step2_CECs_BatchAnnotator_v1.0.py

# ==================== CECs BatchAnnotator v1.0 (Desktop Version) ====================
# Supports batch query (AC50 matching function removed)

import tkinter as tk
from tkinter import ttk, filedialog, messagebox, scrolledtext
import pandas as pd
import requests
import json
import os
import time
from typing import Optional, Dict, List
from datetime import datetime
import threading
import sys


# ==================== Core Function Module ====================
class DifyBasicChat:
    """Dify Basic Chat Function Encapsulation"""

    def __init__(self, api_key: str, base_url: str = "http://localhost/v1"):
        self.api_key = api_key
        self.base_url = base_url.rstrip("/")
        self.headers = {
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json"
        }

    def send_message(
            self,
            query: str,
            user: str,
            conversation_id: Optional[str] = None,
            response_mode: str = "blocking",
            inputs: dict = None
    ) -> tuple:
        """Send chat message"""
        url = f"{self.base_url}/chat-messages"
        payload = {
            "query": query,
            "user": user,
            "response_mode": response_mode,
            "inputs": inputs or {}
        }
        if conversation_id:
            payload["conversation_id"] = conversation_id

        full_response = None
        try:
            if response_mode == "blocking":
                res = requests.post(url, headers=self.headers, json=payload, timeout=120)
                res.raise_for_status()
                full_response = res.json()
                answer = full_response.get("answer", "")
                conv_id = full_response.get("conversation_id")
                return answer, conv_id, full_response

            else:
                full_answer = ""
                conv_id = None
                res = requests.post(url, headers=self.headers, json=payload, stream=True, timeout=120)
                res.raise_for_status()
                for line in res.iter_lines():
                    if line:
                        line_data = line.decode("utf-8").lstrip("data: ")
                        if line_data:
                            try:
                                data = json.loads(line_data)
                                full_response = data
                                if data.get("event") == "message":
                                    full_answer += data.get("answer", "")
                                elif data.get("event") == "message_end":
                                    conv_id = data.get("conversation_id")
                                    break
                                elif data.get("event") == "error":
                                    raise Exception(f"Streaming Error: {data.get('message')}")
                            except json.JSONDecodeError:
                                continue
                return full_answer, conv_id, full_response

        except requests.exceptions.RequestException as e:
            error_msg = f"Request Failed: {str(e)}"
            return error_msg, None, {"error": error_msg}


def parse_dify_response(answer_text: str) -> dict:
    """Parse classification and complete information returned by Dify"""
    result = {
        #"CASRN": "",
        "Main Category": "",
        "Additional Category 1": "",
        "Additional Category 2": "",
        "EndpointName": [],  # Keep for compatibility, no longer used for matching
        "XLogP": "",
        "BioPathway": "",
        "ToxicityInfo": "",
        "KnownUse": "",
        "DisorderDisease": ""
    }

    try:
        clean_text = answer_text.strip()
        # Clean code block markers
        if clean_text.startswith("```json"):
            clean_text = clean_text.replace("```json", "").replace("```", "").strip()
        elif clean_text.startswith("```"):
            clean_text = clean_text.replace("```", "").strip()

        # Parse JSON
        response_json = json.loads(clean_text)

        if isinstance(response_json, dict):
            # Get compound name (first key)
            compound_name = next(iter(response_json.keys())) if response_json else ""

            if compound_name and isinstance(response_json.get(compound_name), dict):
                # Nested format: {"CompoundName": {...}}
                category_info = response_json[compound_name]

                # Extract all fields
                #result["CASRN"] = category_info.get("CASRN", "")
                result["Main Category"] = category_info.get("Main Category", "")
                result["Additional Category 1"] = category_info.get("Additional Category 1", "")
                result["Additional Category 2"] = category_info.get("Additional Category 2", "")

