Upload app.py

app.py

@@ -12,16 +12,16 @@ try:
 except Exception:
     HAS_FUZZ = False
 
-APP_TITLE = "Ward Ranking Random Assigner"
+APP_TITLE = "Ward Ranking Cleaner & Random Assigner (Auto-map + Stepwise Reports)"
 DESCRIPTION = """
 **Flow**
 1) Upload .csv/.xlsx  
 2) Choose wards + set capacity  
 3) Check Available columns  
 4) Map by Auto-detect (Thai/English + fuzzy) or by numbers (1-based)  
-5) Clean → keep NAME/ID + selected wards; convert ranks to integers  
-6) Assign → round-by-rank with random tie-breaking; respect capacity  
-   - Check: #students <= total capacity (shortage allowed, not exceed)
+5) **Clean** → keep NAME/ID + selected wards; convert ranks to integers  
+6) **Show Stepwise Report** → Rank 1 results → random assign rank 1 → Rank 2 results → random assign … until done  
+7) (Optional) Assign button runs the full allocation too and provides CSVs to download
 """
 
 WARD_CHOICES = [
@@ -56,9 +56,9 @@ AUTO_MAP = {
     "NAME": ["ชื่อ-สกุล", "ชื่อ - สกุล", "fullname", "full name", "name", "student name"],
     "ID": ["รหัสนักศึกษา", "รหัส", "student id", "id", "studentid"],
     "Medical": ["อายุรศาสตร์", "medical"],
-    "Medical_1": ["อายุรศาสตร์_1", "medical_1", "med_1", "med1"],
-    "Medical_2": ["อายุรศาสตร์_2", "medical_2", "med_2", "med2"],
-    "Surgical": ["ศัลยศาสตร์", "surgical", "surgery", "surg"],
+    "Medical_1": ["อายุรศาสตร์_1", "medical_1", "med_1"],
+    "Medical_2": ["อายุรศาสตร์_2", "medical_2", "med_2"],
+    "Surgical": ["ศัลยศาสตร์", "surgical", "surgery"],
     "Pediatric": ["เด็ก", "pediatric", "pediatrics"],
     "Community": ["ชุมชน", "community"],
     "Psychiatric": ["จิตเวช", "psychiatric"],
@@ -176,20 +176,92 @@ def build_cleaned_from_indices(df: pd.DataFrame,
     cleaned = cleaned[ordered]
     return cleaned
 
-def random_assign(cleaned: pd.DataFrame,
-                  capacities: Dict[str, int]) -> Tuple[pd.DataFrame, pd.DataFrame, Dict[str, int]]:
+def max_rank_in(cleaned: pd.DataFrame) -> int:
     wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
-    cap = {w: int(capacities.get(w, 0)) for w in wards}
-
-    assigned = pd.Series(index=cleaned.index, data=pd.NA, dtype="object")
-    choice_no = pd.Series(index=cleaned.index, data=pd.NA, dtype="Int64")
-
     mr = 0
     for w in wards:
         m = cleaned[w].max(skipna=True)
         if pd.notna(m):
             mr = max(mr, int(m))
+    return int(mr)
+
+# ===== Stepwise simulation & reports (random each round) =====
+def simulate_stepwise_report(cleaned: pd.DataFrame, capacities: Dict[str, int]) -> str:
+    """Round-by-round report: show rank r results, then randomly assign rank r, update remaining capacity, continue."""
+    wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
+    total_students = len(cleaned)
+    cap = {w: int(capacities.get(w, 0)) for w in wards}
+    assigned = pd.Series(index=cleaned.index, data=False)  # True if assigned already
+    assigned_ward = pd.Series(index=cleaned.index, data="", dtype="object")
+
+    mr = max_rank_in(cleaned)
+    lines = []
+    lines.append(f"### Total Students (จำนวนนักศึกษาทั้งหมด): {total_students} students (คน)")
+    lines.append(f"### Total Capacity (ความจุรวม): {sum(cap.values())} people (คน)")
+    lines.append("")
+
+    for r in range(1, mr + 1):
+        lines.append("\n---\n")
+        lines.append(f"## Rank {r} Results (การแสดงผลอันดับที่ {r})\n")
+        header = "| Ward (วอร์ด) | Remaining Capacity (ความจุคงเหลือ) | Rank {r} Count (จำนวนเลือกอันดับ {r}) | Students (รายชื่อนักศึกษา) |".format(r=r)
+        sep = "|---|---:|---:|---|"
+        lines += [header, sep]
+
+        # Show BEFORE assignment
+        for w in wards:
+            names = cleaned.loc[(~assigned) & (cleaned[w] == r), "NAME"].astype(str).tolist()
+            cnt = len(names)
+            sample = ", ".join(names[:3]) + ("..." if cnt > 3 else "")
+            lines.append(f"| {ward_display(w)} | {cap[w]} | {cnt} | {sample} |")
+
+        # Now perform random assignment at this rank
+        lines.append("")
+        lines.append(f"### Allocation at Rank {r} (การสุ่มจัดสรรในอันดับที่ {r})")
+        for w in wards:
+            candidates_idx = cleaned.index[(~assigned) & (cleaned[w] == r)].tolist()
+            if not candidates_idx or cap[w] <= 0:
+                lines.append(f"- {ward_display(w)}: No allocation (ไม่มีการจัดสรร)")
+                continue
+            if len(candidates_idx) <= cap[w]:
+                chosen = candidates_idx
+            else:
+                chosen = list(np.random.choice(candidates_idx, size=cap[w], replace=False))
+            assigned.loc[chosen] = True
+            assigned_ward.loc[chosen] = w
+            cap[w] -= len(chosen)
 
