return None
+    m = re.findall(r"[-+]?\d*\.?\d+", v)
+    if not m:
+        return None
+    try:
+        val = float(m[0])
+        if val < 0 or val > 120:
+            return None
+        return int(val) if val.is_integer() else val
+    except Exception:
+        return None
+
+def convert_gender(x):
+    v = _after_colon(x)
+    if v is None or v == "":
+        return None
+    s = v.strip().lower()
+    if s in ["female", "f", "woman", "girl", "xx"]:
+        return 0
+    if s in ["male", "m", "man", "boy", "xy"]:
+        return 1
+    return None
+
+# 3) Initial filtering metadata save
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip because trait_row is None)
+if trait_row is not None:
+    selected_clinical = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    _ = preview_df(selected_clinical)
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical.to_csv(out_clinical_data_file, index=True)
+
+# Step 3: Gene Data Extraction
+import io
+import gzip
+import pandas as pd
+
+# 1. Try the standard extractor
+gene_data = get_genetic_data(matrix_file)
+
+# Diagnostics
+print("After get_genetic_data:")
+print("  shape:", gene_data.shape)
+print("  first columns:", list(gene_data.columns[:10]))
+print("  first index values:", list(gene_data.index[:5]))
+
+# 2. If empty, try a robust manual parser for the matrix section
+if gene_data.shape[0] == 0 or len(gene_data.index) == 0:
+    print("Standard parsing returned empty. Attempting manual parsing between matrix begin/end markers...")
+    lines = []
+    in_table = False
+    with gzip.open(matrix_file, 'rt') as fh:
+        for line in fh:
+            s = line.rstrip('\n')
+            if "!series_matrix_table_begin" in s:
+                in_table = True
+                continue
+            if "!series_matrix_table_end" in s:
+                break
+            if in_table:
+                lines.append(s)
+    if lines:
+        text = "\n".join(lines)
+        try:
+            df = pd.read_csv(io.StringIO(text), sep='\t', low_memory=False)
+            # Normalize ID column
+            if 'ID_REF' in df.columns:
+                df = df.rename(columns={'ID_REF': 'ID'})
+            elif 'ID' not in df.columns:
+                # Fallback: assume first column is the identifier
+                first_col = df.columns[0]
+                df = df.rename(columns={first_col: 'ID'})
+            df['ID'] = df['ID'].astype(str)
+            df = df.set_index('ID')
+            gene_data = df
+        except Exception as e:
+            print(f"Manual parsing failed with error: {e}")
+
+# Final diagnostics and first 20 IDs
+print("Final gene_data diagnostics:")
+print("  shape:", gene_data.shape)
+print("  first columns:", list(gene_data.columns[:10]))
+print("  first index values:", list(gene_data.index[:5]))
+print(gene_data.index[:20])
+
+if gene_data.shape[0] == 0:
+    print("WARNING: The series matrix appears to be missing or empty (common for some scRNA-Seq GEO entries).")
+    print("Consider checking supplementary files or platform annotations for expression data.")

output/preprocess/Type_2_Diabetes/code/TCGA.py

@@ -0,0 +1,106 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Type_2_Diabetes"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Type_2_Diabetes/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Type_2_Diabetes/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Type_2_Diabetes/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Type_2_Diabetes/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import pandas as pd
+
+# Discover available subdirectories
+all_subdirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+# Identify cohort directories that could match Type 2 Diabetes (none expected in TCGA cancer cohorts)
+keywords = [
+    "diabetes", "type_2", "type 2", "t2d", "dm2", "type ii", "non-insulin-dependent", "niddm"
+]
+candidate_dirs = [d for d in all_subdirs if any(k in d.lower() for k in keywords)]
+
+selected_dir = None
+if len(candidate_dirs) > 0:
+    # If multiple, choose the most specific (longest name)
+    selected_dir = sorted(candidate_dirs, key=lambda x: len(x), reverse=True)[0]
+
+if selected_dir is None:
+    # No suitable TCGA cohort for Type 2 Diabetes; record and stop further processing in this step
+    _ = validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    # Load dataframes (handle gz if present)
+    compression_clin = 'gzip' if clinical_file_path.endswith('.gz') else None
+    compression_gen = 'gzip' if genetic_file_path.endswith('.gz') else None
+
+    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False, compression=compression_clin)
+    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False, compression=compression_gen)
+
+    print(clinical_df.columns.tolist())
+
+# Step 2: Initial Data Loading
+import os
+import pandas as pd
+
+# Provided subdirectories (from instruction)
+available_subdirs = [
+    'TCGA_lower_grade_glioma_and_glioblastoma_(GBMLGG)', 'TCGA_Uterine_Carcinosarcoma_(UCS)', 'TCGA_Thyroid_Cancer_(THCA)',
+    'TCGA_Thymoma_(THYM)', 'TCGA_Testicular_Cancer_(TGCT)', 'TCGA_Stomach_Cancer_(STAD)', 'TCGA_Sarcoma_(SARC)',
+    'TCGA_Rectal_Cancer_(READ)', 'TCGA_Prostate_Cancer_(PRAD)', 'TCGA_Pheochromocytoma_Paraganglioma_(PCPG)',
+    'TCGA_Pancreatic_Cancer_(PAAD)', 'TCGA_Ovarian_Cancer_(OV)', 'TCGA_Ocular_melanomas_(UVM)', 'TCGA_Mesothelioma_(MESO)',
+    'TCGA_Melanoma_(SKCM)', 'TCGA_Lung_Squamous_Cell_Carcinoma_(LUSC)', 'TCGA_Lung_Cancer_(LUNG)',
+    'TCGA_Lung_Adenocarcinoma_(LUAD)', 'TCGA_Lower_Grade_Glioma_(LGG)', 'TCGA_Liver_Cancer_(LIHC)',
+    'TCGA_Large_Bcell_Lymphoma_(DLBC)', 'TCGA_Kidney_Papillary_Cell_Carcinoma_(KIRP)', 'TCGA_Kidney_Clear_Cell_Carcinoma_(KIRC)',
+    'TCGA_Kidney_Chromophobe_(KICH)', 'TCGA_Head_and_Neck_Cancer_(HNSC)', 'TCGA_Glioblastoma_(GBM)', 'TCGA_Esophageal_Cancer_(ESCA)',
+    'TCGA_Endometrioid_Cancer_(UCEC)', 'TCGA_Colon_and_Rectal_Cancer_(COADREAD)', 'TCGA_Colon_Cancer_(COAD)',
+    'TCGA_Cervical_Cancer_(CESC)', 'TCGA_Breast_Cancer_(BRCA)', 'TCGA_Bladder_Cancer_(BLCA)', 'TCGA_Bile_Duct_Cancer_(CHOL)',
+    'TCGA_Adrenocortical_Cancer_(ACC)', 'TCGA_Acute_Myeloid_Leukemia_(LAML)'
+]
+
+# Filter to those that actually exist in filesystem (defensive)
+existing_subdirs = [d for d in available_subdirs if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+# Identify cohort directories that could match Type 2 Diabetes (none expected in TCGA cancer cohorts)
+keywords = ["diabetes", "type_2", "type 2", "t2d", "dm2", "type ii", "non-insulin-dependent", "niddm"]
+candidate_dirs = [d for d in existing_subdirs if any(k in d.lower() for k in keywords)]
+
+selected_dir = None
+if len(candidate_dirs) > 0:
+    selected_dir = sorted(candidate_dirs, key=lambda x: len(x), reverse=True)[0]
+
+if selected_dir is None:
+    # No suitable TCGA cohort for Type 2 Diabetes; record and stop further processing in this step
+    _ = validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    compression_clin = 'gzip' if clinical_file_path.endswith('.gz') else None
+    compression_gen = 'gzip' if genetic_file_path.endswith('.gz') else None
+
+    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False, compression=compression_clin)
+    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False, compression=compression_gen)
+
+    print(clinical_df.columns.tolist())

output/preprocess/Underweight/code/GSE130563.py

@@ -0,0 +1,218 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+cohort = "GSE130563"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Underweight"
+in_cohort_dir = "../DATA/GEO/Underweight/GSE130563"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/GSE130563.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/GSE130563.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/GSE130563.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+
+# 1) Gene expression data availability
+is_gene_available = True  # Microarray gene expression data per series description
+
+# 2) Variable availability and converters
+# Trait: Underweight is not available in this dataset (no BMI/weight status)
+trait_row = None
+
+# Age: available at key 4
+age_row = 4
+
+# Gender: available at key 1
+gender_row = 1
+
+def convert_trait(v):
+    # Underweight not available in this cohort
+    return None
+
+def _extract_value_after_colon(v):
+    if v is None:
+        return None
+    s = str(v)
+    if ':' in s:
+        s = s.split(':', 1)[1]
+    return s.strip()
+
+def convert_age(v):
+    s = _extract_value_after_colon(v)
+    if not s or s.lower() in {"n/a", "na", "nan", "unknown", "n.d. (not determined)", "nd", "n.d."}:
+        return None
+    # extract first number (integer or float)
+    m = re.search(r"[-+]?\d*\.?\d+", s)
+    if not m:
+        return None
+    try:
+        val = float(m.group(0))
+        return val
+    except Exception:
+        return None
+
+def convert_gender(v):
+    s = _extract_value_after_colon(v)
+    if not s:
+        return None
+    s = s.strip().lower()
+    # Map female -> 0, male -> 1
+    if s in {"f", "female"}:
+        return 0
+    if s in {"m", "male"}:
+        return 1
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip if trait not available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    _preview = preview_df(selected_clinical_df)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Identify the probe ID column in the annotation
+id_col = None
+for cand in ['ID', 'ID_REF', 'ProbeID', 'PROBEID']:
+    if cand in gene_annotation.columns:
+        id_col = cand
+        break
+
+# Heuristically identify a gene symbol column: pick the column where a reasonable fraction of entries
+# yield human-like gene symbols via extract_human_gene_symbols
+gene_col = None
+best_frac = 0.0
+threshold = 0.1  # require at least 10% of sampled entries to look like gene symbols
+
+if id_col is not None:
+    candidate_cols = [c for c in gene_annotation.columns if c != id_col]
+else:
+    candidate_cols = list(gene_annotation.columns)
+
+for c in candidate_cols:
+    ser = gene_annotation[c].dropna().astype(str).head(1000)
+    if ser.empty:
+        continue
+    frac = ser.map(lambda x: len(extract_human_gene_symbols(x)) > 0).mean()
+    if frac > best_frac:
+        best_frac = frac
+        gene_col = c
+
+# Build mapping and convert probe-level data to gene-level if the selected gene column looks valid
+mapping_succeeded = False
+if id_col is not None and gene_col is not None and best_frac >= threshold:
+    mapping_df = get_gene_mapping(gene_annotation, prob_col=id_col, gene_col=gene_col)
+
+    n_total_probes = gene_data.shape[0]
+    n_mapping_rows = len(mapping_df)
+    n_mapping_rows_in_expr = (mapping_df['ID'].isin(gene_data.index)).sum()
+    n_unique_probes_in_expr = mapping_df.loc[mapping_df['ID'].isin(gene_data.index), 'ID'].nunique()
+
+    # Estimate number of genes that can be mapped
+    tmp = mapping_df.copy()
+    tmp['Gene'] = tmp['Gene'].apply(extract_human_gene_symbols)
+    tmp = tmp.explode('Gene').dropna(subset=['Gene'])
+    n_unique_genes_est = tmp['Gene'].nunique()
+
+    print(f"Probe IDs in expression: {n_total_probes}")
+    print(f"Annotation rows: {n_mapping_rows}")
+    print(f"Annotation rows matching expression probes: {n_mapping_rows_in_expr} (unique probes: {n_unique_probes_in_expr})")
+    print(f"Estimated unique genes from annotation column '{gene_col}': {n_unique_genes_est}")
+
+    # Only proceed if we have a reasonable number of genes
+    if n_unique_genes_est >= 100:
+        probe_data = gene_data
+        gene_data_mapped = apply_gene_mapping(probe_data, mapping_df)
+        print(f"Mapped gene-level dataframe shape: {gene_data_mapped.shape}")
+        # Sanity check: ensure non-empty and fewer rows than probes (typical for mapping)
+        if gene_data_mapped.shape[0] > 0:
+            gene_data = gene_data_mapped
+            mapping_succeeded = True
+
+if not mapping_succeeded:
+    print("WARNING: No valid gene symbol column found or mapping produced too few genes. "
+          "Keeping probe-level data without mapping.")
+
+# Step 7: Data Normalization and Linking
+# Normalize gene symbols; if normalization yields no genes (likely due to probe-level IDs), fall back to probe-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+if normalized_gene_data.shape[0] > 0:
+    normalized_gene_data.to_csv(out_gene_data_file)
+else:
+    # Fallback: save original probe-level expression since trait is unavailable and mapping failed
+    gene_data.to_csv(out_gene_data_file)
+
+# Trait is unavailable (from Step 2), so skip linking and downstream steps.
+is_trait_available = False
+is_gene_available = True  # Gene expression data is present (at least at probe-level)
+
+# Record metadata using initial filtering mode to avoid abnormality override
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)

output/preprocess/Underweight/code/GSE131835.py

@@ -0,0 +1,175 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+cohort = "GSE131835"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Underweight"
+in_cohort_dir = "../DATA/GEO/Underweight/GSE131835"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/GSE131835.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/GSE131835.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/GSE131835.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Step 1: Determine gene expression availability
+is_gene_available = True  # Affymetrix Clariom S Microarray indicates mRNA gene expression data.
+
+# Step 2: Identify variable availability
+# Trait: Underweight is not explicitly provided; BMI cannot be computed with the single-row extraction constraint.
+trait_row = None
+
+# Age and Gender are available
+age_row = 3
+gender_row = 2
+
+# Step 2.2: Define converters
+def _extract_value(x):
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        val = parts[1].strip() if len(parts) > 1 else x.strip()
+        if val in ["", "N/A", "NA", "na", "NaN", "nan", "None", "null"]:
+            return None
+        return val
+    return x
+
+def convert_trait(x):
+    # Underweight not available in this dataset under a single key; return None
+    return None
+
+def convert_age(x):
+    v = _extract_value(x)
+    if v is None:
+        return None
+    try:
+        # Age in years, continuous
+        return float(v)
+    except Exception:
+        return None
+
+def convert_gender(x):
+    v = _extract_value(x)
+    if v is None:
+        return None
+    v_lower = str(v).strip().lower()
+    if v_lower in ["male", "m"]:
+        return 1
+    if v_lower in ["female", "f"]:
+        return 0
+    return None
+
+# Step 3: Initial filtering metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Step 4: Clinical feature extraction (skip since trait_row is None)
+# If trait_row were available:
+# selected_clinical_df = geo_select_clinical_features(
+#     clinical_df=clinical_data,
+#     trait=trait,
+#     trait_row=trait_row,
+#     convert_trait=convert_trait,
+#     age_row=age_row,
+#     convert_age=convert_age,
+#     gender_row=gender_row,
+#     convert_gender=convert_gender
+# )
+# preview = preview_df(selected_clinical_df, n=5)
+# os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+# selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Identify the appropriate columns for mapping: 'ID' for probe IDs and 'ORF' for gene symbols
+mapping_df = get_gene_mapping(gene_annotation, prob_col='ID', gene_col='ORF')
+
+# Apply the mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(gene_data, mapping_df)
+
+# Step 7: Data Normalization and Linking
+# 1. Normalize gene symbols and save gene-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+import os
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2-6. Since trait data was not extracted earlier (trait_row was None), skip linking and finalize metadata
+try:
+    trait_available = (trait_row is not None)
+except NameError:
+    trait_available = False
+
+if trait_available and ('selected_clinical_data' in locals()) and (selected_clinical_data is not None):
+    # Full pipeline if trait and clinical data are available (not expected for this cohort)
+    linked_data = geo_link_clinical_genetic_data(selected_clinical_data, normalized_gene_data)
+    linked_data = handle_missing_values(linked_data, trait)
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+    is_usable = validate_and_save_cohort_info(
+        True, cohort, json_path, True, True, is_trait_biased, unbiased_linked_data
+    )
+    if is_usable:
+        os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+        unbiased_linked_data.to_csv(out_data_file)
+else:
+    # Trait unavailable: perform final validation, mark dataset as not available for analysis
+    note = "INFO: Trait not available; cannot link clinical and genetic data."
+    temp_df = normalized_gene_data.T  # non-empty df to avoid false abnormality override
+    _ = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=False,
+        is_biased=True,  # ignored since dataset is not available
+        df=temp_df,
+        note=note
+    )

output/preprocess/Underweight/code/GSE50982.py

@@ -0,0 +1,234 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+cohort = "GSE50982"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Underweight"
+in_cohort_dir = "../DATA/GEO/Underweight/GSE50982"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/GSE50982.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/GSE50982.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/GSE50982.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+
+# 1) Gene expression data availability
+# This series is a cell-line experiment with knockdown and EGF treatment, consistent with mRNA expression profiling.
+is_gene_available = True
+
+# 2) Variable availability (from Sample Characteristics Dictionary)
+# No human clinical variables are present; only cell line, knockdown days, and treatment.
+trait_row = None      # Underweight status not available
+age_row = None        # Age not available
+gender_row = None     # Gender not available
+
+# 2.2) Converters
+def _after_colon(x):
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        x = parts[1] if len(parts) > 1 else parts[0]
+        return x.strip()
+    return x
+
+def convert_trait(x):
+    # Binary: underweight (1) vs not (0)
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower()
+
+    # Numeric BMI extraction
+    nums = re.findall(r"[-+]?\d*\.?\d+", s)
+    if nums:
+        try:
+            val = float(nums[0])
+            # Heuristic: treat as BMI if in a plausible range
+            if 10 <= val <= 60:
+                return 1 if val < 18.5 else 0
+        except Exception:
+            pass
+
+    # Keyword heuristics
+    positive_kw = ["underweight", "cachexia", "malnourished", "low bmi", "bmi<18.5"]
+    negative_kw = ["normal weight", "healthy weight", "non-underweight"]
+    if any(k in s for k in positive_kw):
+        return 1
+    if any(k in s for k in negative_kw):
+        return 0
+
+    # Avoid mapping generic "case/control" or "yes/no" without context
+    return None
+
+def convert_age(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).lower()
+    nums = re.findall(r"[-+]?\d*\.?\d+", s)
+    if not nums:
+        return None
+    try:
+        age = float(nums[0])
+        # Basic plausibility for human age
+        if 0 <= age <= 120:
+            return age
+    except Exception:
+        return None
+    return None
+
+def convert_gender(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower()
+    # female -> 0, male -> 1
+    female_terms = {"female", "f", "woman", "girl"}
+    male_terms = {"male", "m", "man", "boy"}
+    if s in female_terms:
+        return 0
+    if s in male_terms:
+        return 1
+    return None
+
+# 3) Initial filtering and save metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip since trait_row is None)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    _ = preview_df(selected_clinical_df)
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file, index=True)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Extract gene annotation data from the SOFT file
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Preview column names and first few values to identify probe and gene symbol columns
+print("Gene annotation columns:")
+print(list(gene_annotation.columns))
+
+print("\nGene annotation preview (first 5 rows as dict):")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# 1-2. Decide mapping columns and extract mapping dataframe
+# Probe identifiers match the 'ID' column; gene symbols are in the 'Symbol' column.
+mapping_df = get_gene_mapping(gene_annotation, prob_col='ID', gene_col='Symbol')
+
+# 3. Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+import pandas as pd
+
+# 1) Normalize gene symbols and save gene data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2-6) Handle absence of clinical trait data gracefully
+if 'selected_clinical_data' in globals() and isinstance(selected_clinical_data, pd.DataFrame):
+    # Link clinical and genetic data
+    linked_data = geo_link_clinical_genetic_data(selected_clinical_data, normalized_gene_data)
+
+    # Ensure trait column exists before proceeding
+    if trait in linked_data.columns:
+        # 3. Handle missing values
+        linked_data = handle_missing_values(linked_data, trait)
+
+        # 4. Bias checking and remove biased demographic features
+        is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+        # 5. Final validation and save cohort info
+        note = "INFO: Linked data generated with clinical features and normalized gene expression."
+        is_usable = validate_and_save_cohort_info(
+            is_final=True,
+            cohort=cohort,
+            info_path=json_path,
+            is_gene_available=True,
+            is_trait_available=True,
+            is_biased=is_trait_biased,
+            df=unbiased_linked_data,
+            note=note
+        )
+
+        # 6. Save usable linked data
+        if is_usable:
+            os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+            unbiased_linked_data.to_csv(out_data_file)
+    else:
+        # Trait column missing unexpectedly; record unusable with initial filtering to avoid abnormality override
+        _ = validate_and_save_cohort_info(
+            is_final=False,
+            cohort=cohort,
+            info_path=json_path,
+            is_gene_available=True,
+            is_trait_available=False
+        )
+else:
+    # No clinical data extracted (cell line model); record unusable with initial filtering to avoid abnormality override
+    _ = validate_and_save_cohort_info(
+        is_final=False,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=False
+    )