                # Process EndpointName - may be list or string
                endpoint_value = category_info.get("EndpointName", [])
                if isinstance(endpoint_value, list):
                    result["EndpointName"] = endpoint_value
                elif isinstance(endpoint_value, str):
                    result["EndpointName"] = [endpoint_value] if endpoint_value else []

                result["XLogP"] = category_info.get("XLogP", "")
                result["BioPathway"] = category_info.get("BioPathway", "")
                result["ToxicityInfo"] = category_info.get("ToxicityInfo", "")
                result["KnownUse"] = category_info.get("KnownUse", "")
                result["DisorderDisease"] = category_info.get("DisorderDisease", "")

            else:
                # Flat format (compatible with old format)
                result["Main Category"] = response_json.get("Main Category", "")
                result["Additional Category 1"] = response_json.get("Additional Category 1", "")
                result["Additional Category 2"] = response_json.get("Additional Category 2", "")

    except json.JSONDecodeError as e:
        result["Main Category"] = f"JSON Parsing Error: {str(e)}"
    except Exception as e:
        result["Main Category"] = f"Parsing Failed: {str(e)}"

    return result


def normalize_compound_name(name: str) -> str:
    """Normalize compound name (remove quotes, etc.)"""
    if not isinstance(name, str):
        return ""

    # Remove quotes
    name = name.strip()
    if name.startswith('"') and name.endswith('"'):
        name = name[1:-1]
    elif name.startswith("'") and name.endswith("'"):
        name = name[1:-1]

    # Remove extra spaces
    name = ' '.join(name.split())

    return name


def expand_endpoint_rows(parsed_result: dict, compound_name: str) -> list:
    """
    Expand EndpointName into multiple rows (without AC50 matching)
    """
    rows = []

    # Normalize compound name
    compound_clean = normalize_compound_name(compound_name)

    endpoint_names = parsed_result.get("EndpointName", [])

    if not endpoint_names:
        # Create one row if no EndpointName
        row = {
            "CompoundName": compound_clean,
            "OriginalCompoundName": compound_name,
            #"CASRN": parsed_result.get("CASRN", ""),
            "MainCategory": parsed_result.get("Main Category", ""),
            "AdditionalCategory1": parsed_result.get("Additional Category 1", ""),
            "AdditionalCategory2": parsed_result.get("Additional Category 2", ""),
            "EndpointName": "",
            "XLogP": parsed_result.get("XLogP", ""),
            "BioPathway": parsed_result.get("BioPathway", ""),
            "ToxicityInfo": parsed_result.get("ToxicityInfo", ""),
            "KnownUse": parsed_result.get("KnownUse", ""),
            "DisorderDisease": parsed_result.get("DisorderDisease", "")
        }
        rows.append(row)
    else:
        # Create one row per endpoint
        for endpoint in endpoint_names:
            row = {
                "CompoundName": compound_clean,
                "OriginalCompoundName": compound_name,
                #"CASRN": parsed_result.get("CASRN", ""),
                "MainCategory": parsed_result.get("Main Category", ""),
                "AdditionalCategory1": parsed_result.get("Additional Category 1", ""),
                "AdditionalCategory2": parsed_result.get("Additional Category 2", ""),
                "EndpointName": endpoint,
                "XLogP": parsed_result.get("XLogP", ""),
                "BioPathway": parsed_result.get("BioPathway", ""),
                "ToxicityInfo": parsed_result.get("ToxicityInfo", ""),
                "KnownUse": parsed_result.get("KnownUse", ""),
                "DisorderDisease": parsed_result.get("DisorderDisease", "")
            }
            rows.append(row)

    return rows


def batch_process_compounds_gui(
        csv_path: str,
        save_root: str,
        api_key: str,
        base_url: str,
        log_text: tk.Text,
        progress_var: tk.DoubleVar,
        user_id: str = "batch_compound_user",
        compound_col: str = "IUPAC_name",
        batch_num: int = 1,
        csv_encoding: str = "utf-8",
        csv_sep: str = ","
):
    """Batch process compounds (adapted for GUI)"""

    def log(message, color="black"):
        """Output log to GUI text box"""
        log_text.config(state=tk.NORMAL)
        log_text.insert(tk.END, f"[{datetime.now().strftime('%Y-%m-%d %H:%M:%S')}] {message}\n")
        log_text.tag_add(color, f"end-2l", f"end-1l")
        log_text.tag_config(color, foreground=color)
        log_text.see(tk.END)
        log_text.config(state=tk.DISABLED)
        log_text.update()

    try:
        # Initialize Dify client
        log("Initializing Dify connection...", "blue")
        chat = DifyBasicChat(api_key=api_key, base_url=base_url)

        # Create save folder
        result_folder = os.path.join(save_root,
                                     f"Compound_Classification_Results_Batch{batch_num}_{datetime.now().strftime('%Y%m%d%H%M%S')}")
        os.makedirs(result_folder, exist_ok=True)
        log(f"Result save folder: {result_folder}", "blue")