+            chosen_names = cleaned.loc[chosen, "NAME"].astype(str).tolist()
+            sample = ", ".join(chosen_names[:10]) + ("..." if len(chosen_names) > 10 else "")
+            lines.append(f"- {ward_display(w)} → Selected (คัดเลือก): {len(chosen_names)} | {sample}")
+
+        # After assignment at this rank
+        lines.append("")
+        lines.append("**Remaining capacity (จำนวนรับที่เหลือหลังรอบนี้):**")
+        for w in wards:
+            lines.append(f"- {ward_display(w)}: {cap[w]}")
+
+    # Final summary
+    lines.append("\n---\n")
+    lines.append("## Final Summary (สรุปสุดท้าย)")
+    for w in wards:
+        sel_names = assigned_ward[assigned_ward == w]
+        lines.append(f"- {ward_display(w)}: {len(sel_names)} assigned (จัดสรรแล้ว)")
+    unassigned = cleaned.loc[~assigned, "NAME"].astype(str).tolist()
+    lines.append(f"- Not assigned (ยังไม่ได้รับการจัดสรร): {len(unassigned)}")
+    if unassigned:
+        sample_un = ", ".join(unassigned[:15]) + ("..." if len(unassigned) > 15 else "")
+        lines.append(f"  - {sample_un}")
+    return "\n".join(lines)
+
+def random_assign(cleaned: pd.DataFrame,
+                  capacities: Dict[str, int]) -> Tuple[pd.DataFrame, pd.DataFrame, Dict[str, int]]:
+    """Full allocation using the same stepwise logic (for CSV outputs)."""
+    wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
+    cap = {w: int(capacities.get(w, 0)) for w in wards}
+    assigned = pd.Series(index=cleaned.index, data=pd.NA, dtype="object")
+    choice_no = pd.Series(index=cleaned.index, data=pd.NA, dtype="Int64")
+
+    mr = max_rank_in(cleaned)
     for r in range(1, mr + 1):
         if all(c <= 0 for c in cap.values()):
             break
@@ -211,112 +283,9 @@ def random_assign(cleaned: pd.DataFrame,
     result = cleaned.copy()
     result["AssignedWard"] = assigned
     result["ChoiceNumber"] = choice_no
-
     not_assigned = result[result["AssignedWard"].isna()].copy()
     return result.fillna(""), not_assigned.fillna(""), cap
 