output/preprocess/Underweight/code/GSE57802.py

@@ -0,0 +1,205 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+cohort = "GSE57802"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Underweight"
+in_cohort_dir = "../DATA/GEO/Underweight/GSE57802"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/GSE57802.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/GSE57802.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/GSE57802.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+import pandas as pd
+
+# 1. Gene Expression Data Availability
+is_gene_available = True  # Transcriptome profiling of LCLs => gene expression data likely available
+
+# 2. Variable Availability and Data Type Conversion
+
+# Keys in Sample Characteristics Dictionary:
+# trait (Underweight inferred from genotype: duplication -> underweight)
+trait_row = 4  # 'genotype: 600kbdel/600kbdup/Control'
+# age
+age_row = 2    # 'age: ...'
+# gender
+gender_row = 1 # 'gender: M/F'
+
+def _after_colon(x):
+    if x is None:
+        return None
+    s = str(x)
+    parts = s.split(':', 1)
+    return parts[1].strip() if len(parts) > 1 else s.strip()
+
+def convert_trait(x):
+    """
+    Binary: Underweight (1) vs not (0)
+    Heuristic:
+      - genotype: '600kbdup' => 1; '600kbdel' or 'Control' => 0
+      - copy number 16p11.2: 3 => 1; 1 or 2 => 0
+    """
+    val = _after_colon(x)
+    if val is None or val == '':
+        return None
+    lv = val.lower()
+    # Textual mapping
+    if 'dup' in lv:
+        return 1
+    if 'del' in lv or 'control' in lv or 'wt' in lv:
+        return 0
+    # Numeric mapping (e.g., copy number)
+    try:
+        n = float(val)
+        if n == 3:
+            return 1
+        if n in (1, 2):
+            return 0
+    except Exception:
+        pass
+    return None
+
+def convert_age(x):
+    """
+    Continuous: age in years (float). 'NA' -> None
+    """
+    val = _after_colon(x)
+    if val is None:
+        return None
+    lv = val.strip().lower()
+    if lv in {'na', 'n/a', '', 'nan'}:
+        return None
+    try:
+        return float(val)
+    except Exception:
+        return None
+
+def convert_gender(x):
+    """
+    Binary: female -> 0, male -> 1
+    """
+    val = _after_colon(x)
+    if val is None:
+        return None
+    lv = val.strip().lower()
+    if lv in {'m', 'male'}:
+        return 1
+    if lv in {'f', 'female'}:
+        return 0
+    return None
+
+# 3. Save Metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4. Clinical Feature Extraction (only if clinical data is available)
+if is_trait_available:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    # Preview and save
+    print(preview_df(selected_clinical_df))
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# 1-2. Identify the appropriate columns for probe IDs and gene symbols and extract mapping dataframe
+probe_col = 'ID'
+gene_symbol_col = 'Gene Symbol'
+mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_col, gene_col=gene_symbol_col)
+
+# 3. Apply the mapping to convert probe-level data to gene-level expression data
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize the obtained gene data with the 'normalize_gene_symbols_in_index' function from the library.
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link the clinical and genetic data with the 'geo_link_clinical_genetic_data' function from the library.
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values in the linked data
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Determine whether the trait and some demographic features are severely biased, and remove biased features.
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Conduct quality check and save the cohort information.
+is_usable = validate_and_save_cohort_info(
+    True,
+    cohort,
+    json_path,
+    is_gene_available,
+    is_trait_available,
+    is_trait_biased,
+    unbiased_linked_data,
+    note="INFO: Probe-to-gene mapping applied; trait inferred from 16p11.2 duplication vs others."
+)
+
+# 6. If the linked data is usable, save it as a CSV file to 'out_data_file'.
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Underweight/code/GSE84954.py

@@ -0,0 +1,305 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+cohort = "GSE84954"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Underweight"
+in_cohort_dir = "../DATA/GEO/Underweight/GSE84954"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/GSE84954.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/GSE84954.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/GSE84954.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+
+# 1. Gene Expression Data Availability
+is_gene_available = True  # Microarray gene expression data from tissues (not miRNA/methylation)
+
+# 2. Variable Availability and Data Type Conversion
+
+# Heuristic: Use 'disease' to infer Underweight/cachexia status.
+# Rationale: The study compares end-stage liver disease (cachexia context) vs. Crigler-Najjar controls.
+# Map end-stage liver disease subtypes/Alagille to 1 (underweight/cachexia context) and Crigler-Najjar to 0 (controls).
+trait_row = 1  # 'disease' field
+age_row = None
+gender_row = None
+
+def _extract_value(x):
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        val = parts[1].strip() if len(parts) == 2 else x.strip()
+        return val if val != "" else None
+    return x
+
+def convert_trait(x):
+    # Binary: 1 = underweight/cachexia context (chronic liver disease subtypes, Alagille)
+    #         0 = control (Crigler-Najjar)
+    val = _extract_value(x)
+    if val is None:
+        return None
+    v = val.lower()
+    if "crigler" in v:  # controls
+        return 0
+    positive_signals = [
+        "chronic liver disease", "alagille", "biliary atresia", "ba",
+        "biliary cirrhosis", "bc", "neonatal sclerosing cholangitis", "nsc",
+        "alpha-1-antitrypsin", "a1at", "a1-at", "a1 at"
+    ]
+    if any(sig in v for sig in positive_signals):
+        return 1
+    return None
+
+def convert_age(x):
+    # Continuous age in years if possible
+    val = _extract_value(x)
+    if val is None:
+        return None
+    v = val.lower()
+    try:
+        return float(v)
+    except:
+        pass
+    m = re.search(r'([0-9]*\.?[0-9]+)\s*(year|yr|y|month|mo|m(?!ale)|day|d)s?', v)
+    if m:
+        num = float(m.group(1))
+        unit = m.group(2)
+        if unit in ["year", "yr", "y"]:
+            return num
+        if unit in ["month", "mo", "m"]:
+            return num / 12.0
+        if unit in ["day", "d"]:
+            return num / 365.0
+    return None
+
+def convert_gender(x):
+    # Binary: female=0, male=1
+    val = _extract_value(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if v in ["female", "f", "girl", "woman", "women"]:
+        return 0
+    if v in ["male", "m", "boy", "man", "men"]:
+        return 1
+    return None
+
+# 3. Save Metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4. Clinical Feature Extraction
+if trait_row is not None:
+    # Ensure clinical_data exists (expected from previous step)
+    if "clinical_data" not in globals() or clinical_data is None:
+        raise RuntimeError("clinical_data not found; please ensure it is loaded from the previous step before running feature extraction.")
+
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected_clinical_df)
+    print(preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+# The provided identifiers are numeric probe IDs, not human gene symbols.
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+import os
+import re
+import json
+import pandas as pd
+
+# Preserve the original probe-level matrix for mapping
+probe_level_data = gene_data.copy()
+
+# 1) Identify probe ID column and try to find a usable gene symbol column
+probe_col = 'ID'  # matches both annotation and probe-level data
+
+def pick_gene_symbol_column(df, min_symbol_frac: float = 0.05):
+    priority_names = [
+        'Gene Symbol', 'GENE_SYMBOL', 'Gene symbol', 'SYMBOL', 'Symbol',
+        'Gene Symbols', 'GENE_SYMBOLS', 'gene_symbol', 'gene symbols',
+        'Gene name', 'GENE_NAME', 'gene_name', 'gene assignment', 'GENE_ASSIGNMENT'
+    ]
+    cols = [c for c in df.columns if c != probe_col]
+    prioritized = [c for c in cols if any(pn.lower() == str(c).lower() for pn in priority_names)]
+    if not prioritized:
+        prioritized = [c for c in cols if re.search(r'symbol|gene|assign', str(c), re.I)]
+    candidates = prioritized if prioritized else cols
+
+    scores = {}
+    best_col = None
+    best_score = -1.0
+    for c in candidates:
+        s = df[c].dropna().astype(str)
+        if s.empty:
+            scores[c] = 0.0
+            continue
+        sample = s.head(10000)
+        hits = sample.map(lambda x: 1 if len(extract_human_gene_symbols(x)) > 0 else 0)
+        score = float(hits.mean())
+        scores[c] = score
+        if score > best_score:
+            best_score = score
+            best_col = c
+
+    print("Candidate gene symbol columns and scores (fraction with >=1 symbol):")
+    for cname, sc in sorted(scores.items(), key=lambda x: -x[1]):
+        print(f"  {cname}: {sc:.3f}")
+
+    if best_col is None or best_score < min_symbol_frac:
+        return None, 0.0
+    return best_col, best_score
+
+gene_symbol_col, score = pick_gene_symbol_column(gene_annotation)
+if gene_symbol_col:
+    print(f"Selected gene symbol column: {gene_symbol_col} (score={score:.3f})")
+else:
+    print("No direct gene symbol column found in the platform annotation. Attempting fallbacks...")
+
+mapping_df = pd.DataFrame(columns=['ID', 'Gene'])
+
+# 2) Primary mapping if a gene symbol column is available
+if gene_symbol_col:
+    mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_col, gene_col=gene_symbol_col)
+
+# 2b) Fallback: try RefSeq accession (GB_ACC) -> symbol via metadata mapping file if available
+if mapping_df.empty:
+    if 'GB_ACC' in gene_annotation.columns:
+        refseq_series = gene_annotation[['ID', 'GB_ACC']].dropna()
+        refseq_series['GB_ACC'] = refseq_series['GB_ACC'].astype(str)
+        refseq_map_path = "./metadata/refseq_to_gene_symbol.json"
+        symbol_from_refseq = None
+        if os.path.exists(refseq_map_path):
+            try:
+                with open(refseq_map_path, "r") as f:
+                    symbol_from_refseq = json.load(f)
+            except Exception as e:
+                print(f"WARNING: Failed to load RefSeq->Symbol mapping file: {e}")
+        if symbol_from_refseq:
+            refseq_series['Gene'] = refseq_series['GB_ACC'].map(symbol_from_refseq)
+            mapping_df = refseq_series[['ID', 'Gene']].dropna()
+            # Keep only plausible human gene symbols
+            mapping_df['Gene'] = mapping_df['Gene'].astype(str)
+            mapping_df = mapping_df[mapping_df['Gene'].map(lambda x: len(extract_human_gene_symbols(x)) > 0)]
+        else:
+            print("INFO: No RefSeq->Symbol mapping file found. Skipping RefSeq-based mapping fallback.")
+
+# 2c) If still empty, fail gracefully: record unavailability and stop mapping
+if mapping_df.empty:
+    print("ERROR: Could not derive probe->gene mapping for this platform (likely a tiling array without gene symbols).")
+    # Ensure we have clinical data to record metadata
+    try:
+        clinical_df_for_meta = selected_clinical_df.T if 'selected_clinical_df' in globals() else None
+        if clinical_df_for_meta is None or clinical_df_for_meta.empty:
+            if os.path.exists(out_clinical_data_file):
+                clinical_df_for_meta = pd.read_csv(out_clinical_data_file, index_col=0).T
+    except Exception:
+        clinical_df_for_meta = None
+
+    note = ("ERROR: Platform annotation lacks usable gene symbols; IDs appear to be genomic tiling probes. "
+            "Probe-to-gene mapping not feasible.")
+    _ = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=True,
+        is_biased=False,
+        df=(clinical_df_for_meta if clinical_df_for_meta is not None else pd.DataFrame({"Underweight": []})),
+        note=note
+    )
+else:
+    # 3) Apply mapping to convert probe-level data to gene-level expression
+    print("Mapping preview (first 5 rows):")
+    print(mapping_df.head())
+    gene_data = apply_gene_mapping(expression_df=probe_level_data, mapping_df=mapping_df)
+
+    if gene_data.empty or gene_data.shape[1] == 0:
+        print("ERROR: Resulting gene-level expression is empty after applying mapping.")
+        # Record metadata as above
+        try:
+            clinical_df_for_meta = selected_clinical_df.T if 'selected_clinical_df' in globals() else None
+            if clinical_df_for_meta is None or clinical_df_for_meta.empty:
+                if os.path.exists(out_clinical_data_file):
+                    clinical_df_for_meta = pd.read_csv(out_clinical_data_file, index_col=0).T
+        except Exception:
+            clinical_df_for_meta = None
+
+        note = ("ERROR: Mapping produced empty gene-level data. "
+                "This likely reflects lack of valid gene identifiers in platform annotation.")
+        _ = validate_and_save_cohort_info(
+            is_final=True,
+            cohort=cohort,
+            info_path=json_path,
+            is_gene_available=False,
+            is_trait_available=True,
+            is_biased=False,
+            df=(clinical_df_for_meta if clinical_df_for_meta is not None else pd.DataFrame({"Underweight": []})),
+            note=note
+        )
+    else:
+        print(f"Gene-level data shape: {gene_data.shape}")
+        # Optionally save gene-level data for downstream steps
+        os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+        gene_data.to_csv(out_gene_data_file)

output/preprocess/Underweight/code/TCGA.py

@@ -0,0 +1,14 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Underweight"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Underweight/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Underweight/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Underweight/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Underweight/cohort_info.json"

output/preprocess/Underweight/cohort_info.json

@@ -1,62 +1 @@
-{
-  "GSE84954": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 37
-  },
-  "GSE57802": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 99
-  },
-  "GSE50982": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE131835": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 48
-  },
-  "GSE130563": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 38
-  },
-  "TCGA": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 183
-  }
-}
+{"GSE84954": {"is_usable": false, "is_gene_available": false, "is_trait_available": true, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "ERROR: Platform annotation lacks usable gene symbols; IDs appear to be genomic tiling probes. Probe-to-gene mapping not feasible."}, "GSE57802": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": true, "has_gender": true, "sample_size": 99, "note": "INFO: Probe-to-gene mapping applied; trait inferred from 16p11.2 duplication vs others."}, "GSE50982": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE131835": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "INFO: Trait not available; cannot link clinical and genetic data."}, "GSE130563": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}}

output/preprocess/Uterine_Carcinosarcoma/clinical_data/GSE32507.csv

@@ -1,2 +1,2 @@
-,0
-Uterine_Carcinosarcoma,1.0
+,GSM804806,GSM804807,GSM804808,GSM804809,GSM804810,GSM804811,GSM804812,GSM804813,GSM804814,GSM804815,GSM804816,GSM804817,GSM804818,GSM804819,GSM804820,GSM804821,GSM804822,GSM804823,GSM804824,GSM804825,GSM804826,GSM804827,GSM804828,GSM804829,GSM804830,GSM804831,GSM804832,GSM804833,GSM804834,GSM804835,GSM804836,GSM804837,GSM804838,GSM804839,GSM804840,GSM804841,GSM804842,GSM804843,GSM804844,GSM804845,GSM804846,GSM804847,GSM804848,GSM804849,GSM804850,GSM804851
+Uterine_Carcinosarcoma,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,1.0,0.0,0.0,0.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0

output/preprocess/Uterine_Carcinosarcoma/code/GSE32507.py

@@ -0,0 +1,193 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Carcinosarcoma"
+cohort = "GSE32507"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Uterine_Carcinosarcoma"
+in_cohort_dir = "../DATA/GEO/Uterine_Carcinosarcoma/GSE32507"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/GSE32507.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/gene_data/GSE32507.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/clinical_data/GSE32507.csv"
+json_path = "./output/z6/preprocess/Uterine_Carcinosarcoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import pandas as pd
+
+# 1) Gene expression availability (cDNA microarray => gene expression is available)
+is_gene_available = True
+
+# 2) Variable availability
+# Trait (Uterine_Carcinosarcoma): can be inferred from "tissue" field
+trait_row = 0
+
+# Age and Gender are not available in the provided characteristics
+age_row = None
+gender_row = None
+
+# 2.2) Converters
+def _extract_value(cell):
+    if pd.isna(cell):
+        return None
+    s = str(cell)
+    # Take the substring after the last colon, which is usually the value
+    if ':' in s:
+        s = s.split(':')[-1]
+    return s.strip()
+
+def convert_trait(cell):
+    val = _extract_value(cell)
+    if val is None:
+        return None
+    low = val.lower()
+    # Prioritize carcinosarcoma mapping to avoid substring issues with 'sarcoma'
+    if ('carcinosarcoma' in low) or ('malignant mixed müllerian' in low) or ('malignant mixed mullerian' in low) or (low == 'cs'):
+        return 1
+    # Non-CS uterine tumors in this study act as controls
+    if ('endometrioid' in low) or ('adenocarcinoma' in low) or ('sarcoma' in low) or (low in {'ec','us'}):
+        return 0
+    return None
+
+def convert_age(cell):
+    # Not used (age_row is None), but provided for completeness
+    val = _extract_value(cell)
+    if val is None:
+        return None
+    # Extract first numeric token (e.g., "65 years" -> 65)
+    import re
+    m = re.search(r'(\d+(\.\d+)?)', val)
+    if m:
+        try:
+            return float(m.group(1))
+        except Exception:
+            return None
+    return None
+
+def convert_gender(cell):
+    # Not used (gender_row is None), but provided for completeness
+    val = _extract_value(cell)
+    if val is None:
+        return None
+    low = val.lower()
+    if 'female' in low or low in {'f'}:
+        return 0
+    if 'male' in low or low in {'m'}:
+        return 1
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (only if trait_row is available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=None,
+        gender_row=gender_row,
+        convert_gender=None
+    )
+    preview = preview_df(selected_clinical_df)
+    print("Preview of selected clinical features:", preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Decide the columns for probe IDs and gene symbols based on annotation preview
+probe_id_col = 'ID'
+gene_symbol_col = 'GENE_SYMBOL'
+
+# 2) Build the mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_id_col, gene_col=gene_symbol_col)
+
+# 3) Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save gene-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Bias assessment and removal of biased covariates
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Final validation and metadata saving
+note_msg = "INFO: Only trait available; no age or gender annotations. Gene symbols normalized via NCBI synonym mapping."
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=True,
+    is_trait_available=True,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note_msg
+)
+
+# 6. Save linked dataset if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Uterine_Carcinosarcoma/code/GSE36133.py