        # Read CSV
        log("Reading CSV file...", "blue")
        df = pd.read_csv(
            csv_path,
            encoding=csv_encoding,
            sep=csv_sep,
            na_filter=True
        )
        df = df.reset_index(drop=True)

        # Check if column exists
        if compound_col not in df.columns:
            raise ValueError(
                f"Column not found in CSV: [{compound_col}]\n"
                f"Current CSV columns: {list(df.columns)}"
            )

        # Remove duplicates and empty values
        compounds = df[compound_col].dropna().unique()
        total = len(compounds)
        log(f"Successfully read {total} non-empty and unique compound names", "green")

        all_rows = []  # Store all row data
        failed_list = []

        # Batch processing
        for idx, compound in enumerate(compounds, 1):
            compound = str(compound).strip()
            if not compound:
                continue

            # Update progress
            progress = (idx / total) * 100
            progress_var.set(progress)
            log(f"Processing {idx}/{total}：{compound}", "black")

            try:
                # Call Dify API
                answer, _, full_response = chat.send_message(
                    query=compound,
                    user=f"{user_id}_batch{batch_num}",
                    response_mode="blocking"
                )

                # Parse results
                parsed_categories = parse_dify_response(answer)

                # Expand EndpointName into multiple rows
                expanded_rows = expand_endpoint_rows(parsed_categories, compound)
                all_rows.extend(expanded_rows)

                # Save original record (for debugging)
                record_file = os.path.join(result_folder, f"Original_Record_{idx}.json")
                with open(record_file, "w", encoding="utf-8") as f:
                    json.dump({
                        "Input Compound": compound,
                        "Dify Original Response": answer,
                        "Complete Response": full_response,
                        "Parsed Classification": parsed_categories,
                        "Expanded Rows Count": len(expanded_rows),
                        "Timestamp": datetime.now().strftime("%Y-%m-%d %H:%M:%S")
                    }, f, ensure_ascii=False, indent=4)

                log(f"✅ Processing completed: {compound} | Main Category: {parsed_categories['Main Category']} | Generated {len(expanded_rows)} rows",
                    "green")
                time.sleep(0.5)  # Avoid too fast requests

            except Exception as e:
                error_msg = str(e)
                log(f"❌ Processing failed: {compound} | Error: {error_msg}", "red")
                failed_list.append({
                    "CompoundName": normalize_compound_name(compound),
                    "OriginalCompoundName": compound,
                    #"CASRN": "",
                    "MainCategory": f"Processing Failed: {error_msg}",
                    "AdditionalCategory1": "",
                    "AdditionalCategory2": "",
                    "EndpointName": "",
                    "XLogP": "",
                    "BioPathway": "",
                    "ToxicityInfo": "",
                    "KnownUse": "",
                    "DisorderDisease": ""
                })

        # Merge results and save
        result_df = pd.DataFrame(all_rows)

        # Add failed records
        if failed_list:
            failed_df = pd.DataFrame(failed_list)
            result_df = pd.concat([result_df, failed_df], ignore_index=True)

        # Define column order
        column_order = [
            "CompoundName",
            "OriginalCompoundName",
            #"CASRN",
            "MainCategory",
            "AdditionalCategory1",
            "AdditionalCategory2",
            "EndpointName",
            "XLogP",
            "BioPathway",
            "ToxicityInfo",
            "KnownUse",
            "DisorderDisease"
        ]

        # Ensure all columns exist
        for col in column_order:
            if col not in result_df.columns:
                result_df[col] = ""

        # Reorder columns
        result_df = result_df.reindex(columns=column_order)

        # Save final CSV
        csv_filename = f"Compound_Query_Results_Batch{batch_num}.csv"
        csv_path_out = os.path.join(result_folder, csv_filename)
        result_df.to_csv(csv_path_out, index=False, encoding="utf-8-sig")
        log(f"📄 Result file saved to: {csv_path_out}", "blue")
        log(f"📊 Total Rows: {len(result_df)} rows", "blue")

        # Save failed list (separate file)
        if failed_list:
            fail_file = os.path.join(result_folder, f"Failed_List_Batch{batch_num}.csv")
            pd.DataFrame(failed_list).to_csv(fail_file, index=False, encoding="utf-8-sig")
            log(f"❌ {len(failed_list)} compounds failed to process, details: {fail_file}", "red")