@@ -0,0 +1,134 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Carcinosarcoma"
+cohort = "GSE36133"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Uterine_Carcinosarcoma"
+in_cohort_dir = "../DATA/GEO/Uterine_Carcinosarcoma/GSE36133"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/GSE36133.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/gene_data/GSE36133.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/clinical_data/GSE36133.csv"
+json_path = "./output/z6/preprocess/Uterine_Carcinosarcoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Step 1: Determine gene expression data availability
+is_gene_available = True  # CCLE series contains gene expression data (not pure miRNA/methylation)
+
+# Step 2: Variable availability and converters
+
+# Based on the provided sample characteristics dictionary, there is no explicit field for uterine carcinosarcoma.
+# Available keys:
+# 0 -> primary site (e.g., 'primary site: endometrium', etc.)
+# 1 -> histology (e.g., 'histology: carcinoma', 'histology: sarcoma', etc.)
+# 2 -> histology subtype1 (does not include carcinosarcoma)
+# Therefore, trait (Uterine Carcinosarcoma) cannot be reliably inferred from any single key. Age and gender are absent.
+trait_row = None
+age_row = None
+gender_row = None
+
+def _get_value_after_colon(x):
+    if x is None:
+        return None
+    s = str(x)
+    if ':' in s:
+        s = s.split(':', 1)[1]
+    return s.strip().strip('"').strip()
+
+def convert_trait(x):
+    # Not used since trait_row is None, but implemented for completeness:
+    # Map to 1 if value clearly indicates carcinosarcoma/MMMT, else 0; unknown -> None.
+    val = _get_value_after_colon(x)
+    if val is None or val == '' or val.lower() in {'na', 'n/a', 'null', 'none', 'unknown'}:
+        return None
+    v = val.lower().replace(' ', '_')
+    keywords = ['carcinosarcoma', 'malignant_mixed_mullerian', 'mmmt', 'malignant_mixed_müllerian']
+    if any(k in v for k in keywords):
+        return 1
+    # If we are confident it's not carcinosarcoma
+    return 0
+
+def convert_age(x):
+    # Continuous age if present; return float or None
+    val = _get_value_after_colon(x)
+    if val is None:
+        return None
+    v = val.lower().replace('years', '').replace('year', '').strip()
+    if v in {'na', 'n/a', 'null', 'none', 'unknown', ''}:
+        return None
+    try:
+        return float(v)
+    except Exception:
+        # try to extract number
+        import re
+        m = re.search(r'[-+]?\d*\.?\d+', v)
+        if m:
+            try:
+                return float(m.group(0))
+            except Exception:
+                return None
+        return None
+
+def convert_gender(x):
+    # Binary: female -> 0, male -> 1; unknown -> None
+    val = _get_value_after_colon(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if v in {'female', 'f', 'woman', 'women'}:
+        return 0
+    if v in {'male', 'm', 'man', 'men'}:
+        return 1
+    if v in {'na', 'n/a', 'null', 'none', 'unknown', ''}:
+        return None
+    return None
+
+# Step 3: Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Step 4: Clinical Feature Extraction (skip because trait_row is None)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected_clinical_df, n=5)
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file, index=True)

output/preprocess/Uterine_Carcinosarcoma/code/GSE36138.py

@@ -0,0 +1,120 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Carcinosarcoma"
+cohort = "GSE36138"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Uterine_Carcinosarcoma"
+in_cohort_dir = "../DATA/GEO/Uterine_Carcinosarcoma/GSE36138"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/GSE36138.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/gene_data/GSE36138.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/clinical_data/GSE36138.csv"
+json_path = "./output/z6/preprocess/Uterine_Carcinosarcoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Step 1: Determine gene expression availability based on background info (SNP array -> not gene expression)
+is_gene_available = False
+
+# Step 2: Variable availability and conversion functions
+# Based on the sample characteristics, there is no explicit or inferable uterine carcinosarcoma label, nor age/gender for human subjects in CCLE cell lines.
+trait_row = None
+age_row = None
+gender_row = None
+
+def _after_colon(x):
+    if x is None:
+        return None
+    s = str(x)
+    parts = s.split(":", 1)
+    val = parts[1] if len(parts) > 1 else parts[0]
+    val = val.strip().strip('"').strip()
+    return val if val != "" else None
+
+def convert_trait(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    # Positive mappings
+    positive_terms = [
+        "carcinosarcoma",
+        "mmmt",
+        "malignant mixed mullerian tumor",
+        "malignant mixed müllerian tumor",
+        "uterine carcinosarcoma"
+    ]
+    if any(t in v_low for t in positive_terms):
+        return 1
+    # Otherwise treat as non-trait if clearly a different histology/site label
+    return 0
+
+def convert_age(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v = v.replace("years", "").replace("year", "").replace("yrs", "").replace("yr", "").strip()
+    try:
+        return float(v)
+    except Exception:
+        return None
+
+def convert_gender(x):
+    v = _after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    if v_low in ["female", "f", "0", "woman", "women"]:
+        return 0
+    if v_low in ["male", "m", "1", "man", "men"]:
+        return 1
+    return None
+
+# Step 3: Initial filtering and save metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Step 4: Clinical feature extraction (skip because trait_row is None)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    _ = preview_df(selected_clinical_df)
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)

output/preprocess/Uterine_Carcinosarcoma/code/GSE68950.py

@@ -0,0 +1,222 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Carcinosarcoma"
+cohort = "GSE68950"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Uterine_Carcinosarcoma"
+in_cohort_dir = "../DATA/GEO/Uterine_Carcinosarcoma/GSE68950"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/GSE68950.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/gene_data/GSE68950.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/clinical_data/GSE68950.csv"
+json_path = "./output/z6/preprocess/Uterine_Carcinosarcoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import pandas as pd
+
+# 1) Gene expression availability
+is_gene_available = True  # Assay Type indicates Gene Expression (Affymetrix HT_HG-U133A)
+
+# 2) Variable availability setup
+candidate_trait_row = 1   # 'disease state' is the closest field for trait mapping
+age_row = None            # Not available in the provided characteristics
+gender_row = None         # Not available in the provided characteristics
+
+# Converters
+def _extract_value(x):
+    if x is None or (isinstance(x, float) and pd.isna(x)):
+        return None
+    s = str(x)
+    if ':' in s:
+        s = s.split(':', 1)[1]
+    return s.strip()
+
+def convert_trait(x):
+    v = _extract_value(x)
+    if v is None:
+        return None
+    s = v.lower()
+    # Map carcinosarcoma cases to 1; others to 0
+    if 'carcinosarcoma' in s:
+        return 1
+    return 0
+
+def convert_age(x):
+    return None  # Not available
+
+def convert_gender(x):
+    return None  # Not available
+
+# 2.1/2.2) Determine actual trait availability; enforce constant-feature check
+trait_row = None
+if candidate_trait_row in clinical_data.index:
+    raw_vals = clinical_data.loc[candidate_trait_row]
+    converted_vals = pd.Series([convert_trait(v) for v in raw_vals])
+    unique_non_missing = converted_vals.dropna().nunique()
+    # Require at least two classes (e.g., both 0 and 1) for availability
+    if unique_non_missing >= 2:
+        trait_row = candidate_trait_row
+
+# 3) Initial filtering metadata saving
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction: only proceed if trait is available
+if is_trait_available:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df)
+    print("Clinical data preview:", clinical_preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Identify annotation columns for probe IDs and gene symbols based on the preview
+probe_col = 'ID'
+gene_symbol_col = 'Gene Symbol'
+
+# Build the mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_col, gene_col=gene_symbol_col)
+
+# Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(gene_data, mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+import json
+import pandas as pd
+
+# 1. Normalize gene symbols and save gene expression data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# Ensure clinical features are available (reload if not in memory)
+if 'selected_clinical_df' not in globals() or selected_clinical_df is None:
+    if os.path.exists(out_clinical_data_file):
+        selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)
+
+# If clinical data isn't available, we cannot proceed with linking and final validation
+if 'selected_clinical_df' in globals() and selected_clinical_df is not None:
+    # 2. Link clinical and genetic data
+    raw_linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+    # 3. Handle missing values
+    linked_data = handle_missing_values(raw_linked_data, trait)
+
+    # 4. Determine bias and remove biased demographic features
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+    # 5. Final validation and save cohort info
+    # Cast to native Python bools to avoid serialization issues
+    is_gene_available_flag = bool((normalized_gene_data.shape[0] > 0) and (normalized_gene_data.shape[1] > 0))
+    is_trait_available_flag = bool((trait in raw_linked_data.columns) and raw_linked_data[trait].notna().any())
+    is_trait_biased = bool(is_trait_biased)
+
+    note = ("INFO: Trait derived from 'disease state' in cancer cell lines; distribution is extremely imbalanced "
+            "for Uterine_Carcinosarcoma (only a very small number of positives), likely unusable for association.")
+
+    try:
+        is_usable = validate_and_save_cohort_info(
+            is_final=True,
+            cohort=cohort,
+            info_path=json_path,
+            is_gene_available=is_gene_available_flag,
+            is_trait_available=is_trait_available_flag,
+            is_biased=is_trait_biased,
+            df=unbiased_linked_data,
+            note=note
+        )
+    except TypeError:
+        # Fallback: write a sanitized record converting booleans to ints
+        is_available = bool(is_gene_available_flag and is_trait_available_flag)
+        is_usable_calc = bool(is_available and (is_trait_biased is False))
+        sanitized_record = {
+            "is_usable": int(is_usable_calc),
+            "is_gene_available": int(is_gene_available_flag),
+            "is_trait_available": int(is_trait_available_flag),
+            "is_available": int(is_available),
+            "is_biased": int(is_trait_biased) if is_available else None,
+            "has_age": int('Age' in unbiased_linked_data.columns) if is_available else None,
+            "has_gender": int('Gender' in unbiased_linked_data.columns) if is_available else None,
+            "sample_size": int(len(unbiased_linked_data)) if is_available else None,
+            "note": note
+        }
+
+        trait_directory = os.path.dirname(json_path)
+        os.makedirs(trait_directory, exist_ok=True)
+        if not os.path.exists(json_path):
+            with open(json_path, 'w') as f:
+                json.dump({}, f)
+        with open(json_path, 'r') as f:
+            records = json.load(f)
+        records[cohort] = sanitized_record
+        temp_path = json_path + ".tmp"
+        with open(temp_path, 'w') as f:
+            json.dump(records, f)
+        os.replace(temp_path, json_path)
+        is_usable = is_usable_calc
+
+    # 6. Save linked data if usable
+    if is_usable:
+        os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+        unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Uterine_Carcinosarcoma/code/TCGA.py

@@ -0,0 +1,358 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Carcinosarcoma"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Carcinosarcoma/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Uterine_Carcinosarcoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import pandas as pd
+
+# 1) Select the most relevant TCGA cohort directory
+all_entries = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+priority_matches = [d for d in all_entries if 'uterine_carcinosarcoma' in d.lower()]
+if not priority_matches:
+    priority_matches = [d for d in all_entries if '(ucs)' in d.lower() or 'ucs' in d.lower()]
+
+selected_dir = priority_matches[0] if priority_matches else None
+
+if selected_dir is None:
+    # No suitable directory found -> mark as unavailable and skip
+    _ = validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+    clinical_df = None
+    genetic_df = None
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+    # 2) Identify file paths for clinical and genetic data
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    # 3) Load both files
+    clinical_df = pd.read_csv(clinical_file_path, sep="\t", index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_file_path, sep="\t", index_col=0, low_memory=False)
+
+    # 4) Print clinical column names
+    print(clinical_df.columns.tolist())
+
+# Step 2: Find Candidate Demographic Features
+import os
+import re
+import pandas as pd
+
+def find_cohort_dir(root):
+    dirs = [os.path.join(root, d) for d in os.listdir(root) if os.path.isdir(os.path.join(root, d))]
+    prioritized = [d for d in dirs if any(k in os.path.basename(d).lower() for k in ['ucs', 'uterine', 'carcinosarcoma'])]
+    return prioritized[0] if prioritized else (dirs[0] if dirs else root)
+
+def load_tcga_clinical_df(root_dir):
+    cohort_dir = find_cohort_dir(root_dir)
+    clinical_fp, _ = tcga_get_relevant_filepaths(cohort_dir)
+    try:
+        df = pd.read_csv(clinical_fp, sep="\t", index_col=0, dtype=str, low_memory=False)
+    except Exception:
+        df = pd.read_csv(clinical_fp, sep=",", index_col=0, dtype=str, low_memory=False)
+    return df
+
+clinical_df = load_tcga_clinical_df(tcga_root_dir)
+all_cols = list(clinical_df.columns)
+
+# Improved heuristics to avoid false positives like "usage" and "stage"
+age_token = re.compile(r'(^|[^a-z0-9])age([^a-z0-9]|$)')
+
+candidate_age_cols = []
+for c in all_cols:
+    cl = c.lower()
+    if cl == "days_to_birth":
+        candidate_age_cols.append(c)
+    elif age_token.search(cl):
+        candidate_age_cols.append(c)
+
+candidate_gender_cols = [c for c in all_cols if c.lower() in {"gender", "sex"}]
+
+def dedup(seq):
+    seen = set()
+    out = []
+    for x in seq:
+        if x not in seen:
+            seen.add(x)
+            out.append(x)
+    return out
+
+candidate_age_cols = dedup(candidate_age_cols)
+candidate_gender_cols = dedup(candidate_gender_cols)
+
+print(f"candidate_age_cols = {candidate_age_cols}")
+print(f"candidate_gender_cols = {candidate_gender_cols}")
+
+if candidate_age_cols:
+    age_preview = preview_df(clinical_df[candidate_age_cols], n=5)
+    print(age_preview)
+
+if candidate_gender_cols:
+    gender_preview = preview_df(clinical_df[candidate_gender_cols], n=5)
+    print(gender_preview)
+
+# Step 3: Select Demographic Features
+# Select the best age and gender columns from candidates using preview dictionaries only.
+# If the preview dictionary is missing or empty, set the corresponding selection to None.
+
+# Retrieve candidate lists
+candidate_age_cols = locals().get("candidate_age_cols", []) or []
+candidate_gender_cols = locals().get("candidate_gender_cols", []) or []
+
+# Retrieve preview dictionaries (expect standardized names from previous step)
+age_preview_dict = locals().get("age_values_dict", {}) or {}
+gender_preview_dict = locals().get("gender_values_dict", {}) or {}
+
+def _is_missing(x):
+    if x is None:
+        return True
+    s = str(x).strip()
+    return s == "" or s.lower() in {"na", "nan", "none", "null", "unknown", "not reported"}
+
+def _safe_list(x):
+    return list(x) if isinstance(x, (list, tuple)) else []
+
+def _age_value(x):
+    # Use library converter; returns an int if any digits are present
+    try:
+        v = tcga_convert_age(x)
+        return v if isinstance(v, int) else None
+    except Exception:
+        return None
+
+def _age_plausible(x):
+    v = _age_value(x)
+    return v is not None and 0 <= v <= 120
+
+def _gender_recognized(x):
+    try:
+        v = tcga_convert_gender(x)
+        return v in (0, 1)
+    except Exception:
+        return False
+
+# Select age column using preview dict only
+age_col = None
+if isinstance(age_preview_dict, dict) and len(age_preview_dict) > 0 and candidate_age_cols:
+    best_score = None
+    best_col = None
+    for col in candidate_age_cols:
+        samples = _safe_list(age_preview_dict.get(col, []))
+        if not samples:
+            continue
+        n = len(samples)
+        missing = sum(_is_missing(v) for v in samples)
+        plausible = sum(_age_plausible(v) for v in samples)
+        # Require at least 3 plausible values and <=60% missing
+        if plausible >= 3 and (missing / max(n, 1)) <= 0.6:
+            name_bonus = 2 if col == "age_at_initial_pathologic_diagnosis" else (1 if "age" in col.lower() else 0)
+            score = (plausible * 2) - missing + name_bonus
+            if best_score is None or score > best_score:
+                best_score = score
+                best_col = col
+    age_col = best_col  # May remain None if no suitable column
+
+# Select gender column using preview dict only
+gender_col = None
+if isinstance(gender_preview_dict, dict) and len(gender_preview_dict) > 0 and candidate_gender_cols:
+    best_score = None
+    best_col = None
+    for col in candidate_gender_cols:
+        samples = _safe_list(gender_preview_dict.get(col, []))
+        if not samples:
+            continue
+        n = len(samples)
+        missing = sum(_is_missing(v) for v in samples)
+        recognized = sum(_gender_recognized(v) for v in samples)
+        # Require at least 3 recognizable values and <=60% missing
+        if recognized >= 3 and (missing / max(n, 1)) <= 0.6:
+            name_bonus = 2 if col.lower() == "gender" else (1 if "sex" in col.lower() else 0)
+            score = (recognized * 2) - missing + name_bonus
+            if best_score is None or score > best_score:
+                best_score = score
+                best_col = col
+    gender_col = best_col  # May remain None if no suitable column
+
+# Explicitly print chosen columns
+print(f"Selected age_col: {age_col}")
+print(f"Selected gender_col: {gender_col}")
+
+# Step 4: Select Demographic Features
+import math
+
+# Initialize defaults
+age_col = None
+gender_col = None
+
+def is_missing(val):
+    if val is None:
+        return True
+    if isinstance(val, float) and math.isnan(val):
+        return True
+    if isinstance(val, str) and val.strip() == "":
+        return True
+    return False
+
+def iter_candidate_dicts(target='age'):
+    # Search globals for dicts that likely store candidate samples
+    # Preference: variable names containing the target keyword
+    for var_name, var_val in globals().items():
+        if not isinstance(var_val, dict):
+            continue
+        name_lower = var_name.lower()
+        if target in name_lower and any(k for k in var_val.keys()):
+            yield var_name, var_val
+    # Fallback: consider any dicts that look like {col_name: [samples...]}
+    for var_name, var_val in globals().items():
+        if not isinstance(var_val, dict):
+            continue
+        if any(isinstance(k, str) for k in var_val.keys()) and any(
+            isinstance(v, (list, tuple)) for v in var_val.values()
+        ):
+            yield var_name, var_val
+
+def choose_best_column(candidate_dicts, converter, min_valid=3):
+    best_col = None
+    best_score = (-1, -1)  # (valid_count, total_non_missing)
+    seen_cols = set()
+
+    for _, d in candidate_dicts:
+        for col, samples in d.items():
+            if col in seen_cols:
+                continue
+            seen_cols.add(col)
+            if not isinstance(samples, (list, tuple)) or len(samples) == 0:
+                continue
+
+            total = 0
+            valid = 0
+            for s in samples:
+                if is_missing(s):
+                    continue
+                total += 1
+                conv = converter(s)
+                if conv is not None:
+                    valid += 1
+
+            # Require minimum validity across the previewed items
+            if total == 0 or valid < min_valid:
+                continue
+
+            score = (valid, total)
+            if score > best_score:
+                best_score = score
+                best_col = col
+
+    return best_col
+
+# Collect candidate dictionaries and pick columns
+age_candidate_dicts = list(iter_candidate_dicts(target='age'))
+gender_candidate_dicts = list(iter_candidate_dicts(target='gender'))
+
+# Choose age column using tcga_convert_age
+age_col_selected = choose_best_column(age_candidate_dicts, tcga_convert_age, min_valid=2)
+# Choose gender column using tcga_convert_gender
+gender_col_selected = choose_best_column(gender_candidate_dicts, tcga_convert_gender, min_valid=2)
+
+# Assign to required variables, defaulting to None if not found
+age_col = age_col_selected if age_col_selected else None
+gender_col = gender_col_selected if gender_col_selected else None
+
+# Explicitly print out chosen columns
+print(f"Selected age_col: {age_col}")
+print(f"Selected gender_col: {gender_col}")
+
+# Step 5: Feature Engineering and Validation
+import os
+import pandas as pd
+
+# 1) Extract and standardize clinical features
+age_col = locals().get("age_col", None)
+gender_col = locals().get("gender_col", None)
+
+selected_clinical_df = tcga_select_clinical_features(
+    clinical_df=clinical_df,
+    trait=trait,
+    age_col=age_col,
+    gender_col=gender_col
+)
+
+# 2) Normalize gene symbols in the gene expression data
+def _normalize_gene_axis(df: pd.DataFrame) -> pd.DataFrame:
+    # Try as-is (assume genes are rows)
+    try:
+        ng1 = normalize_gene_symbols_in_index(df.copy())
+    except Exception:
+        ng1 = pd.DataFrame()
+
+    # If very few genes recognized, try transposed (assume genes are columns)
+    if ng1.shape[0] < 100:
+        try:
+            ng2 = normalize_gene_symbols_in_index(df.T.copy())
+        except Exception:
+            ng2 = pd.DataFrame()
+        return ng2 if ng2.shape[0] > ng1.shape[0] else ng1
+    else:
+        return ng1
+
+normalized_gene_df = _normalize_gene_axis(genetic_df)
+
+# Save normalized gene data
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_df.to_csv(out_gene_data_file)
+
+# 3) Link clinical and genetic data on sample IDs
+common_samples = selected_clinical_df.index.intersection(normalized_gene_df.columns)
+linked_data = pd.concat(
+    [selected_clinical_df.loc[common_samples], normalized_gene_df.loc[:, common_samples].T],
+    axis=1
+)
+
+# 4) Handle missing values
+processed_df = handle_missing_values(linked_data, trait_col=trait)
+
+# 5) Determine bias and remove biased demographic features
+is_biased, processed_df = judge_and_remove_biased_features(processed_df, trait=trait)
+is_biased = bool(is_biased)
+
+# 6) Final validation and save cohort info (ensure native Python bools for JSON)
+is_gene_available = bool((normalized_gene_df.shape[0] > 0) and (normalized_gene_df.shape[1] > 0))
+is_trait_available = bool((trait in linked_data.columns) and (int(linked_data[trait].notna().sum()) > 0))
+
+note = (
+    f"INFO: Selected age column '{age_col}', gender column '{gender_col}'. "
+    f"Linked {len(common_samples)} of {len(selected_clinical_df)} clinical samples with gene expression. "
+    f"Gene symbols normalized using NCBI synonyms; retained {normalized_gene_df.shape[0]} genes."
+)
+
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort="TCGA",
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available,
+    is_biased=is_biased,
+    df=processed_df,
+    note=note
+)
+
+# 7) Save linked data only if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    processed_df.to_csv(out_data_file)

output/preprocess/Uterine_Carcinosarcoma/cohort_info.json

@@ -1,52 +1 @@
-{
-  "GSE68950": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": true,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 798
-  },
-  "GSE36138": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE36133": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE32507": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 46
-  },
-  "TCGA": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": true,
-    "has_age": true,
-    "has_gender": false,
-    "sample_size": 57
-  }
-}
+{"GSE68950": {"is_usable": false, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": true, "has_age": false, "has_gender": false, "sample_size": 798, "note": "INFO: Trait derived from 'disease state' in cancer cell lines; distribution is extremely imbalanced for Uterine_Carcinosarcoma (only a very small number of positives), likely unusable for association."}, "GSE36138": {"is_usable": false, "is_gene_available": false, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE36133": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE32507": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 46, "note": "INFO: Only trait available; no age or gender annotations. Gene symbols normalized via NCBI synonym mapping."}, "TCGA": {"is_usable": false, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": true, "has_age": true, "has_gender": false, "sample_size": 57, "note": "INFO: Selected age column 'age_at_initial_pathologic_diagnosis', gender column 'gender'. Linked 57 of 57 clinical samples with gene expression. Gene symbols normalized using NCBI synonyms; retained 19848 genes."}}

output/preprocess/Uterine_Corpus_Endometrial_Carcinoma/clinical_data/GSE32507.csv

@@ -1,2 +1,2 @@
-Sample_ID,characteristics_ch1
-0.0,0.0
+,GSM804806,GSM804807,GSM804808,GSM804809,GSM804810,GSM804811,GSM804812,GSM804813,GSM804814,GSM804815,GSM804816,GSM804817,GSM804818,GSM804819,GSM804820,GSM804821,GSM804822,GSM804823,GSM804824,GSM804825,GSM804826,GSM804827,GSM804828,GSM804829,GSM804830,GSM804831,GSM804832,GSM804833,GSM804834,GSM804835,GSM804836,GSM804837,GSM804838,GSM804839,GSM804840,GSM804841,GSM804842,GSM804843,GSM804844,GSM804845,GSM804846,GSM804847,GSM804848,GSM804849,GSM804850,GSM804851
+Uterine_Corpus_Endometrial_Carcinoma,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0

output/preprocess/Uterine_Corpus_Endometrial_Carcinoma/code/GSE32507.py

@@ -0,0 +1,211 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Corpus_Endometrial_Carcinoma"
+cohort = "GSE32507"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Uterine_Corpus_Endometrial_Carcinoma"
+in_cohort_dir = "../DATA/GEO/Uterine_Corpus_Endometrial_Carcinoma/GSE32507"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/GSE32507.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/gene_data/GSE32507.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/clinical_data/GSE32507.csv"
+json_path = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+import pandas as pd
+
+# 1) Gene expression availability based on background: cDNA microarray => gene expression available
+is_gene_available = True
+
+# 2) Identify rows for variables from Sample Characteristics Dictionary
+# From the provided dictionary:
+# {0: ['tissue: endometrioid adenocarcinoma', 'tissue: carcinosarcoma', 'tissue: sarcoma'],
+#  1: ['carcinosarcoma status: : heterologous', 'carcinosarcoma status: : homologous']}
+trait_row = 0  # 'tissue' distinguishes EC vs CS/US
+age_row = None
+gender_row = None
+
+# 2.2 Conversion functions
+def _parse_after_colon(x):
+    if x is None:
+        return None
+    try:
+        # Handle cases with multiple colons, take the last segment
+        return str(x).split(':')[-1].strip()
+    except Exception:
+        return None
+
+def convert_trait(x):
+    v = _parse_after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    # Map endometrioid adenocarcinoma (EC) to 1; carcinosarcoma (CS) and sarcoma (US) to 0
+    if ('endometrioid' in v_low) and ('adenocarcinoma' in v_low or 'carcinoma' in v_low):
+        return 1
+    if ('carcinosarcoma' in v_low) or ('sarcoma' in v_low):
+        return 0
+    # Conservative fallback
+    return None
+
+def convert_age(x):
+    # Not used (no age available), but provided for completeness
+    v = _parse_after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    # Extract numeric value
+    m = re.search(r'(\d+(\.\d+)?)', v_low)
+    if not m:
+        return None
+    val = float(m.group(1))
+    # Heuristics for units if present
+    if 'month' in v_low:
+        return val / 12.0
+    # default to years
+    return val
+
+def convert_gender(x):
+    # Not used (no gender available), but provided for completeness
+    v = _parse_after_colon(x)
+    if v is None:
+        return None
+    v_low = v.lower()
+    if v_low in ['male', 'm', 'man']:
+        return 1
+    if v_low in ['female', 'f', 'woman', 'women']:
+        return 0
+    return None
+
+# 3) Save metadata using initial filtering
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (only if trait is available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected_clinical_df, n=5)
+    print(preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Identify the appropriate columns in the annotation for probe IDs and gene symbols
+probe_col = 'ID'
+symbol_col = 'GENE_SYMBOL'
+
+# Build mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, probe_col, symbol_col)
+
+# Apply mapping to convert probe-level data to gene-level data
+gene_data = apply_gene_mapping(gene_data, mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+import pandas as pd
+
+# 1. Normalize gene symbols and save gene data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# Ensure we have the clinical dataframe in memory; fallback to loading from disk if needed
+try:
+    selected_clinical_df
+except NameError:
+    selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)
+
+# 2. Link the clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Bias checking; remove biased demographic features if necessary
+is_trait_biased_raw, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+is_trait_biased = bool(is_trait_biased_raw)
+
+# 5. Final quality validation and save metadata
+gene_cols_in_linked = [c for c in unbiased_linked_data.columns if c not in [trait, 'Age', 'Gender']]
+is_gene_available_final = bool(len(gene_cols_in_linked) > 0)
+is_trait_available_final = bool((trait in unbiased_linked_data.columns) and (unbiased_linked_data[trait].notna().any()))
+
+note = ("INFO: Trait derived from 'tissue': endometrioid adenocarcinoma mapped to 1; "
+        "carcinosarcoma/sarcoma mapped to 0. No age/gender available in clinical. "
+        "Probes mapped via GENE_SYMBOL and normalized to NCBI standard symbols.")
+
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available_final,
+    is_trait_available=is_trait_available_final,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note
+)
+
+# 6. Save linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Uterine_Corpus_Endometrial_Carcinoma/code/TCGA.py

@@ -0,0 +1,250 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Uterine_Corpus_Endometrial_Carcinoma"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Uterine_Corpus_Endometrial_Carcinoma/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import pandas as pd
+
+# Step 1: Select the most relevant TCGA cohort directory for Uterine Corpus Endometrial Carcinoma (UCEC)
+subdirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+def score_dir(name: str) -> int:
+    lname = name.lower()
+    score = 0
+    if "(ucec)" in lname:
+        score += 100
+    if "ucec" in lname:
+        score += 50
+    if "endometrial" in lname or "endometrioid" in lname:
+        score += 30
+    if "uterine" in lname:
+        score += 10
+    if "carcinosarcoma" in lname or "(ucs)" in lname:
+        score -= 1000
+    return score
+
+scored = [(d, score_dir(d)) for d in subdirs]
+scored.sort(key=lambda x: x[1], reverse=True)
+selected_dir = scored[0][0] if scored and scored[0][1] > 0 else None
+
+if selected_dir is None:
+    # No suitable directory found; record and exit early for this trait
+    validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+
+    # Step 2: Identify clinicalMatrix and PANCAN files
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    # Step 3: Load both files as DataFrames
+    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False)
+
+    # Step 4: Print clinical column names
+    print(clinical_df.columns.tolist())
+
+# Step 2: Find Candidate Demographic Features
+# Identify candidate columns for age and gender from the provided column list
+column_list = ['CDE_ID_3226963', '_INTEGRATION', '_PANCAN_CNA_PANCAN_K8', '_PANCAN_Cluster_Cluster_PANCAN', '_PANCAN_DNAMethyl_PANCAN', '_PANCAN_DNAMethyl_UCEC', '_PANCAN_RPPA_PANCAN_K8', '_PANCAN_UNC_RNAseq_PANCAN_K16', '_PANCAN_miRNA_PANCAN', '_PANCAN_mirna_UCEC', '_PANCAN_mutation_PANCAN', '_PATIENT', '_cohort', '_primary_disease', '_primary_site', 'additional_pharmaceutical_therapy', 'additional_radiation_therapy', 'additional_surgery_locoregional_procedure', 'additional_treatment_completion_success_outcome', 'age_at_initial_pathologic_diagnosis', 'aln_pos_ihc', 'aln_pos_light_micro', 'bcr_followup_barcode', 'bcr_patient_barcode', 'bcr_sample_barcode', 'birth_control_pill_history_usage_category', 'clinical_stage', 'colorectal_cancer', 'days_to_additional_surgery_metastatic_procedure', 'days_to_birth', 'days_to_collection', 'days_to_death', 'days_to_initial_pathologic_diagnosis', 'days_to_last_followup', 'days_to_last_known_alive', 'days_to_new_tumor_event_additional_surgery_procedure', 'days_to_new_tumor_event_after_initial_treatment', 'diabetes', 'disease_code', 'followup_case_report_form_submission_reason', 'form_completion_date', 'gender', 'height', 'histological_type', 'history_of_neoadjuvant_treatment', 'horm_ther', 'hypertension', 'icd_10', 'icd_o_3_histology', 'icd_o_3_site', 'informed_consent_verified', 'init_pathology_dx_method_other', 'initial_pathologic_diagnosis_method', 'initial_weight', 'is_ffpe', 'lost_follow_up', 'menopause_status', 'neoplasm_histologic_grade', 'new_neoplasm_event_occurrence_anatomic_site', 'new_neoplasm_event_type', 'new_neoplasm_occurrence_anatomic_site_text', 'new_tumor_event_additional_surgery_procedure', 'new_tumor_event_after_initial_treatment', 'oct_embedded', 'other_dx', 'pathology_report_file_name', 'patient_id', 'pct_tumor_invasion', 'peritoneal_wash', 'person_neoplasm_cancer_status', 'pln_pos_ihc', 'pln_pos_light_micro', 'postoperative_rx_tx', 'pregnancies', 'primary_therapy_outcome_success', 'prior_tamoxifen_administered_usage_category', 'project_code', 'radiation_therapy', 'recurrence_second_surgery_neoplasm_surgical_procedure_name', 'recurrence_second_surgery_neoplasm_surgical_procedure_name_other', 'residual_disease_post_new_tumor_event_margin_status', 'residual_tumor', 'sample_type', 'sample_type_id', 'surgical_approach', 'system_version', 'targeted_molecular_therapy', 'tissue_prospective_collection_indicator', 'tissue_retrospective_collection_indicator', 'tissue_source_site', 'total_aor_lnp', 'total_aor_lnr', 'total_pelv_lnp', 'total_pelv_lnr', 'tumor_tissue_site', 'vial_number', 'vital_status', 'weight', 'year_of_initial_pathologic_diagnosis', '_GENOMIC_ID_TCGA_UCEC_exp_HiSeqV2_PANCAN', '_GENOMIC_ID_data/public/TCGA/UCEC/miRNA_GA_gene', '_GENOMIC_ID_TCGA_UCEC_PDMRNAseq', '_GENOMIC_ID_TCGA_UCEC_exp_HiSeqV2_percentile', '_GENOMIC_ID_TCGA_UCEC_RPPA_RBN', '_GENOMIC_ID_TCGA_UCEC_RPPA', '_GENOMIC_ID_TCGA_UCEC_PDMarrayCNV', '_GENOMIC_ID_TCGA_UCEC_miRNA_GA', '_GENOMIC_ID_TCGA_UCEC_exp_HiSeqV2_exon', '_GENOMIC_ID_TCGA_UCEC_mutation_broad_gene', '_GENOMIC_ID_TCGA_UCEC_mutation_wustl_gene', '_GENOMIC_ID_TCGA_UCEC_mutation', '_GENOMIC_ID_TCGA_UCEC_exp_HiSeqV2', '_GENOMIC_ID_TCGA_UCEC_PDMarray', '_GENOMIC_ID_TCGA_UCEC_miRNA_HiSeq', '_GENOMIC_ID_TCGA_UCEC_exp_GAV2', '_GENOMIC_ID_TCGA_UCEC_gistic2thd', '_GENOMIC_ID_TCGA_UCEC_G4502A_07_3', '_GENOMIC_ID_TCGA_UCEC_gistic2', '_GENOMIC_ID_data/public/TCGA/UCEC/miRNA_HiSeq_gene', '_GENOMIC_ID_TCGA_UCEC_PDMRNAseqCNV', '_GENOMIC_ID_TCGA_UCEC_hMethyl450', '_GENOMIC_ID_TCGA_UCEC_hMethyl27', '_GENOMIC_ID_TCGA_UCEC_exp_GAV2_exon']
+
+columns_set = set(column_list)
+
+candidate_age_cols = []
+if 'age_at_initial_pathologic_diagnosis' in columns_set:
+    candidate_age_cols.append('age_at_initial_pathologic_diagnosis')
+if 'days_to_birth' in columns_set:
+    candidate_age_cols.append('days_to_birth')
+
+candidate_gender_cols = []
+if 'gender' in columns_set:
+    candidate_gender_cols.append('gender')
+
+print(f"candidate_age_cols = {candidate_age_cols}")
+print(f"candidate_gender_cols = {candidate_gender_cols}")
+
+# Extract and preview candidate columns if clinical data is available or can be loaded
+selected_cols = list(dict.fromkeys(candidate_age_cols + candidate_gender_cols))
+
+if selected_cols:
+    import os
+    import glob
+    import pandas as pd
+
+    clinical_df_available = 'clinical_df' in globals() and isinstance(globals().get('clinical_df', None), pd.DataFrame)
+    if not clinical_df_available:
+        # Try to locate and load the clinical matrix for UCEC
+        dirs = [d for d in glob.glob(os.path.join(tcga_root_dir, '*')) if os.path.isdir(d)]
+        cohort_dir = None
+        for d in dirs:
+            if os.path.basename(d).upper() == 'UCEC':
+                cohort_dir = d
+                break
+        if cohort_dir is None:
+            for d in dirs:
+                base = os.path.basename(d)
+                if 'UCEC' in base.upper() or 'uterine' in base.lower():
+                    cohort_dir = d
+                    break
+        if cohort_dir is None:
+            cohort_dir = os.path.join(tcga_root_dir, 'UCEC')
+
+        try:
+            clinical_file_path, _ = tcga_get_relevant_filepaths(cohort_dir)
+            clinical_df = pd.read_table(clinical_file_path, sep='\t', header=0)
+        except Exception:
+            clinical_df = None
+
+    # Preview if data loaded successfully and columns exist
+    if isinstance(globals().get('clinical_df', None), pd.DataFrame):
+        existing_cols = [c for c in selected_cols if c in clinical_df.columns]
+        if existing_cols:
+            preview = preview_df(clinical_df[existing_cols], n=5)
+            print(preview)
+        else:
+            print({})
+    else:
+        print({})
+else:
+    print({})
+
+# Step 3: Select Demographic Features
+# Select best demographic columns from candidates based on name heuristics and available previews
+
+# Initialize defaults
+age_col = None
+gender_col = None
+
+# Heuristic scoring for age columns: prefer 'age', avoid 'days'/'birth'
+def score_age_col(col_name: str) -> int:
+    name = col_name.lower()
+    score = 0
+    if "age" in name:
+        score += 2
+    if "pathologic" in name or "diagnosis" in name:
+        score += 1
+    if "days" in name or "birth" in name:
+        score -= 2
+    return score
+
+# Heuristic selection for age
+if 'candidate_age_cols' in globals() and isinstance(candidate_age_cols, list) and len(candidate_age_cols) > 0:
+    scored = sorted(candidate_age_cols, key=lambda c: score_age_col(c), reverse=True)
+    # Ensure the top choice has a non-negative score; otherwise set to None
+    top = scored[0]
+    if score_age_col(top) >= 0:
+        age_col = top
+    else:
+        age_col = None
+else:
+    age_col = None
+
+# Heuristic selection for gender
+if 'candidate_gender_cols' in globals() and isinstance(candidate_gender_cols, list) and len(candidate_gender_cols) > 0:
+    # Prefer 'gender' or 'sex'
+    pref_order = sorted(candidate_gender_cols, key=lambda c: (("gender" in c.lower()) or ("sex" in c.lower())), reverse=True)
+    gender_col = pref_order[0]
+else:
+    gender_col = None
+
+# Helper to retrieve preview values if previously stored in any dict in globals
+def _get_preview_values(col_name):
+    if col_name is None:
+        return None
+    for var_name, obj in list(globals().items()):
+        if isinstance(obj, dict) and col_name in obj:
+            return obj.get(col_name, None)
+    return None
+
+# Print explicit information
+print(f"Selected age_col: {age_col}")
+age_preview_vals = _get_preview_values(age_col)
+if age_preview_vals is not None:
+    print(f"Preview values for {age_col}: {age_preview_vals}")
+
+print(f"Selected gender_col: {gender_col}")
+gender_preview_vals = _get_preview_values(gender_col)
+if gender_preview_vals is not None:
+    print(f"Preview values for {gender_col}: {gender_preview_vals}")
+
+# Step 4: Feature Engineering and Validation
+import os
+import pandas as pd
+
+# 1) Extract and standardize clinical features (trait, Age, Gender)
+selected_clinical_df = tcga_select_clinical_features(
+    clinical_df=clinical_df,
+    trait=trait,
+    age_col=age_col,
+    gender_col=gender_col
+)
+
+# 2) Normalize gene symbols and save normalized gene expression data
+normalized_gene_df = normalize_gene_symbols_in_index(genetic_df)  # genes x samples
+
+# Ensure output directory exists and save normalized gene data
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_df.to_csv(out_gene_data_file)
+
+# 3) Link clinical and genetic data on sample IDs
+# clinical: samples x features; genes: genes x samples -> transpose to samples x genes
+linked_data = selected_clinical_df.join(normalized_gene_df.T, how='inner')
+
+# 4) Handle missing values as specified
+linked_data = handle_missing_values(linked_data, trait)
+
+# 5) Determine bias; remove biased demographic features while keeping trait
+trait_biased, linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 6) Final quality validation and save cohort info
+covariate_cols = [trait, 'Age', 'Gender']
+gene_cols_count = len([c for c in linked_data.columns if c not in covariate_cols])
+is_gene_available = gene_cols_count > 0
+is_trait_available = (trait in linked_data.columns) and (linked_data[trait].notna().sum() > 0) and (len(linked_data) > 0)
+
+# Prepare an informational note
+label_counts = linked_data[trait].value_counts(dropna=False).to_dict() if trait in linked_data.columns else {}
+has_age = 'Age' in linked_data.columns
+has_gender = 'Gender' in linked_data.columns
+note_parts = [
+    f"Samples={len(linked_data)}",
+    f"Genes={gene_cols_count}",
+    f"Trait_counts={label_counts}",
+    f"HasAge={has_age}",
+    f"HasGender={has_gender}"
+]
+note = "INFO: " + "; ".join(note_parts)
+
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort="TCGA",
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available,
+    is_biased=trait_biased,
+    df=linked_data,
+    note=note
+)
+
+# 7) Save linked data only if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    linked_data.to_csv(out_data_file)