        # Update progress and log after completion
        progress_var.set(100)
        log(f"\n{'=' * 40}", "blue")
        log(f"🏁 Processing Complete!", "green")
        log(f"{'=' * 40}", "blue")
        log(f"📊 Statistics: Total Compounds={total} | Successful Rows={len(all_rows)} | Failed Compounds={len(failed_list)}", "blue")
        log(f"📁 All results saved to: {result_folder}", "blue")

        # Ask if open result folder
        if messagebox.askyesno("Processing Complete", f"Batch processing completed!\nTotal {len(result_df)} rows of data generated\nOpen result folder?"):
            if os.name == 'nt':  # Windows
                os.startfile(result_folder)
            elif os.name == 'posix':  # macOS, Linux
                import subprocess
                try:
                    if sys.platform == 'darwin':
                        subprocess.run(['open', result_folder])
                    else:
                        subprocess.run(['xdg-open', result_folder])
                except:
                    pass

    except Exception as e:
        log(f"❌ Overall processing failed: {str(e)}", "red")
        messagebox.showerror("Error", f"Processing failed: {str(e)}")
    finally:
        # Reset progress
        progress_var.set(0)


# ==================== Graphical User Interface Module ====================
class CompoundBatchToolGUI:
    def __init__(self, root):
        self.root = root
        self.root.title("CECs BatchAnnotator v1.0")
        self.root.geometry("850x700")
        self.root.resizable(True, True)

        # Default configuration
        self.default_api_key = "app-QRGuoLVqSksMsG4t9O53cITj"
        self.default_base_url = "http://192.168.0.179:8080/v1"
        self.default_save_root = "./Compound_Query_Results"
        self.default_compound_col = "IUPAC_name"
        self.default_csv_encoding = "utf-8"
        self.default_csv_sep = ","

        # Create main frame
        main_frame = ttk.Frame(root, padding="20")
        main_frame.pack(fill=tk.BOTH, expand=True)

        # 1. File selection area
        file_frame = ttk.LabelFrame(main_frame, text="1. Select CSV File", padding="10")
        file_frame.pack(fill=tk.X, pady=5)

        self.csv_path_var = tk.StringVar()
        ttk.Entry(file_frame, textvariable=self.csv_path_var, state="readonly", width=65).grid(row=0, column=1, padx=5,
                                                                                               pady=5)
        ttk.Button(file_frame, text="Select File", command=self.select_csv_file).grid(row=0, column=0, padx=5, pady=5)

        # 2. Parameter configuration area
        param_frame = ttk.LabelFrame(main_frame, text="2. Parameter Configuration", padding="10")
        param_frame.pack(fill=tk.X, pady=5)

        # 2.1 Dify configuration
        ttk.Label(param_frame, text="Dify API Key：").grid(row=0, column=0, sticky=tk.W, padx=5, pady=3)
        self.api_key_var = tk.StringVar(value=self.default_api_key)
        ttk.Entry(param_frame, textvariable=self.api_key_var, width=60).grid(row=0, column=1, columnspan=3, padx=5,
                                                                             pady=3)

        ttk.Label(param_frame, text="Dify URL：").grid(row=1, column=0, sticky=tk.W, padx=5, pady=3)
        self.base_url_var = tk.StringVar(value=self.default_base_url)
        ttk.Entry(param_frame, textvariable=self.base_url_var, width=60).grid(row=1, column=1, columnspan=3, padx=5,
                                                                              pady=3)

        # 2.2 CSV configuration
        ttk.Label(param_frame, text="Compound Column Name：").grid(row=2, column=0, sticky=tk.W, padx=5, pady=3)
        self.compound_col_var = tk.StringVar(value=self.default_compound_col)
        ttk.Entry(param_frame, textvariable=self.compound_col_var, width=20).grid(row=2, column=1, padx=5, pady=3)

        ttk.Label(param_frame, text="CSV Encoding：").grid(row=2, column=2, sticky=tk.W, padx=5, pady=3)
        self.csv_encoding_var = tk.StringVar(value=self.default_csv_encoding)
        ttk.Entry(param_frame, textvariable=self.csv_encoding_var, width=15).grid(row=2, column=3, padx=5, pady=3)

        ttk.Label(param_frame, text="CSV Separator：").grid(row=3, column=0, sticky=tk.W, padx=5, pady=3)
        self.csv_sep_var = tk.StringVar(value=self.default_csv_sep)
        ttk.Entry(param_frame, textvariable=self.csv_sep_var, width=20).grid(row=3, column=1, padx=5, pady=3)