output/preprocess/Uterine_Corpus_Endometrial_Carcinoma/cohort_info.json

@@ -1,22 +1 @@
-{
-  "GSE32507": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 46
-  },
-  "TCGA": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": true,
-    "has_gender": false,
-    "sample_size": 201
-  }
-}
+{"GSE32507": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 46, "note": "INFO: Trait derived from 'tissue': endometrioid adenocarcinoma mapped to 1; carcinosarcoma/sarcoma mapped to 0. No age/gender available in clinical. Probes mapped via GENE_SYMBOL and normalized to NCBI standard symbols."}, "TCGA": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": true, "has_gender": false, "sample_size": 201, "note": "INFO: Samples=201; Genes=19848; Trait_counts={1: 177, 0: 24}; HasAge=True; HasGender=False"}}

output/preprocess/Vitamin_D_Levels/clinical_data/GSE129604.csv

@@ -1,3 +1,3 @@
-0
-1.0
-""
+,GSM3716832,GSM3716833,GSM3716834,GSM3716835,GSM3716836,GSM3716837,GSM3716838,GSM3716839,GSM3716840,GSM3716841,GSM3716842,GSM3716843,GSM3716844,GSM3716845,GSM3716846,GSM3716847,GSM3716848,GSM3716849,GSM3716850,GSM3716851,GSM3716852,GSM3716853,GSM3716854,GSM3716855,GSM3716856,GSM3716857,GSM3716858,GSM3716859,GSM3716860,GSM3716861,GSM3716862,GSM3716863,GSM3716864,GSM3716865,GSM3716866,GSM3716867,GSM3716868,GSM3716869,GSM3716870,GSM3716871,GSM3716872,GSM3716873,GSM3716874,GSM3716875,GSM3716876,GSM3716877,GSM3716878,GSM3716879,GSM3716880,GSM3716881,GSM3716882,GSM3716883,GSM3716884,GSM3716885,GSM3716886,GSM3716887,GSM3716888,GSM3716889,GSM3716890,GSM3716891,GSM3716892,GSM3716893,GSM3716894,GSM3716895,GSM3716896,GSM3716897,GSM3716898,GSM3716899,GSM3716900,GSM3716901,GSM3716902,GSM3716903,GSM3716904,GSM3716905,GSM3716906,GSM3716907,GSM3716908,GSM3716909,GSM3716910,GSM3716911,GSM3716912,GSM3716913,GSM3716914,GSM3716915,GSM3716916,GSM3716917,GSM3716918,GSM3716919,GSM3716920,GSM3716921,GSM3716922,GSM3716923,GSM3716924,GSM3716925,GSM3716926,GSM3716927,GSM3716928,GSM3716929,GSM3716930,GSM3716931,GSM3716932,GSM3716933,GSM3716934,GSM3716935,GSM3716936,GSM3716937,GSM3716938,GSM3716939,GSM3716940,GSM3716941,GSM3716942,GSM3716943,GSM3716944,GSM3716945,GSM3716946,GSM3716947,GSM3716948,GSM3716949,GSM3716950,GSM3716951,GSM3716952,GSM3716953,GSM3716954,GSM3716955,GSM3716956,GSM3716957,GSM3716958,GSM3716959,GSM3716960,GSM3716961,GSM3716962,GSM3716963,GSM3716964,GSM3716965,GSM3716966,GSM3716967,GSM3716968,GSM3716969,GSM3716970,GSM3716971,GSM3716972
+Vitamin_D_Levels,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,0.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0,0.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,1.0,1.0
+Gender,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0

output/preprocess/Vitamin_D_Levels/clinical_data/GSE34450.csv

@@ -1,2 +1,2 @@
-0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56,57,58,59,60,61,62,63,64,65,66,67,68,69,70,71,72,73,74,75,76,77,78,79,80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99,100,101,102,103,104,105,106,107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,124,125,126,127,128,129,130,131
-,1.0,,,,,,,,,,1.0,1.0,1.0,,1.0,,,1.0,,,0.0,1.0,0.0,,,1.0,,,,1.0,,1.0,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
+,GSM549681,GSM549682,GSM549683,GSM549684,GSM549685,GSM549689,GSM549690,GSM549691,GSM549692,GSM549693,GSM549694,GSM549695,GSM549698,GSM549703,GSM549705,GSM549707,GSM549713,GSM549715,GSM549716,GSM549717,GSM549718,GSM549719,GSM549720,GSM549721,GSM549722,GSM549723,GSM549724,GSM549725,GSM549726,GSM549727,GSM549728,GSM549729,GSM549730,GSM549731,GSM549732,GSM549733,GSM549734,GSM549735,GSM549736,GSM549737,GSM549738,GSM549739,GSM549740,GSM549744,GSM549745,GSM549746,GSM549747,GSM549758,GSM549764,GSM549771,GSM549773,GSM549775,GSM549778,GSM549779,GSM549780,GSM549781,GSM549782,GSM549783,GSM549784,GSM549785,GSM549786,GSM549787,GSM549788,GSM549789,GSM549790,GSM549791,GSM549792,GSM549793,GSM549794,GSM549795,GSM549796,GSM549797,GSM549798,GSM549799,GSM549800,GSM549801,GSM549802,GSM549803,GSM549804,GSM549805,GSM549806,GSM549807,GSM549808,GSM549809,GSM549810,GSM549811,GSM549812,GSM549813,GSM599910,GSM599911,GSM599912,GSM599913,GSM599915,GSM599916,GSM599917,GSM599918,GSM599919,GSM599920,GSM599921,GSM631346,GSM631351,GSM631353,GSM631354,GSM631356,GSM631357,GSM631359,GSM631364,GSM631365,GSM631366,GSM631367,GSM631369,GSM631370,GSM631371,GSM631372,GSM631373,GSM631374,GSM631375,GSM631376,GSM631377,GSM631378,GSM631379,GSM631380,GSM631381,GSM631382,GSM631383,GSM631384,GSM631385,GSM631386,GSM631387,GSM631389,GSM631390,GSM631391
+Vitamin_D_Levels,,,,,,,,,,,,,1.0,,,1.0,,,1.0,,,0.0,1.0,0.0,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

output/preprocess/Vitamin_D_Levels/clinical_data/GSE76324.csv

@@ -1,2 +1,2 @@
-ID_REF,GSM549681,GSM549682,GSM549683,GSM549684,GSM549685,GSM549686,GSM549687,GSM549688,GSM549689,GSM549690,GSM549691,GSM549692,GSM549693,GSM549694,GSM549695,GSM549696,GSM549697,GSM549698,GSM549699,GSM549700,GSM549701,GSM549702,GSM549703,GSM549704,GSM549705,GSM549706,GSM549707,GSM549708,GSM549709,GSM549710,GSM549711,GSM549712,GSM549713,GSM549714,GSM549715,GSM549716,GSM549717,GSM549718,GSM549719,GSM549720,GSM549721,GSM549722,GSM549723,GSM549724,GSM549725,GSM549726,GSM549727,GSM549728,GSM549729,GSM549730,GSM549731,GSM549732,GSM549733,GSM549734,GSM549735,GSM549736,GSM549737,GSM549738,GSM549739,GSM549740,GSM549741,GSM549742,GSM549743,GSM549744,GSM549745,GSM549746,GSM549747,GSM549748,GSM549749,GSM549750,GSM549751,GSM549752,GSM549753,GSM549754,GSM549755,GSM549756,GSM549757,GSM549758,GSM549759,GSM549760,GSM549761,GSM549762,GSM549763,GSM549764,GSM549765,GSM549766,GSM549767,GSM549768,GSM549769,GSM549770,GSM549771,GSM549772,GSM549773,GSM549774,GSM549775,GSM549776,GSM549777,GSM549778,GSM549779,GSM549780,GSM549781,GSM549782,GSM549783,GSM549784,GSM549785,GSM549786,GSM549787,GSM549788,GSM549789,GSM549790,GSM549791,GSM549792,GSM549793,GSM549794,GSM549795,GSM549796,GSM549797,GSM549798,GSM549799,GSM549800,GSM549801,GSM549802,GSM549803,GSM549804,GSM549805,GSM549806,GSM549807,GSM549809,GSM549810,GSM549811,GSM549813,GSM631340,GSM631341,GSM631342,GSM631343,GSM631344,GSM631345,GSM631346,GSM631347,GSM631348,GSM631349,GSM631350,GSM631351,GSM631352,GSM631353,GSM631354,GSM631355,GSM631356,GSM631357,GSM631358,GSM631359,GSM631360,GSM631361,GSM631362,GSM631363,GSM631364,GSM631365,GSM631366,GSM631367,GSM631368,GSM631369,GSM631370,GSM631371,GSM631372,GSM631373,GSM631374,GSM631375,GSM631376,GSM631377,GSM631378,GSM631379,GSM631380,GSM631381,GSM631382,GSM631383,GSM631384,GSM631385,GSM631386,GSM631387,GSM631388,GSM631389,GSM631390,GSM631391,GSM1056118,GSM1056119,GSM1056120,GSM1056121,GSM1056123,GSM1056124,GSM1056125,GSM1056126,GSM1056127,GSM1056128,GSM1056129,GSM1056130,GSM1056131,GSM1056132,GSM1056133,GSM1056134,GSM1056135,GSM1056136,GSM1056138,GSM1056139,GSM1056140,GSM1056141,GSM1056142,GSM1056143,GSM1056144,GSM1056145,GSM1056147,GSM1056148,GSM1056149,GSM1056150,GSM1056153,GSM1575535,GSM1575536,GSM1575537
-,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
+,GSM549681,GSM549682,GSM549683,GSM549684,GSM549685,GSM549686,GSM549687,GSM549688,GSM549689,GSM549690,GSM549691,GSM549692,GSM549693,GSM549694,GSM549695,GSM549696,GSM549697,GSM549698,GSM549699,GSM549700,GSM549701,GSM549702,GSM549703,GSM549704,GSM549705,GSM549706,GSM549707,GSM549708,GSM549709,GSM549710,GSM549711,GSM549712,GSM549713,GSM549714,GSM549715,GSM549716,GSM549717,GSM549718,GSM549719,GSM549720,GSM549721,GSM549722,GSM549723,GSM549724,GSM549725,GSM549726,GSM549727,GSM549728,GSM549729,GSM549730,GSM549731,GSM549732,GSM549733,GSM549734,GSM549735,GSM549736,GSM549737,GSM549738,GSM549739,GSM549740,GSM549741,GSM549742,GSM549743,GSM549744,GSM549745,GSM549746,GSM549747,GSM549748,GSM549749,GSM549750,GSM549751,GSM549752,GSM549753,GSM549754,GSM549755,GSM549756,GSM549757,GSM549758,GSM549759,GSM549760,GSM549761,GSM549762,GSM549763,GSM549764,GSM549765,GSM549766,GSM549767,GSM549768,GSM549769,GSM549770,GSM549771,GSM549772,GSM549773,GSM549774,GSM549775,GSM549776,GSM549777,GSM549778,GSM549779,GSM549780,GSM549781,GSM549782,GSM549783,GSM549784,GSM549785,GSM549786,GSM549787,GSM549788,GSM549789,GSM549790,GSM549791,GSM549792,GSM549793,GSM549794,GSM549795,GSM549796,GSM549797,GSM549798,GSM549799,GSM549800,GSM549801,GSM549802,GSM549803,GSM549804,GSM549805,GSM549806,GSM549807,GSM549809,GSM549810,GSM549811,GSM549813,GSM631340,GSM631341,GSM631342,GSM631343,GSM631344,GSM631345,GSM631346,GSM631347,GSM631348,GSM631349,GSM631350,GSM631351,GSM631352,GSM631353,GSM631354,GSM631355,GSM631356,GSM631357,GSM631358,GSM631359,GSM631360,GSM631361,GSM631362,GSM631363,GSM631364,GSM631365,GSM631366,GSM631367,GSM631368,GSM631369,GSM631370,GSM631371,GSM631372,GSM631373,GSM631374,GSM631375,GSM631376,GSM631377,GSM631378,GSM631379,GSM631380,GSM631381,GSM631382,GSM631383,GSM631384,GSM631385,GSM631386,GSM631387,GSM631388,GSM631389,GSM631390,GSM631391,GSM1056118,GSM1056119,GSM1056120,GSM1056121,GSM1056123,GSM1056124,GSM1056125,GSM1056126,GSM1056127,GSM1056128,GSM1056129,GSM1056130,GSM1056131,GSM1056132,GSM1056133,GSM1056134,GSM1056135,GSM1056136,GSM1056138,GSM1056139,GSM1056140,GSM1056141,GSM1056142,GSM1056143,GSM1056144,GSM1056145,GSM1056147,GSM1056148,GSM1056149,GSM1056150,GSM1056153,GSM1575535,GSM1575536,GSM1575537
+Vitamin_D_Levels,,0.5,,,,,,,,,,,,,0.5,1.0,,1.0,,0.0,1.0,,0.5,,,,1.0,0.0,0.5,,0.0,,,1.0,,1.0,,,0.0,1.0,0.0,,,0.5,,,,0.5,,0.5,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,

output/preprocess/Vitamin_D_Levels/code/GSE118723.py

@@ -0,0 +1,208 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE118723"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE118723"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE118723.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE118723.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE118723.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+
+# 1) Gene expression availability (scRNA-seq of iPSCs => gene expression available)
+is_gene_available = True
+
+# 2) Variable availability (based on provided Sample Characteristics Dictionary)
+# No Vitamin D, age, or gender information present
+trait_row = None
+age_row = None
+gender_row = None
+
+# 2.2) Conversion functions
+
+def _extract_value_after_colon(x: str) -> str:
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        return parts[1].strip() if len(parts) > 1 else x.strip()
+    return x
+
+def convert_trait(x):
+    """
+    Convert Vitamin D levels to continuous (float, e.g., ng/mL or nmol/L if present).
+    Extracts the first numeric value found after the colon; returns None if not parseable.
+    """
+    try:
+        s = _extract_value_after_colon(x)
+        if s is None:
+            return None
+        match = re.search(r"[-+]?\d*\.?\d+(?:[eE][-+]?\d+)?", s)
+        return float(match.group()) if match else None
+    except Exception:
+        return None
+
+def convert_age(x):
+    """
+    Convert age to continuous (float years).
+    Extract the first numeric value after the colon; returns None if not parseable.
+    """
+    try:
+        s = _extract_value_after_colon(x)
+        if s is None:
+            return None
+        match = re.search(r"[-+]?\d*\.?\d+(?:[eE][-+]?\d+)?", s)
+        return float(match.group()) if match else None
+    except Exception:
+        return None
+
+def convert_gender(x):
+    """
+    Convert gender to binary: female -> 0, male -> 1.
+    Supports common synonyms. Returns None if unknown.
+    Uses exact token checks first, then word-boundary regex to avoid substring issues (e.g., 'female' containing 'male').
+    """
+    s = _extract_value_after_colon(x)
+    if s is None:
+        return None
+    v = str(s).strip().lower()
+
+    # Exact tokens first
+    if v in {"female", "f", "woman", "girl"}:
+        return 0
+    if v in {"male", "m", "man", "boy"}:
+        return 1
+
+    # Word-boundary regex fallback
+    if re.search(r"\bfemale\b", v):
+        return 0
+    if re.search(r"\bmale\b", v):
+        return 1
+
+    return None
+
+# 3) Save metadata with initial filtering
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction: skipped because trait_row is None (no clinical trait data available)
+# If trait_row becomes available in future steps, you can uncomment and use the following:
+# selected_clinical_df = geo_select_clinical_features(
+#     clinical_df=clinical_data,
+#     trait=trait,
+#     trait_row=trait_row,
+#     convert_trait=convert_trait,
+#     age_row=age_row,
+#     convert_age=convert_age,
+#     gender_row=gender_row,
+#     convert_gender=convert_gender
+# )
+# preview = preview_df(selected_clinical_df, n=5)
+# os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+# selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+import os
+import gzip
+import pandas as pd
+
+# 1) Extract gene expression data with robust fallback
+print(f"Selected matrix_file: {matrix_file}")
+
+# Primary attempt using library
+try:
+    gene_data = get_genetic_data(matrix_file)
+except Exception as e:
+    print(f"get_genetic_data failed with error: {e}")
+    gene_data = pd.DataFrame()
+
+# Fallback if empty
+if gene_data.shape[0] == 0:
+    skip_rows = None
+    try:
+        with gzip.open(matrix_file, 'rt') as fh:
+            for i, line in enumerate(fh):
+                if "!series_matrix_table_begin" in line:
+                    skip_rows = i + 1
+                    break
+    except Exception as e:
+        print(f"Error locating marker: {e}")
+
+    if skip_rows is not None:
+        try:
+            df_try = pd.read_csv(
+                matrix_file,
+                compression='gzip',
+                sep='\t',
+                skiprows=skip_rows,
+                engine='python',
+                dtype=str,
+                comment='!',
+                on_bad_lines='skip',
+                header=0
+            )
+            # Clean column names (strip quotes/whitespace)
+            df_try.columns = [str(c).strip().strip('"').strip("'") for c in df_try.columns]
+
+            # Normalize first column as ID
+            first_col = df_try.columns[0]
+            if first_col in ["ID_REF", "ID"]:
+                df_try = df_try.rename(columns={first_col: "ID"})
+            else:
+                # Force first column to ID if header is unconventional
+                df_try = df_try.rename(columns={first_col: "ID"})
+
+            # Clean ID column and set index
+            df_try["ID"] = df_try["ID"].astype(str).str.strip().str.strip('"').str.strip("'")
+            df_try = df_try[df_try["ID"].notna() & (df_try["ID"] != "")]
+            df_try = df_try.set_index("ID")
+
+            # Also strip quotes/whitespace in values
+            df_try = df_try.apply(lambda col: col.astype(str).str.strip().str.strip('"').str.strip("'"))
+
+            gene_data = df_try
+        except Exception as e:
+            print(f"Fallback read failed with error: {e}")
+            gene_data = pd.DataFrame()
+
+print("Parsed gene_data shape:", getattr(gene_data, "shape", None))
+
+# 2) Print the first 20 row IDs (gene or probe identifiers)
+first_ids = list(gene_data.index[:20])
+print("First 20 row IDs:")
+print(first_ids)

output/preprocess/Vitamin_D_Levels/code/GSE123993.py

@@ -0,0 +1,215 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE123993"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE123993"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE123993.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE123993.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE123993.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+
+# 1) Gene expression availability
+is_gene_available = True  # Affymetrix HuGene 2.1ST arrays -> whole-genome expression
+
+# 2) Variable availability
+# From the sample characteristics:
+# 0: tissue
+# 1: Sex
+# 2: subject id
+# 3: intervention group
+# 4: time of sampling
+trait_row = None          # No direct Vitamin D level per sample; cannot be mapped to a single key
+age_row = None            # No per-sample age provided
+gender_row = 1            # 'Sex: Male/Female'
+
+# 2.2) Converters and data type choices
+trait_type = 'continuous'
+age_type = 'continuous'
+gender_type = 'binary'
+
+def _extract_value(cell):
+    if cell is None:
+        return None
+    if isinstance(cell, str):
+        parts = cell.split(':', 1)
+        return parts[1].strip() if len(parts) > 1 else cell.strip()
+    return cell
+
+def convert_trait(x):
+    # Expect numeric vitamin D concentration if available; here it's not used since trait_row is None.
+    v = _extract_value(x)
+    if v is None:
+        return None
+    # Extract first float-like number
+    m = re.search(r'[-+]?\d*\.?\d+', str(v))
+    if m:
+        try:
+            return float(m.group())
+        except Exception:
+            return None
+    return None
+
+def convert_age(x):
+    v = _extract_value(x)
+    if v is None:
+        return None
+    # Extract age as the first integer/float found
+    m = re.search(r'[-+]?\d*\.?\d+', str(v))
+    if m:
+        try:
+            return float(m.group())
+        except Exception:
+            return None
+    return None
+
+def convert_gender(x):
+    v = _extract_value(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower()
+    if s in ['male', 'm']:
+        return 1
+    if s in ['female', 'f']:
+        return 0
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip if trait is not available)
+if trait_row is not None:
+    selected = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected)
+    print(preview)
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+print("requires_gene_mapping = True")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Determine columns for mapping based on annotation preview:
+# - Probe identifier matches 'ID' (numeric probeset IDs like '1665xxxx')
+# - Gene symbols are embedded in 'gene_assignment'
+prob_col = 'ID'
+gene_col = 'gene_assignment'
+
+# Build the mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, prob_col=prob_col, gene_col=gene_col)
+
+# Apply mapping: convert probe-level data to gene-level data
+probe_data = gene_data  # preserve original probe-level data
+gene_data = apply_gene_mapping(expression_df=probe_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1) Normalize gene symbols and save gene-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2) Link clinical and genetic data only if clinical features (including trait) were extracted
+is_usable = False
+if 'selected_clinical_data' in locals() and selected_clinical_data is not None and trait in selected_clinical_data.index:
+    linked_data = geo_link_clinical_genetic_data(selected_clinical_data, normalized_gene_data)
+
+    # 3) Handle missing values
+    linked_data = handle_missing_values(linked_data, trait)
+
+    # 4) Bias checks and removal of biased covariates
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+    # 5) Final validation and save cohort info
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=True,
+        is_biased=is_trait_biased,
+        df=unbiased_linked_data,
+        note="INFO: Linked data created from gene-level normalized expression and available clinical features."
+    )
+
+    # 6) Save linked data only if usable
+    if is_usable:
+        os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+        unbiased_linked_data.to_csv(out_data_file)
+
+else:
+    # Trait not available -> cannot link; still record final metadata without marking gene as unavailable
+    # Use gene data (transposed) as df for validation to avoid false flagging gene availability
+    df_for_validation = normalized_gene_data.T
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=False,
+        is_biased=False,
+        df=df_for_validation,
+        note="WARNING: Trait not available; cannot link clinical and genetic data for this cohort."
+    )
+    # Do not save out_data_file when trait is not available

output/preprocess/Vitamin_D_Levels/code/GSE129604.py

@@ -0,0 +1,204 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE129604"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE129604"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE129604.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE129604.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE129604.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+
+# 1. Gene Expression Data Availability
+is_gene_available = True  # Whole blood whole-genome expression per background
+
+# 2. Variable Availability and Data Type Conversion
+
+# Keys from Sample Characteristics Dictionary:
+# 0: Sex
+# 1: tissue
+# 2: agent (Placebo, BPH, VitD+BPH, VitD, VitD-Pep-C-VitDPep)
+# 3: time point
+
+# Trait: Infer Vitamin D exposure from 'agent' (contains VitD -> 1, else -> 0)
+trait_row = 2
+
+# Age: Not available in the sample characteristics
+age_row = None
+
+# Gender: Available under 'Sex'
+gender_row = 0
+
+def _extract_value(x: str) -> str:
+    if x is None:
+        return ""
+    parts = str(x).split(":", 1)
+    val = parts[1] if len(parts) > 1 else parts[0]
+    return val.strip()
+
+def convert_trait(x):
+    val = _extract_value(x).lower()
+    if val == "":
+        return None
+    # Map presence of vitamin D in the agent to 1, otherwise 0
+    if "vitd" in val:
+        return 1
+    if "placebo" in val or "bph" in val:
+        return 0
+    return None
+
+def convert_age(x):
+    # Generic parser: extract numeric value if present, else None
+    val = _extract_value(x)
+    if val == "":
+        return None
+    m = re.search(r"[-+]?\d*\.?\d+", val)
+    return float(m.group()) if m else None
+
+def convert_gender(x):
+    val = _extract_value(x).lower()
+    if val in ["male", "m"]:
+        return 1
+    if val in ["female", "f"]:
+        return 0
+    return None
+
+# 3. Save Metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4. Clinical Feature Extraction (only if trait available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=None if age_row is None else convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df)
+    # Save clinical data
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+    print(clinical_preview)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+# The observed identifiers (e.g., 'AFFX-BkGr-GC03_st') are Affymetrix control/probeset IDs, not human gene symbols.
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Decide columns for probe IDs and gene symbols in the annotation
+id_col = 'probeset_id' if 'probeset_id' in gene_annotation.columns else 'ID'
+if 'SPOT_ID.1' in gene_annotation.columns:
+    gene_col = 'SPOT_ID.1'
+elif 'SPOT_ID' in gene_annotation.columns:
+    gene_col = 'SPOT_ID'
+else:
+    # Fallback: if no obvious gene column, use the most descriptive text column available
+    # This is a safeguard; in this dataset SPOT_ID.1 exists per preview.
+    text_cols = [c for c in gene_annotation.columns if gene_annotation[c].dtype == object]
+    gene_col = text_cols[-1] if text_cols else id_col  # worst case fallback
+
+# 2. Build mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, prob_col=id_col, gene_col=gene_col)
+
+# 3. Apply mapping to convert probe-level to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save gene expression data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Assess bias and remove biased demographic features
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# Compute availability flags for final validation
+is_gene_available_final = (normalized_gene_data.shape[0] > 0) and (normalized_gene_data.shape[1] > 0)
+is_trait_available_final = (trait in selected_clinical_df.index) and (not selected_clinical_df.loc[trait].isna().all())
+
+note = (
+    "INFO: Trait derived from 'agent' field (presence of 'VitD' => 1, otherwise 0). "
+    "Age not provided; Gender parsed from 'Sex'. Randomized crossover design with two time points retained."
+)
+
+# 5. Final validation and save cohort metadata
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available_final,
+    is_trait_available=is_trait_available_final,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note
+)
+
+# 6. Save linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Vitamin_D_Levels/code/GSE33544.py

@@ -0,0 +1,129 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE33544"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE33544"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE33544.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE33544.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE33544.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import re
+
+# 1) Gene expression data availability (based on series description: expression of light chain V genes)
+is_gene_available = True
+
+# 2) Variable availability (from provided Sample Characteristics Dictionary)
+# Trait is Vitamin_D_Levels, which is not present in this dataset
+trait_row = None
+age_row = None
+gender_row = None
+
+# 2.2) Converters (defined for completeness; they won't be used since rows are None)
+def _after_colon(x):
+    if x is None:
+        return None
+    s = str(x)
+    parts = s.split(":", 1)
+    return parts[1].strip() if len(parts) == 2 else s.strip()
+
+def convert_trait(x):
+    # Expecting continuous vitamin D levels (e.g., "vitamin d: 25 ng/ml")
+    v = _after_colon(x)
+    if not v:
+        return None
+    # Extract first float-like token
+    m = re.search(r'[-+]?\d*\.?\d+(?:[eE][-+]?\d+)?', v)
+    if m:
+        try:
+            return float(m.group(0))
+        except:
+            return None
+    # Map qualitative descriptors if ever encountered
+    low_map = {"deficient": 0.0, "insufficient": 0.0}
+    high_map = {"sufficient": 1.0, "normal": 1.0}
+    vl = v.lower()
+    if vl in low_map:
+        return low_map[vl]
+    if vl in high_map:
+        return high_map[vl]
+    return None
+
+def convert_age(x):
+    # Continuous age in years
+    v = _after_colon(x)
+    if not v:
+        return None
+    m = re.search(r'[-+]?\d*\.?\d+(?:[eE][-+]?\d+)?', v)
+    if m:
+        try:
+            return float(m.group(0))
+        except:
+            return None
+    return None
+
+def convert_gender(x):
+    # Binary: female -> 0, male -> 1
+    v = _after_colon(x)
+    if not v:
+        return None
+    vl = v.strip().lower()
+    if vl in {"male", "m", "man", "boy"}:
+        return 1
+    if vl in {"female", "f", "woman", "girl"}:
+        return 0
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip because trait_row is None)
+# If trait data were available, we would run:
+# selected_clinical = geo_select_clinical_features(
+#     clinical_df=clinical_data,
+#     trait=trait,
+#     trait_row=trait_row,
+#     convert_trait=convert_trait,
+#     age_row=age_row,
+#     convert_age=convert_age,
+#     gender_row=gender_row,
+#     convert_gender=convert_gender
+# )
+# preview = preview_df(selected_clinical)
+# os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+# selected_clinical.to_csv(out_clinical_data_file)

output/preprocess/Vitamin_D_Levels/code/GSE34450.py

@@ -0,0 +1,184 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE34450"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE34450"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE34450.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE34450.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE34450.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import pandas as pd
+
+# 1) Gene expression data availability (based on background: microarray gene expression study)
+is_gene_available = True
+
+# 2) Variable availability and conversion functions
+
+# Trait (Vitamin D Levels) identified under "serum 25-oh-d" with categories in row 3
+trait_row = 3
+
+# Age and Gender not available in the provided characteristics
+age_row = None
+gender_row = None
+
+def convert_trait(x):
+    if pd.isna(x):
+        return None
+    s = str(x)
+    parts = s.split(':', 1)
+    key = parts[0].strip().lower()
+    val = parts[1].strip().lower() if len(parts) == 2 else ''
+    # Only process if the key indicates serum 25-OH-D (vitamin D)
+    if 'serum 25-oh-d' in key:
+        if 'low' in val:
+            return 0
+        if 'high' in val:
+            return 1
+        if 'mid' in val or 'medium' in val:
+            return None
+        return None
+    return None
+
+def convert_age(x):
+    # Not available for this cohort
+    return None
+
+def convert_gender(x):
+    # Not available for this cohort
+    return None
+
+# 3) Save initial metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (only if trait_row is available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df, n=5)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file, index=True)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Identify the annotation columns for probe IDs and gene symbols
+probe_col = 'ID'
+gene_symbol_col = 'Gene Symbol'
+
+# Build mapping dataframe from annotation
+mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_col, gene_col=gene_symbol_col)
+
+# Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save gene data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Assess bias and remove biased demographic features (guard against empty data after missing value handling)
+if len(linked_data) > 0:
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+else:
+    is_trait_biased = True
+    unbiased_linked_data = linked_data  # keep structure for validation
+
+# 5. Final validation and metadata saving
+# Cast to built-in bool to avoid numpy.bool_ serialization issues
+is_gene_available_final = bool((normalized_gene_data.shape[0] > 0) and (normalized_gene_data.shape[1] > 0))
+is_trait_available_final = bool((trait in unbiased_linked_data.columns) and bool(unbiased_linked_data[trait].notna().any()))
+is_trait_biased = bool(is_trait_biased)
+
+note = (
+    "INFO: Probe-level data mapped to gene-level and normalized using NCBI synonyms. "
+    "Samples with 'mid' vitamin D were treated as missing for the trait and removed during filtering. "
+    "Note: Series note indicates many samples lack processed data; effective sample size may be reduced."
+)
+
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available_final,
+    is_trait_available=is_trait_available_final,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note
+)
+
+# 6. Save linked dataset only if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Vitamin_D_Levels/code/GSE35925.py

@@ -0,0 +1,215 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE35925"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE35925"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE35925.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE35925.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE35925.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+import pandas as pd
+
+# 1) Gene expression availability
+# Affymetrix U133 Plus 2.0 microarray -> mRNA gene expression data
+is_gene_available = True
+
+# 2) Variable availability based on Sample Characteristics Dictionary
+# Provided dictionary:
+# 0: ['gender: female'] -> constant; exclude
+# 1: ['age: 66', 'age: 63', ...] -> available
+# No explicit Vitamin D levels in characteristics -> not available
+trait_row = None
+age_row = 1
+gender_row = None  # constant 'female' only
+
+# 2.2) Converters
+def _extract_after_colon(value: str) -> str:
+    if value is None:
+        return ""
+    parts = str(value).split(":", 1)
+    return parts[1].strip() if len(parts) == 2 else str(value).strip()
+
+def _extract_number(s: str):
+    if s is None:
+        return None
+    m = re.search(r'[-+]?\d*\.?\d+', str(s))
+    if not m:
+        return None
+    try:
+        num = float(m.group())
+        # Return int when it's an integer value
+        return int(num) if num.is_integer() else num
+    except Exception:
+        return None
+
+# Trait: Vitamin D levels (expected continuous, e.g., ng/mL), but not available here.
+def convert_trait(x):
+    val = _extract_after_colon(x)
+    # Extract numeric value if present (e.g., "28 ng/mL" -> 28)
+    return _extract_number(val)
+
+# Age: continuous
+def convert_age(x):
+    val = _extract_after_colon(x)
+    return _extract_number(val)
+
+# Gender: binary female=0, male=1
+def convert_gender(x):
+    val = _extract_after_colon(x).strip().lower()
+    if val in {"f", "female", "woman", "women"}:
+        return 0
+    if val in {"m", "male", "man", "men"}:
+        return 1
+    return None
+
+# 3) Initial filtering and save metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip if trait not available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected_clinical_df)
+    print(preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# 1-2) Identify annotation columns and build mapping dataframe
+id_col = 'ID'  # Probe identifier column matches gene_data index like '1007_s_at'
+gene_symbol_col = 'Gene Symbol'  # Column containing gene symbols
+
+gene_mapping = get_gene_mapping(gene_annotation, prob_col=id_col, gene_col=gene_symbol_col)
+
+# 3) Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=gene_mapping)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1) Normalize gene symbols and save gene-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# Determine if trait is available from previous steps
+has_trait = ('trait_row' in globals()) and (trait_row is not None)
+
+if has_trait:
+    # Recompute clinical features to ensure availability in this step
+    selected_clinical_data = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row if 'age_row' in globals() else None,
+        convert_age=convert_age if 'convert_age' in globals() else None,
+        gender_row=gender_row if 'gender_row' in globals() else None,
+        convert_gender=convert_gender if 'convert_gender' in globals() else None
+    )
+
+    # 2) Link clinical and genetic data
+    linked_data = geo_link_clinical_genetic_data(selected_clinical_data, normalized_gene_data)
+
+    # 3) Handle missing values
+    linked_data = handle_missing_values(linked_data, trait)
+
+    # 4) Bias checks
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+    # 5) Final validation and save cohort info
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=True,
+        is_biased=is_trait_biased,
+        df=unbiased_linked_data,
+        note="INFO: Linked clinical and gene data with trait available."
+    )
+
+    # 6) Save usable linked data
+    if is_usable:
+        os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+        unbiased_linked_data.to_csv(out_data_file)
+
+else:
+    # Trait not available: skip linking and downstream processing, but record metadata
+    note = f"INFO: Trait '{trait}' not available in clinical annotations; skipping linking. Only gene expression saved."
+    _ = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=True,
+        is_trait_available=False,
+        is_biased=False,
+        df=normalized_gene_data.T,  # Use sample-oriented view for consistency
+        note=note
+    )

output/preprocess/Vitamin_D_Levels/code/GSE76324.py

@@ -0,0 +1,202 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE76324"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE76324"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE76324.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE76324.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE76324.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+
+# 1) Gene expression availability
+is_gene_available = True  # Microarray transcriptome study (mRNA), not miRNA/methylation
+
+# 2) Variable availability and conversion
+
+# From the provided Sample Characteristics Dictionary:
+# Choose row 3 for trait because it contains only serum 25-oh-d categories (row 2 mixes smoking status and vitamin D).
+trait_row = 3
+
+# No explicit age or gender fields in the provided keys
+age_row = None
+gender_row = None
+
+def _extract_value_after_colon(x):
+    if x is None:
+        return None
+    try:
+        s = str(x)
+    except Exception:
+        return None
+    parts = s.split(":", 1)
+    val = parts[1] if len(parts) > 1 else parts[0]
+    return val.strip()
+
+def convert_trait(x):
+    # Map categorical vitamin D levels to a continuous scale: low=0.0, mid=0.5, high=1.0
+    val = _extract_value_after_colon(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if "low" in v:
+        return 0.0
+    if "mid" in v or "intermediate" in v or "medium" in v:
+        return 0.5
+    if "high" in v:
+        return 1.0
+    try:
+        num = float(v)
+        return num
+    except Exception:
+        return None
+
+def convert_age(x):
+    # Not used (age_row is None); implemented for completeness
+    val = _extract_value_after_colon(x)
+    if val is None:
+        return None
+    v = val.lower()
+    import re
+    nums = re.findall(r"[-+]?\d*\.\d+|\d+", v)
+    if not nums:
+        return None
+    try:
+        age = float(nums[0])
+        if 0 <= age <= 120:
+            return age
+        return None
+    except Exception:
+        return None
+
+def convert_gender(x):
+    # Not used (gender_row is None); implemented for completeness
+    val = _extract_value_after_colon(x)
+    if val is None:
+        return None
+    v = val.strip().lower()
+    if v in {"female", "f", "woman", "women"}:
+        return 0
+    if v in {"male", "m", "man", "men"}:
+        return 1
+    if "female" in v:
+        return 0
+    if "male" in v:
+        return 1
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical Feature Extraction (only if trait_row is available)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df)
+    print(clinical_preview)
+    # Save clinical data
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Determine appropriate columns for mapping: probe ID ('ID') and gene symbol ('Gene Symbol')
+mapping_df = get_gene_mapping(gene_annotation, prob_col='ID', gene_col='Gene Symbol')
+
+# Apply the mapping to convert probe-level expression to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+import pandas as pd
+
+# 1. Normalize and save gene data
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link the clinical and genetic data
+if 'selected_clinical_df' not in globals():
+    selected_clinical_df = pd.read_csv(out_clinical_data_file, index_col=0)
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Bias assessment
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Final validation and save metadata
+is_gene_available = (normalized_gene_data.shape[0] > 0) and (normalized_gene_data.shape[1] > 0)
+is_trait_available = trait in selected_clinical_df.index
+note = "INFO: Trait derived from categorical serum 25-OH-D levels; Age/Gender not available in sample characteristics."
+is_usable = validate_and_save_cohort_info(
+    True, cohort, json_path, is_gene_available, is_trait_available, is_trait_biased, unbiased_linked_data, note
+)
+
+# 6. Save linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Vitamin_D_Levels/code/GSE86406.py