        # 2.3 Save configuration (AC50 folder removed)
        ttk.Label(param_frame, text="Result Save Path：").grid(row=4, column=0, sticky=tk.W, padx=5, pady=3)
        self.save_root_var = tk.StringVar(value=self.default_save_root)
        ttk.Entry(param_frame, textvariable=self.save_root_var, width=50).grid(row=4, column=1, columnspan=2, padx=5,
                                                                               pady=3)
        ttk.Button(param_frame, text="Select Path", command=self.select_save_root).grid(row=4, column=3, padx=5, pady=3)

        # 3. Operation area
        op_frame = ttk.LabelFrame(main_frame, text="3. Start Processing", padding="10")
        op_frame.pack(fill=tk.X, pady=5)

        self.progress_var = tk.DoubleVar()
        progress_bar = ttk.Progressbar(op_frame, variable=self.progress_var, maximum=100)
        progress_bar.pack(fill=tk.X, padx=5, pady=5)

        self.start_btn = ttk.Button(op_frame, text="Start Batch Processing", command=self.start_processing)
        self.start_btn.pack(pady=5)

        # 4. Log output area
        log_frame = ttk.LabelFrame(main_frame, text="4. Processing Log", padding="10")
        log_frame.pack(fill=tk.BOTH, expand=True, pady=5)

        self.log_text = scrolledtext.ScrolledText(log_frame, wrap=tk.WORD, state=tk.DISABLED)
        self.log_text.pack(fill=tk.BOTH, expand=True)
        # Set log color tags
        self.log_text.tag_config("red", foreground="red")
        self.log_text.tag_config("green", foreground="green")
        self.log_text.tag_config("blue", foreground="blue")
        self.log_text.tag_config("orange", foreground="orange")
        self.log_text.tag_config("gray", foreground="gray")

        # 5. Bottom tip (AC50 related tip removed)
        tip_label = ttk.Label(main_frame,
                              text="Tip: Each endpoint returned by Dify generates a separate row in the result",
                              foreground="gray")
        tip_label.pack(side=tk.BOTTOM, pady=10)

    def select_csv_file(self):
        """Select CSV file"""
        file_path = filedialog.askopenfilename(
            title="Select Compound CSV File",
            filetypes=[("CSV Files", "*.csv"), ("All Files", "*.*")]
        )
        if file_path:
            self.csv_path_var.set(file_path)

    def select_save_root(self):
        """Select save path"""
        folder_path = filedialog.askdirectory(title="Select Result Save Folder")
        if folder_path:
            self.save_root_var.set(folder_path)

    def start_processing(self):
        """Start batch processing (new thread to avoid UI freezing)"""
        # Verify required parameters
        csv_path = self.csv_path_var.get()
        if not csv_path:
            messagebox.showwarning("Warning", "Please select a CSV file first!")
            return

        api_key = self.api_key_var.get().strip()
        if not api_key:
            messagebox.showwarning("Warning", "Please fill in the Dify API Key!")
            return

        base_url = self.base_url_var.get().strip()
        if not base_url:
            messagebox.showwarning("Warning", "Please fill in the Dify URL!")
            return

        # Disable start button to prevent duplicate clicks
        self.start_btn.config(state=tk.DISABLED)

        # Clear log
        self.log_text.config(state=tk.NORMAL)
        self.log_text.delete(1.0, tk.END)
        self.log_text.config(state=tk.DISABLED)

        # New thread for processing (avoid UI freezing)
        def process_thread():
            try:
                batch_process_compounds_gui(
                    csv_path=csv_path,
                    save_root=self.save_root_var.get(),
                    api_key=api_key,
                    base_url=base_url,
                    log_text=self.log_text,
                    progress_var=self.progress_var,
                    compound_col=self.compound_col_var.get(),
                    csv_encoding=self.csv_encoding_var.get(),
                    csv_sep=self.csv_sep_var.get()
                )
            finally:
                # Restore button state
                self.start_btn.config(state=tk.NORMAL)

        threading.Thread(target=process_thread, daemon=True).start()


# ==================== Start Program ====================
if __name__ == "__main__":
    # Normal GUI startup
    root = tk.Tk()
    app = CompoundBatchToolGUI(root)
    root.mainloop()