@@ -0,0 +1,249 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+cohort = "GSE86406"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Vitamin_D_Levels"
+in_cohort_dir = "../DATA/GEO/Vitamin_D_Levels/GSE86406"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/GSE86406.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/GSE86406.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/GSE86406.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+
+# 1) Determine data availability
+is_gene_available = True  # Microarray gene expression data described in the background
+trait_row = None          # No sample-level Vitamin D level data available in characteristics
+age_row = 1               # 'age: ...'
+gender_row = 2            # 'gender: ...'
+
+# 2) Converters
+def _after_colon(x: str) -> str:
+    if x is None:
+        return ""
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        return parts[1].strip() if len(parts) == 2 else x.strip()
+    return str(x)
+
+def convert_trait(x):
+    # Not used since trait_row is None; provide a generic numeric extractor for robustness.
+    val = _after_colon(x).lower()
+    # Extract first numeric value if present (e.g., "35 ng/mL" -> 35.0)
+    m = re.search(r'-?\d+\.?\d*', val)
+    if m:
+        try:
+            return float(m.group())
+        except Exception:
+            return None
+    return None
+
+def convert_age(x):
+    val = _after_colon(x)
+    m = re.search(r'-?\d+\.?\d*', val)
+    if m:
+        try:
+            return float(m.group())
+        except Exception:
+            return None
+    return None
+
+def convert_gender(x):
+    val = _after_colon(x).strip().lower()
+    if val in {"m", "male", "man", "men"}:
+        return 1
+    if val in {"f", "female", "woman", "women"}:
+        return 0
+    return None
+
+# 3) Save metadata (initial filtering)
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (skip because trait_row is None)
+if trait_row is not None:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    preview = preview_df(selected_clinical_df)
+    print(preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+# The observed identifiers like '16650001', '16650003', etc., are numeric feature IDs (e.g., Agilent Feature Numbers),
+# not human gene symbols. They require mapping to official gene symbols via platform annotations.
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+import os
+import json
+import pandas as pd
+
+# Try to use the platform (GPL) SOFT file for richer annotation; fall back to the earlier soft_file.
+files = os.listdir(in_cohort_dir)
+gpl_softs = [f for f in files if ('soft' in f.lower()) and ('gpl' in f.lower())]
+soft_for_annotation = os.path.join(in_cohort_dir, gpl_softs[0]) if gpl_softs else soft_file
+
+# Load annotation from the chosen SOFT
+gene_annotation = get_gene_annotation(soft_for_annotation)
+
+# 1) Decide probe ID column
+probe_col_candidates = [
+    'ID', 'ID_REF', 'PROBE_ID', 'ProbeName', 'PROBEID', 'Composite Element REF', 'CompositeElementID', 'SPOT_ID'
+]
+probe_col = next((c for c in probe_col_candidates if c in gene_annotation.columns), None)
+if probe_col is None:
+    raise ValueError(f"Probe/ID column not found in gene annotation. Available columns: {list(gene_annotation.columns)}")
+
+# 2) Decide gene symbol column by testing contents with extract_human_gene_symbols
+symbol_col_candidates_ordered = [
+    # highly preferred symbol columns
+    'Gene Symbol', 'GENE_SYMBOL', 'GENE_SYMBOLS', 'Symbol', 'SYMBOL', 'Gene symbol', 'GENE SYMBOL',
+    'GeneSymbols', 'GENE_SYMBOL (primary)',
+    # common fallbacks that often include symbols in mixed text
+    'gene_assignment', 'GENE_ASSIGNMENT', 'Gene Assignment',
+    'GENE_NAME', 'Gene Name', 'GENE TITLE', 'GENE_TITLE', 'DESCRIPTION', 'Description'
+]
+existing_symbol_cols = [c for c in symbol_col_candidates_ordered if c in gene_annotation.columns]
+
+def column_has_symbol_like(col_name: str, df, sample_n: int = 2000, min_fraction: float = 0.05) -> bool:
+    s = df[col_name].dropna().astype(str)
+    if s.empty:
+        return False
+    s = s.sample(min(len(s), sample_n), random_state=0) if len(s) > sample_n else s
+    hits = sum(1 for x in s if len(extract_human_gene_symbols(x)) > 0)
+    return (hits / len(s)) >= min_fraction
+
+gene_col = None
+for c in existing_symbol_cols:
+    if column_has_symbol_like(c, gene_annotation):
+        gene_col = c
+        break
+
+mapping_performed = False
+note = None
+
+def try_apply_mapping_with_columns(annotation_df: pd.DataFrame, id_col: str, symbol_col: str) -> pd.DataFrame:
+    mapping_df_local = get_gene_mapping(annotation_df, prob_col=id_col, gene_col=symbol_col)
+    if mapping_df_local.shape[0] == 0:
+        return pd.DataFrame()
+    mapped_gene_df_local = apply_gene_mapping(gene_data, mapping_df_local)
+    return mapped_gene_df_local
+
+mapped_gene_df = pd.DataFrame()
+
+# First attempt: use a true symbol-like column if found
+if gene_col is not None:
+    mapped_gene_df = try_apply_mapping_with_columns(gene_annotation, probe_col, gene_col)
+
+# Fallback: attempt RefSeq accession -> symbol mapping if first attempt failed
+def load_refseq_map() -> dict:
+    candidates = [
+        "./metadata/refseq_to_symbol.json",
+        "./metadata/refseq_to_symbol.tsv",
+        "./metadata/refseq2symbol.tsv",
+        "./metadata/refseq_symbol_map.tsv"
+    ]
+    for p in candidates:
+        if os.path.exists(p):
+            if p.endswith(".json"):
+                with open(p, "r") as f:
+                    d = json.load(f)
+                return d
+            else:
+                try:
+                    df_map = pd.read_csv(p, sep=None, engine='python', dtype=str)
+                except Exception:
+                    df_map = pd.read_csv(p, sep='\t', dtype=str)
+                df_map.columns = [c.strip().lower() for c in df_map.columns]
+                # Heuristic: find refseq and symbol columns
+                ref_col = next((c for c in df_map.columns if 'refseq' in c or c in {'gb_acc', 'accession'}), None)
+                sym_col = next((c for c in df_map.columns if 'symbol' in c), None)
+                if ref_col and sym_col:
+                    return dict(zip(df_map[ref_col].astype(str), df_map[sym_col].astype(str)))
+    return {}
+
+if (mapped_gene_df.shape[0] == 0 or mapped_gene_df.shape[1] == 0):
+    if 'GB_ACC' in gene_annotation.columns:
+        refseq_map = load_refseq_map()
+        if refseq_map:
+            temp = gene_annotation[[probe_col, 'GB_ACC']].dropna().copy()
+            temp['Gene'] = temp['GB_ACC'].map(refseq_map)
+            temp = temp.dropna(subset=['Gene'])
+            if temp.shape[0] > 0:
+                # Build mapping_df in the expected format
+                mapping_df = temp.rename(columns={probe_col: 'ID'})[['ID', 'Gene']]
+                mapped_gene_df = apply_gene_mapping(gene_data, mapping_df)
+
+# Finalize: replace gene_data only if mapping produced a valid matrix
+if mapped_gene_df is not None and mapped_gene_df.shape[0] > 0 and mapped_gene_df.shape[1] > 0:
+    gene_data = mapped_gene_df
+    mapping_performed = True
+    print(f"Mapping succeeded using probe_col='{probe_col}' and "
+          f"gene_col='{gene_col if gene_col is not None else 'GB_ACC->Symbol (fallback)'}'. "
+          f"Gene-level matrix shape: {gene_data.shape}")
+else:
+    note = ("WARNING: Gene symbol mapping could not be completed. "
+            "Keeping probe-level data for downstream steps.")
+    print(note)
+
+print(f"mapping_performed = {mapping_performed}")

output/preprocess/Vitamin_D_Levels/code/TCGA.py

@@ -0,0 +1,60 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Vitamin_D_Levels"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Vitamin_D_Levels/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Vitamin_D_Levels/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Vitamin_D_Levels/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Vitamin_D_Levels/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+
+# Step 1: Select the most relevant TCGA cohort directory for Vitamin D related phenotypes
+subdirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+# Prioritized keywords related to Vitamin D
+keywords = [
+    "vitamin d", "25-hydroxyvitamin d", "25(oh)d", "calcidiol", "calcitriol",
+    "cholecalciferol", "vit d", "vitamin_d", "vit_d"
+]
+
+selected_dir = None
+for kw in keywords:
+    for d in subdirs:
+        if kw in d.lower():
+            selected_dir = d
+            break
+    if selected_dir:
+        break
+
+if not selected_dir:
+    # No suitable cohort found; mark as unavailable and skip
+    validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+    clinical_df = None
+    genetic_df = None
+    print(f"No suitable TCGA cohort found for trait: {trait}. Skipping.")
+else:
+    # Step 2: Identify clinical and genetic file paths
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    # Step 3: Load both files
+    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False)
+
+    # Step 4: Print clinical column names
+    print(clinical_df.columns.tolist())

output/preprocess/Vitamin_D_Levels/cohort_info.json

@@ -1,72 +1 @@
-{
-  "GSE86406": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE35925": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE34450": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 132
-  },
-  "GSE33544": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "GSE123993": {
-    "is_usable": true,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": false,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 44
-  },
-  "GSE118723": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  },
-  "TCGA": {
-    "is_usable": false,
-    "is_gene_available": false,
-    "is_trait_available": false,
-    "is_available": false,
-    "is_biased": null,
-    "has_age": null,
-    "has_gender": null,
-    "sample_size": null
-  }
-}
+{"GSE86406": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE76324": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 19, "note": "INFO: Trait derived from categorical serum 25-OH-D levels; Age/Gender not available in sample characteristics."}, "GSE35925": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "INFO: Trait 'Vitamin_D_Levels' not available in clinical annotations; skipping linking. Only gene expression saved."}, "GSE34450": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": false, "sample_size": 6, "note": "INFO: Probe-level data mapped to gene-level and normalized using NCBI synonyms. Samples with 'mid' vitamin D were treated as missing for the trait and removed during filtering. Note: Series note indicates many samples lack processed data; effective sample size may be reduced."}, "GSE33544": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "GSE129604": {"is_usable": true, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": false, "has_age": false, "has_gender": true, "sample_size": 141, "note": "INFO: Trait derived from 'agent' field (presence of 'VitD' => 1, otherwise 0). Age not provided; Gender parsed from 'Sex'. Randomized crossover design with two time points retained."}, "GSE123993": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "WARNING: Trait not available; cannot link clinical and genetic data for this cohort."}, "GSE118723": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}, "TCGA": {"is_usable": false, "is_gene_available": false, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": null}}

output/preprocess/Von_Hippel_Lindau/code/GSE33093.py

@@ -0,0 +1,213 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Von_Hippel_Lindau"
+cohort = "GSE33093"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Von_Hippel_Lindau"
+in_cohort_dir = "../DATA/GEO/Von_Hippel_Lindau/GSE33093"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/GSE33093.csv"
+out_gene_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/gene_data/GSE33093.csv"
+out_clinical_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/clinical_data/GSE33093.csv"
+json_path = "./output/z6/preprocess/Von_Hippel_Lindau/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+# Determine data availability based on provided background information and characteristics
+is_gene_available = True  # Gene expression is indicated in the series title/summary; not miRNA/methylation-specific.
+
+# No explicit VHL status, age, or gender in the provided sample characteristics dictionary.
+trait_row = None
+age_row = None
+gender_row = None
+
+# Converters
+def _after_colon(x: str) -> str:
+    if x is None:
+        return None
+    if isinstance(x, str):
+        parts = x.split(":", 1)
+        val = parts[1] if len(parts) > 1 else parts[0]
+        return val.strip() if isinstance(val, str) else val
+    return x
+
+def convert_trait(x):
+    # Binary: 1 = VHL mutated/inactivated/loss; 0 = VHL wild type; None = unknown
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower()
+    if s in {"vhl wt", "vhl wild type", "wild type", "wt", "wild-type"}:
+        return 0
+    if any(tok in s for tok in ["mut", "mutation", "mutant", "inactivat", "loss", "deleted", "del", "loh"]):
+        return 1
+    if "vhl" in s:
+        # Heuristic: if mentions VHL without explicit WT keywords, treat as unknown
+        return None
+    return None
+
+def convert_age(x):
+    # Continuous age in years
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower().replace("years", "").replace("year", "").replace("yrs", "").replace("yr", "")
+    s = s.replace("yo", "").replace("y/o", "").strip()
+    try:
+        return float(s)
+    except:
+        # Extract first number if present
+        import re
+        m = re.search(r"(\d+(\.\d+)?)", s)
+        return float(m.group(1)) if m else None
+
+def convert_gender(x):
+    # Binary: female -> 0, male -> 1
+    v = _after_colon(x)
+    if v is None:
+        return None
+    s = str(v).strip().lower()
+    if s in {"male", "m"}:
+        return 1
+    if s in {"female", "f"}:
+        return 0
+    return None
+
+# Initial filtering and save cohort metadata
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# Clinical feature extraction is skipped because trait_row is None (no clinical trait data available).
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# Decide the appropriate columns for mapping based on previous previews
+probe_id_col = 'ID'           # Matches the probe identifiers in gene_data (e.g., '1', '2', ...)
+gene_symbol_col = 'GENE_SYMBOL'  # Column containing gene symbols
+
+# 2. Build the mapping dataframe
+mapping_df = get_gene_mapping(gene_annotation, prob_col=probe_id_col, gene_col=gene_symbol_col)
+
+# 3. Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save gene-level data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# Determine availability flags based on earlier steps
+is_gene_available = True
+try:
+    is_trait_available = (trait_row is not None)
+except NameError:
+    is_trait_available = False
+
+if is_trait_available:
+    # Ensure clinical features are extracted if not already present
+    if 'selected_clinical_data' not in globals() and 'selected_clinical_data' not in locals():
+        selected_clinical_data = geo_select_clinical_features(
+            clinical_df=clinical_data,
+            trait=trait,
+            trait_row=trait_row,
+            convert_trait=convert_trait,
+            age_row=age_row,
+            convert_age=convert_age,
+            gender_row=gender_row,
+            convert_gender=convert_gender
+        )
+
+    # 2. Link clinical and genetic data
+    linked_data = geo_link_clinical_genetic_data(selected_clinical_data, normalized_gene_data)
+
+    # 3. Handle missing values
+    linked_data = handle_missing_values(linked_data, trait)
+
+    # 4. Bias assessment (trait required to validate usability)
+    is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+    # 5. Final validation and cohort info saving
+    note = "INFO: Linked data generated with gene symbol normalization; clinical trait available."
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=is_gene_available,
+        is_trait_available=is_trait_available,
+        is_biased=is_trait_biased,
+        df=unbiased_linked_data,
+        note=note
+    )
+
+    # 6. Save linked data only if usable
+    if is_usable:
+        os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+        unbiased_linked_data.to_csv(out_data_file)
+
+else:
+    # No clinical trait available: skip linking and downstream steps
+    note = "INFO: Trait data not available in this cohort; only gene data were processed and saved."
+    # Provide a non-empty df to pass validation; it won't be considered available due to missing trait
+    placeholder_df = normalized_gene_data.T  # samples x genes
+    _ = validate_and_save_cohort_info(
+        is_final=True,
+        cohort=cohort,
+        info_path=json_path,
+        is_gene_available=is_gene_available,
+        is_trait_available=False,
+        is_biased=False,  # ignored because is_available will be False
+        df=placeholder_df,
+        note=note
+    )
+    # Do NOT save out_data_file since dataset is not usable without trait

output/preprocess/Von_Hippel_Lindau/code/TCGA.py

@@ -0,0 +1,272 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Von_Hippel_Lindau"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Von_Hippel_Lindau/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Von_Hippel_Lindau/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import pandas as pd
+
+# List available TCGA cohort subdirectories
+subdirs = [d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))]
+
+# Select the most relevant cohort for Von Hippel-Lindau (prioritize PCPG, then KIRC)
+preferred_order = [
+    'TCGA_Pheochromocytoma_Paraganglioma_(PCPG)',
+    'TCGA_Kidney_Clear_Cell_Carcinoma_(KIRC)'
+]
+selected_subdir = None
+for name in preferred_order:
+    if name in subdirs:
+        selected_subdir = name
+        break
+
+# Fallback: try to find by keyword if exact names not present
+if selected_subdir is None:
+    for d in subdirs:
+        d_low = d.lower()
+        if ('pheochromocytoma' in d_low) or ('paraganglioma' in d_low) or ('kirc' in d_low) or ('kidney_clear_cell' in d_low):
+            selected_subdir = d
+            break
+
+# If no suitable cohort found, record and exit
+if selected_subdir is None:
+    validate_and_save_cohort_info(
+        is_final=False,
+        cohort='TCGA',
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_subdir)
+    clinical_path, genetic_path = tcga_get_relevant_filepaths(cohort_dir)
+
+    # Load dataframes
+    clinical_df = pd.read_csv(clinical_path, sep='\t', index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_path, sep='\t', index_col=0, low_memory=False)
+
+    # Print clinical column names
+    print(list(clinical_df.columns))
+
+# Step 2: Find Candidate Demographic Features
+# Identify candidate demographic feature columns from the provided list
+provided_columns = ['_INTEGRATION', '_PATIENT', '_cohort', '_primary_disease', '_primary_site', 'age_at_initial_pathologic_diagnosis', 'bcr_followup_barcode', 'bcr_patient_barcode', 'bcr_sample_barcode', 'ct_scan', 'days_to_birth', 'days_to_collection', 'days_to_death', 'days_to_initial_pathologic_diagnosis', 'days_to_last_followup', 'days_to_new_tumor_event_after_initial_treatment', 'disease_detected_on_screening', 'eastern_cancer_oncology_group', 'form_completion_date', 'gender', 'histological_type', 'history_of_neoadjuvant_treatment', 'history_pheo_or_para_anatomic_site', 'history_pheo_or_para_include_benign', 'icd_10', 'icd_o_3_histology', 'icd_o_3_site', 'informed_consent_verified', 'initial_weight', 'is_ffpe', 'karnofsky_performance_score', 'laterality', 'lost_follow_up', 'lymph_node_examined_count', 'new_neoplasm_confirmed_diagnosis_method_name', 'new_neoplasm_event_occurrence_anatomic_site', 'new_neoplasm_event_type', 'new_neoplasm_occurrence_anatomic_site_text', 'new_tumor_event_after_initial_treatment', 'number_of_lymphnodes_positive_by_he', 'oct_embedded', 'other_dx', 'outside_adrenal', 'pathology_report_file_name', 'patient_id', 'performance_status_scale_timing', 'person_neoplasm_cancer_status', 'postoperative_rx_tx', 'primary_lymph_node_presentation_assessment', 'primary_therapy_outcome_success', 'radiation_therapy', 'sample_type', 'sample_type_id', 'tissue_prospective_collection_indicator', 'tissue_retrospective_collection_indicator', 'tissue_source_site', 'tumor_tissue_site', 'tumor_tissue_site_other', 'vial_number', 'vital_status', 'year_of_initial_pathologic_diagnosis', '_GENOMIC_ID_TCGA_PCPG_exp_HiSeqV2_PANCAN', '_GENOMIC_ID_TCGA_PCPG_mutation_bcm_gene', '_GENOMIC_ID_TCGA_PCPG_mutation_broad_gene', '_GENOMIC_ID_TCGA_PCPG_hMethyl450', '_GENOMIC_ID_TCGA_PCPG_gistic2thd', '_GENOMIC_ID_TCGA_PCPG_exp_HiSeqV2', '_GENOMIC_ID_TCGA_PCPG_exp_HiSeqV2_exon', '_GENOMIC_ID_TCGA_PCPG_PDMRNAseqCNV', '_GENOMIC_ID_TCGA_PCPG_miRNA_HiSeq', '_GENOMIC_ID_data/public/TCGA/PCPG/miRNA_HiSeq_gene', '_GENOMIC_ID_TCGA_PCPG_mutation_bcgsc_gene', '_GENOMIC_ID_TCGA_PCPG_RPPA', '_GENOMIC_ID_TCGA_PCPG_mutation_ucsc_maf_gene', '_GENOMIC_ID_TCGA_PCPG_gistic2', '_GENOMIC_ID_TCGA_PCPG_PDMRNAseq', '_GENOMIC_ID_TCGA_PCPG_exp_HiSeqV2_percentile']
+
+candidate_age_cols = [c for c in provided_columns if ('age' in c.lower()) or ('birth' in c.lower())]
+candidate_gender_cols = [c for c in provided_columns if ('gender' in c.lower()) or ('sex' in c.lower())]
+
+print(f"candidate_age_cols = {candidate_age_cols}")
+print(f"candidate_gender_cols = {candidate_gender_cols}")
+
+# Try to locate a clinical dataframe from previous steps
+import pandas as pd
+
+df_candidates = []
+for var_name in ['clinical_df', 'clin_df', 'clinical_data', 'tcga_clinical_df', 'df_clinical']:
+    if var_name in globals() and isinstance(globals()[var_name], pd.DataFrame):
+        df_candidates.append(globals()[var_name])
+
+if not df_candidates:
+    # Fallback: search any DataFrame in globals that has overlap with candidate columns
+    for name, obj in list(globals().items()):
+        if isinstance(obj, pd.DataFrame):
+            if any(col in obj.columns for col in (candidate_age_cols + candidate_gender_cols)):
+                df_candidates.append(obj)
+
+clinical_df_to_use = df_candidates[0] if df_candidates else None
+
+# Extract and preview candidate columns if clinical data is available
+age_preview = {}
+gender_preview = {}
+if clinical_df_to_use is not None:
+    selected_age_cols = [c for c in candidate_age_cols if c in clinical_df_to_use.columns]
+    selected_gender_cols = [c for c in candidate_gender_cols if c in clinical_df_to_use.columns]
+    if selected_age_cols:
+        age_preview = preview_df(clinical_df_to_use[selected_age_cols], n=5)
+    if selected_gender_cols:
+        gender_preview = preview_df(clinical_df_to_use[selected_gender_cols], n=5)
+
+print(f"age_preview = {age_preview}")
+print(f"gender_preview = {gender_preview}")
+
+# Step 3: Select Demographic Features
+# Heuristic selection of age and gender columns from previews
+
+def is_number(x):
+    try:
+        float(x)
+        return True
+    except Exception:
+        return False
+
+def select_age_column(candidate_cols, preview_dict):
+    if not preview_dict or not candidate_cols:
+        return None
+
+    best_col = None
+    best_score = -1
+
+    for col in candidate_cols:
+        vals = preview_dict.get(col, [])
+        if not vals:
+            continue
+
+        # Clean numeric values
+        nums = [float(v) for v in vals if is_number(v)]
+
+        if not nums:
+            continue
+
+        # Compute scores
+        # age-like: values between 0 and 120
+        age_like = sum(0 <= v <= 120 for v in nums)
+
+        # days_to_birth-like: negative values with plausible range up to 120 years
+        days_like = sum((v < 0) and (abs(v) <= 120 * 365.25) for v in nums)
+
+        # Prefer explicit age columns when available
+        name_bonus = 1 if 'age' in col.lower() else 0
+
+        # Final scoring logic:
+        # - If column name indicates age, emphasize age_like
+        # - Otherwise, consider both patterns
+        score = age_like * 2 + days_like + name_bonus
+
+        if score > best_score:
+            best_score = score
+            best_col = col
+
+    return best_col
+
+def select_gender_column(candidate_cols, preview_dict):
+    if not preview_dict or not candidate_cols:
+        return None
+
+    best_col = None
+    best_score = -1
+
+    valid_tokens = {'male', 'female', 'm', 'f'}
+    for col in candidate_cols:
+        vals = preview_dict.get(col, [])
+        if not vals:
+            continue
+        cleaned = [str(v).strip().lower() for v in vals if v is not None]
+        recognized = sum(v in valid_tokens for v in cleaned)
+        name_bonus = 1 if any(tok in col.lower() for tok in ['gender', 'sex']) else 0
+        score = recognized * 2 + name_bonus
+
+        if score > best_score:
+            best_score = score
+            best_col = col
+
+    return best_col
+
+# Use previews from previous step
+age_col = select_age_column(candidate_age_cols, age_preview) if 'age_preview' in globals() else None
+gender_col = select_gender_column(candidate_gender_cols, gender_preview) if 'gender_preview' in globals() else None
+
+# Fallbacks: ensure selected columns have meaningful preview values; else set to None
+if age_col is not None:
+    vals = age_preview.get(age_col, [])
+    if not vals or all(v is None for v in vals):
+        age_col = None
+
+if gender_col is not None:
+    vals = gender_preview.get(gender_col, [])
+    if not vals or all(v is None for v in vals):
+        gender_col = None
+
+# Explicitly print chosen columns and their preview values
+print("Chosen age_col:", age_col)
+print("Age preview:", age_preview.get(age_col) if age_col else None)
+print("Chosen gender_col:", gender_col)
+print("Gender preview:", gender_preview.get(gender_col) if gender_col else None)
+
+# Step 4: Feature Engineering and Validation
+import os
+import pandas as pd
+
+# 1) Extract and standardize clinical features (trait, Age, Gender)
+age_col_use = age_col if ('age_col' in globals() and isinstance(age_col, str) and age_col in clinical_df.columns) else None
+gender_col_use = gender_col if ('gender_col' in globals() and isinstance(gender_col, str) and gender_col in clinical_df.columns) else None
+
+selected_clinical_df = tcga_select_clinical_features(
+    clinical_df=clinical_df,
+    trait=trait,
+    age_col=age_col_use,
+    gender_col=gender_col_use
+)
+
+# 2) Normalize gene symbols and save normalized gene data
+gene_df_norm = normalize_gene_symbols_in_index(genetic_df.copy())
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+gene_df_norm.to_csv(out_gene_data_file)
+
+# 3) Link clinical and genetic data on sample IDs
+linked_data = selected_clinical_df.join(gene_df_norm.T, how='inner')
+
+# 4) Handle missing values systematically
+processed_df = handle_missing_values(linked_data, trait_col=trait)
+
+# 5) Determine trait/demographic bias and remove biased demographic features
+trait_biased, processed_df = judge_and_remove_biased_features(processed_df, trait)
+
+# 6) Final quality validation and save cohort info
+# Ensure native Python bools
+is_gene_available = bool(gene_df_norm.shape[0] > 0)
+is_trait_available = bool((trait in processed_df.columns) and (processed_df[trait].notna().sum() > 0))
+trait_biased_bool = bool(trait_biased)
+
+genes_before = int(genetic_df.shape[0]) if isinstance(genetic_df, pd.DataFrame) else None
+genes_after = int(gene_df_norm.shape[0])
+note = (
+    f"INFO: Selected cohort TCGA. Trait from TCGA barcode; Age column='{age_col_use}'; "
+    f"Gender column='{gender_col_use}'. Genes normalized using NCBI synonyms; kept {genes_after} of {genes_before} rows."
+)
+
+# Defensive retry with explicit casting if JSON serialization fails
+try:
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort='TCGA',
+        info_path=json_path,
+        is_gene_available=is_gene_available,
+        is_trait_available=is_trait_available,
+        is_biased=trait_biased_bool,
+        df=processed_df,
+        note=note
+    )
+except TypeError as e:
+    print(f"DEBUG: validate_and_save_cohort_info failed with {e}. Retrying after enforcing basic Python types.")
+    processed_df_retry = processed_df.copy()
+    processed_df_retry.columns = [str(c) for c in processed_df_retry.columns]
+    is_usable = validate_and_save_cohort_info(
+        is_final=True,
+        cohort='TCGA',
+        info_path=json_path,
+        is_gene_available=bool(is_gene_available),
+        is_trait_available=bool(is_trait_available),
+        is_biased=bool(trait_biased_bool),
+        df=processed_df_retry,
+        note=str(note)
+    )
+
+# 7) Save linked data (and clinical data) if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    processed_df.to_csv(out_data_file)
+    # Optionally save the standardized clinical table
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file)

output/preprocess/Von_Hippel_Lindau/cohort_info.json

@@ -1,22 +1 @@
-{
-  "GSE33093": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": true,
-    "has_age": false,
-    "has_gender": false,
-    "sample_size": 44
-  },
-  "TCGA": {
-    "is_usable": false,
-    "is_gene_available": true,
-    "is_trait_available": true,
-    "is_available": true,
-    "is_biased": true,
-    "has_age": true,
-    "has_gender": true,
-    "sample_size": 187
-  }
-}
+{"GSE33093": {"is_usable": false, "is_gene_available": true, "is_trait_available": false, "is_available": false, "is_biased": null, "has_age": null, "has_gender": null, "sample_size": null, "note": "INFO: Trait data not available in this cohort; only gene data were processed and saved."}, "TCGA": {"is_usable": false, "is_gene_available": true, "is_trait_available": true, "is_available": true, "is_biased": true, "has_age": true, "has_gender": true, "sample_size": 187, "note": "INFO: Selected cohort TCGA. Trait from TCGA barcode; Age column='age_at_initial_pathologic_diagnosis'; Gender column='gender'. Genes normalized using NCBI synonyms; kept 19848 of 20530 rows."}}

output/preprocess/Von_Willebrand_Disease/clinical_data/GSE27597.csv

@@ -1,4 +1,4 @@
-0,1,2,3,4,5,6,7
-0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
-,,,,,,,
-,,,,1.0,,,
+GSM684089,GSM684090,GSM684091,GSM684092,GSM684093,GSM684094,GSM684095,GSM684096,GSM684097,GSM684098,GSM684101,GSM684103,GSM684105,GSM684107,GSM684109,GSM684112,GSM684114,GSM684117,GSM684119,GSM684120,GSM684121,GSM684122,GSM684123,GSM684124,GSM684125,GSM684126,GSM684127,GSM684128,GSM684129,GSM684130,GSM684132,GSM684133,GSM684135,GSM684136,GSM684139,GSM684141,GSM684143,GSM684144,GSM684145,GSM684146,GSM684147,GSM684148,GSM684149,GSM684150,GSM684151,GSM684152,GSM684153,GSM684154,GSM684155,GSM684156,GSM684157,GSM684158,GSM684159,GSM684160,GSM684161,GSM684162,GSM684163,GSM684164,GSM684165,GSM684166,GSM684167,GSM684168,GSM684169,GSM684170
+0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0
+62.0,62.0,62.0,62.0,62.0,62.0,62.0,62.0,61.0,61.0,61.0,61.0,61.0,61.0,61.0,61.0,63.0,63.0,63.0,63.0,63.0,63.0,63.0,63.0,56.0,56.0,56.0,56.0,56.0,56.0,56.0,56.0,55.0,55.0,55.0,55.0,55.0,55.0,55.0,55.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,59.0,62.0,62.0,62.0,62.0,62.0,62.0,62.0,62.0
+1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,0.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0,1.0

output/preprocess/Von_Willebrand_Disease/code/GSE27597.py

@@ -0,0 +1,181 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Von_Willebrand_Disease"
+cohort = "GSE27597"
+
+# Input paths
+in_trait_dir = "../DATA/GEO/Von_Willebrand_Disease"
+in_cohort_dir = "../DATA/GEO/Von_Willebrand_Disease/GSE27597"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/GSE27597.csv"
+out_gene_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/gene_data/GSE27597.csv"
+out_clinical_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/clinical_data/GSE27597.csv"
+json_path = "./output/z6/preprocess/Von_Willebrand_Disease/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+from tools.preprocess import *
+# 1. Identify the paths to the SOFT file and the matrix file
+soft_file, matrix_file = geo_get_relevant_filepaths(in_cohort_dir)
+
+# 2. Read the matrix file to obtain background information and sample characteristics data
+background_prefixes = ['!Series_title', '!Series_summary', '!Series_overall_design']
+clinical_prefixes = ['!Sample_geo_accession', '!Sample_characteristics_ch1']
+background_info, clinical_data = get_background_and_clinical_data(matrix_file, background_prefixes, clinical_prefixes)
+
+# 3. Obtain the sample characteristics dictionary from the clinical dataframe
+sample_characteristics_dict = get_unique_values_by_row(clinical_data)
+
+# 4. Explicitly print out all the background information and the sample characteristics dictionary
+print("Background Information:")
+print(background_info)
+print("Sample Characteristics Dictionary:")
+print(sample_characteristics_dict)
+
+# Step 2: Dataset Analysis and Clinical Feature Extraction
+import os
+import re
+
+# 1) Gene expression data availability
+is_gene_available = True  # Affymetrix Human Exon/Gene ST arrays indicate mRNA gene expression data
+
+# 2) Variable availability and converters based on provided Sample Characteristics Dictionary
+# Keys identified from the dictionary:
+# 7 -> notes: contains 'von Willebrand disease' in some samples (usable for trait)
+# 5 -> age
+# 4 -> Sex
+
+trait_row = 7
+age_row = 5
+gender_row = 4
+
+def _after_colon(x: str) -> str:
+    if x is None:
+        return ""
+    parts = str(x).split(":", 1)
+    return parts[1].strip() if len(parts) > 1 else str(x).strip()
+
+def convert_trait(x):
+    # Binary: 1 if notes mention von Willebrand disease, else 0
+    val = _after_colon(x).lower()
+    if val == "" or val in {"na", "n/a", "none", "unknown"}:
+        # If explicitly unknown/none, treat as absence (0) rather than missing, since 'none' is informative here.
+        return 0
+    return 1 if "von willebrand" in val else 0
+
+def convert_age(x):
+    # Continuous: extract numeric age
+    val = _after_colon(x)
+    if val is None or val == "":
+        return None
+    m = re.search(r"[-+]?\d*\.?\d+", val)
+    if not m:
+        return None
+    try:
+        return float(m.group(0))
+    except Exception:
+        return None
+
+def convert_gender(x):
+    # Binary: female -> 0, male -> 1
+    val = _after_colon(x).lower()
+    if val.startswith("m"):
+        return 1
+    if val.startswith("f"):
+        return 0
+    return None
+
+# 3) Initial filtering metadata save
+is_trait_available = trait_row is not None
+_ = validate_and_save_cohort_info(
+    is_final=False,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=is_gene_available,
+    is_trait_available=is_trait_available
+)
+
+# 4) Clinical feature extraction (only if trait is available)
+if is_trait_available:
+    selected_clinical_df = geo_select_clinical_features(
+        clinical_df=clinical_data,
+        trait=trait,
+        trait_row=trait_row,
+        convert_trait=convert_trait,
+        age_row=age_row,
+        convert_age=convert_age,
+        gender_row=gender_row,
+        convert_gender=convert_gender
+    )
+    clinical_preview = preview_df(selected_clinical_df)
+    print(clinical_preview)
+
+    os.makedirs(os.path.dirname(out_clinical_data_file), exist_ok=True)
+    selected_clinical_df.to_csv(out_clinical_data_file, index=False)
+
+# Step 3: Gene Data Extraction
+# 1. Use the get_genetic_data function from the library to get the gene_data from the matrix_file previously defined.
+gene_data = get_genetic_data(matrix_file)
+
+# 2. Print the first 20 row IDs (gene or probe identifiers) for future observation.
+print(gene_data.index[:20])
+
+# Step 4: Gene Identifier Review
+# The observed identifiers are numeric probe IDs (e.g., Affymetrix probesets), not human gene symbols.
+requires_gene_mapping = True
+print(f"requires_gene_mapping = {requires_gene_mapping}")
+
+# Step 5: Gene Annotation
+# 1. Use the 'get_gene_annotation' function from the library to get gene annotation data from the SOFT file.
+gene_annotation = get_gene_annotation(soft_file)
+
+# 2. Use the 'preview_df' function from the library to preview the data and print out the results.
+print("Gene annotation preview:")
+print(preview_df(gene_annotation))
+
+# Step 6: Gene Identifier Mapping
+# 1-2. Decide identifier and gene symbol columns and build mapping dataframe
+# Probe identifier column: 'ID' matches the probe IDs in the expression data
+# Gene symbol information column: 'gene_assignment' (contains gene symbols within annotation text)
+mapping_df = get_gene_mapping(gene_annotation, prob_col='ID', gene_col='gene_assignment')
+
+# 3. Apply mapping to convert probe-level data to gene-level expression
+gene_data = apply_gene_mapping(expression_df=gene_data, mapping_df=mapping_df)
+
+# Step 7: Data Normalization and Linking
+import os
+
+# 1. Normalize gene symbols and save the normalized gene expression data
+normalized_gene_data = normalize_gene_symbols_in_index(gene_data)
+os.makedirs(os.path.dirname(out_gene_data_file), exist_ok=True)
+normalized_gene_data.to_csv(out_gene_data_file)
+
+# 2. Link the clinical and genetic data
+linked_data = geo_link_clinical_genetic_data(selected_clinical_df, normalized_gene_data)
+
+# 3. Handle missing values systematically
+linked_data = handle_missing_values(linked_data, trait)
+
+# 4. Assess bias and remove biased demographic features if necessary
+is_trait_biased, unbiased_linked_data = judge_and_remove_biased_features(linked_data, trait)
+
+# 5. Final validation and save cohort metadata
+note = "INFO: Trait inferred from 'notes'; highly imbalanced across samples."
+is_usable = validate_and_save_cohort_info(
+    is_final=True,
+    cohort=cohort,
+    info_path=json_path,
+    is_gene_available=True,
+    is_trait_available=True,
+    is_biased=is_trait_biased,
+    df=unbiased_linked_data,
+    note=note
+)
+
+# 6. Save the linked data if usable
+if is_usable:
+    os.makedirs(os.path.dirname(out_data_file), exist_ok=True)
+    unbiased_linked_data.to_csv(out_data_file)

output/preprocess/Von_Willebrand_Disease/code/TCGA.py

@@ -0,0 +1,65 @@
+# Path Configuration
+from tools.preprocess import *
+
+# Processing context
+trait = "Von_Willebrand_Disease"
+
+# Input paths
+tcga_root_dir = "../DATA/TCGA"
+
+# Output paths
+out_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/TCGA.csv"
+out_gene_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/gene_data/TCGA.csv"
+out_clinical_data_file = "./output/z6/preprocess/Von_Willebrand_Disease/clinical_data/TCGA.csv"
+json_path = "./output/z6/preprocess/Von_Willebrand_Disease/cohort_info.json"
+
+
+# Step 1: Initial Data Loading
+import os
+import re
+import pandas as pd
+
+# Step 1: Identify the most relevant TCGA cohort directory for Von Willebrand Disease (VWD)
+dirs = sorted([d for d in os.listdir(tcga_root_dir) if os.path.isdir(os.path.join(tcga_root_dir, d))])
+
+trait_lower = trait.lower()
+# Heuristic keyword set for VWD and related hemostasis/bleeding disorders
+keywords = [
+    "von", "willebrand", "vwd", "bleed", "coagul", "hemost", "platelet", "thrombo", "hemorr"
+]
+
+def score_dir(name: str, kws):
+    name_l = name.lower()
+    return sum(1 for k in kws if k in name_l)
+
+scored = [(d, score_dir(d, keywords)) for d in dirs]
+# Select directory with highest score (>0), prefer most specific (highest score), else None
+scored_sorted = sorted(scored, key=lambda x: (-x[1], len(x[0])))
+selected_dir = scored_sorted[0][0] if scored_sorted and scored_sorted[0][1] > 0 else None
+
+if selected_dir is None:
+    print("No suitable TCGA cohort found for Von Willebrand Disease; skipping TCGA for this trait.")
+    # Record unavailability in cohort metadata
+    validate_and_save_cohort_info(
+        is_final=False,
+        cohort="TCGA",
+        info_path=json_path,
+        is_gene_available=False,
+        is_trait_available=False
+    )
+else:
+    cohort_dir = os.path.join(tcga_root_dir, selected_dir)
+    print(f"Selected cohort directory: {selected_dir}")
+
+    # Step 2: Identify clinicalMatrix and PANCAN file paths
+    clinical_file_path, genetic_file_path = tcga_get_relevant_filepaths(cohort_dir)
+    print(f"Clinical file: {os.path.basename(clinical_file_path)}")
+    print(f"Genetic file: {os.path.basename(genetic_file_path)}")
+
+    # Step 3: Load both files as DataFrames
+    clinical_df = pd.read_csv(clinical_file_path, sep='\t', index_col=0, low_memory=False)
+    genetic_df = pd.read_csv(genetic_file_path, sep='\t', index_col=0, low_memory=False)
+
+    # Step 4: Print clinical column names
+    print("Clinical data columns:")
+    print(list(clinical_df.columns))