-# ===== Reporting helpers =====
-def ward_display(ward_key: str) -> str:
-    en, th = WARD_LABELS.get(ward_key, (ward_key, ward_key))
-    return f"{en} ({th})"
-
-def max_rank_in(cleaned: pd.DataFrame) -> int:
-    wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
-    mr = 0
-    for w in wards:
-        m = cleaned[w].max(skipna=True)
-        if pd.notna(m):
-            mr = max(mr, int(m))
-    return int(mr)
-
-def make_rank1_report(cleaned: pd.DataFrame, capacities: Dict[str, int]) -> str:
-    wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
-    total_students = len(cleaned)
-    total_capacity = sum(int(capacities.get(w, 0)) for w in wards)
-    lines = []
-    lines.append("## Rank 1 Results (การแสดงผลอันดับที่ 1)")
-    lines.append("")
-    lines.append(f"- **Total Students (จำนวนนักศึกษาทั้งหมด):** {total_students} students (คน)")
-    lines.append(f"- **Total Capacity (ความจุรวม):** {total_capacity} people (คน)")
-    lines.append("")
-    header = "| Ward (วอร์ด) | Capacity (ความจุ) | Rank 1 Count (จำนวนเลือกอันดับ 1) | Students (รายชื่อนักศึกษา) |"
-    sep = "|---|---:|---:|---|"
-    lines += [header, sep]
-    over = []
-    under = []
-    for w in wards:
-        cap = int(capacities.get(w, 0))
-        rank1_students = cleaned.loc[cleaned[w] == 1, "NAME"].astype(str).tolist()
-        r1_count = len(rank1_students)
-        display_students = ", ".join(rank1_students[:3]) + ("..." if r1_count > 3 else "")
-        lines.append(f"| {ward_display(w)} | {cap} | {r1_count} | {display_students} |")
-        if r1_count > cap:
-            over.append((w, r1_count, cap))
-        elif r1_count < cap:
-            under.append((w, r1_count, cap))
-    lines.append("")
-    lines.append("### Additional Statistics (สถิติเพิ่มเติม)")
-    lines.append("")
-    if over:
-        lines.append("**Wards where Rank 1 count exceeds capacity (วอร์ดที่มีคนเลือกอันดับ 1 เกินความจุ):**")
-        for w, c, cap in over:
-            lines.append(f"- {ward_display(w)}: {c} selected (capacity {cap})")
-    else:
-        lines.append("- No wards exceed capacity at Rank 1. (ไม่มีวอร์ดใดเกินความจุในอันดับ 1)")
-    if under:
-        lines.append("")
-        lines.append("**Wards where Rank 1 count below capacity (วอร์ดที่มีคนเลือกอันดับ 1 น้อยกว่าความจุ):**")
-        for w, c, cap in under:
-            lines.append(f"- {ward_display(w)}: {c} selected (capacity {cap})")
-    return "\n".join(lines)
-
-def make_rank_report(cleaned: pd.DataFrame, capacities: Dict[str, int], rank: int) -> str:
-    wards = [w for w in cleaned.columns if w not in ("NAME", "ID")]
-    lines = []
-    lines.append(f"## Rank {rank} Results (การแสดงผลอันดับที่ {rank})")
-    total_students = len(cleaned)
-    total_capacity = sum(int(capacities.get(w, 0)) for w in wards)
-    lines.append(f"- **Total Students (จำนวนนักศึกษาทั้งหมด):** {total_students} students (คน)")
-    lines.append(f"- **Total Capacity (ความจุรวม):** {total_capacity} people (คน)")
-    lines.append("")
-    header = "| Ward (วอร์ด) | Capacity (ความจุ) | Rank {rank} Count (จำนวนเลือกอันดับ {rank}) | Students (รายชื่อนักศึกษา) |".format(rank=rank)
-    sep = "|---|---:|---:|---|"
-    lines += [header, sep]
-    over, under = [], []
-    for w in wards:
-        cap = int(capacities.get(w, 0))
-        names = cleaned.loc[cleaned[w] == rank, "NAME"].astype(str).tolist()
-        cnt = len(names)
-        sample = ", ".join(names[:3]) + ("..." if cnt > 3 else "")
-        lines.append(f"| {ward_display(w)} | {cap} | {cnt} | {sample} |")
-        if cnt > cap:
-            over.append((w, cnt, cap))
-        elif cnt < cap:
-            under.append((w, cnt, cap))
-    lines.append("")
-    lines.append("**Additional Statistics (สถิติเพิ่มเติม):**")
-    if over:
-        lines.append("- Wards where count exceeds capacity (เกินความจุ):")
-        for w, c, cap in over:
-            lines.append(f"  - {ward_display(w)}: {c} selected (capacity {cap})")
-    else:
-        lines.append("- No wards exceed capacity at this rank. (ไม่มีวอร์ดเกินความจุในอันดับนี้)")
-    if under:
-        lines.append("- Wards where count below capacity (ต่ำกว่าความจุ):")
-        for w, c, cap in under:
-            lines.append(f"  - {ward_display(w)}: {c} selected (capacity {cap})")
-    return "\n".join(lines)
-
-def make_all_ranks_report(cleaned: pd.DataFrame, capacities: Dict[str, int]) -> str:
-    mr = max_rank_in(cleaned)
-    if mr == 0:
-        return "No ranking numbers found. (ไม่พบข้อมูลอันดับเป็นตัวเลข)"
-    parts = []
-    for r in range(1, mr + 1):
-        parts.append(make_rank_report(cleaned, capacities, r))
-        parts.append("\n---\n")
-    return "\n".join(parts)
-
 # ===== Helpers for temp file paths =====
 def _tmp(name: str) -> str:
     os.makedirs("/tmp", exist_ok=True)
@@ -424,29 +393,9 @@ def _capacities_from_df(cleaned: pd.DataFrame, capacity_df: Optional[pd.DataFram
             capacities[str(row["Ward"])] = 0
     return capacities
 
-def on_rank1_report(file, selected_wards, capacity_df, name_num, id_num, 
-                    med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num):
-    df, msg = read_table(file)
-    if df is None:
-        return "Please upload a valid file."
-    n_cols = len(df.columns)
-    ward_nums = {
-        "Medical": med_num, "Medical_1": med1_num, "Medical_2": med2_num,
-        "Surgical": surg_num, "Pediatric": ped_num, "Community": comm_num,
-        "Psychiatric": psy_num, "Obstetrics": obs_num
-    }
-    errors, mapping_idx = collect_mapping_numbers(name_num, id_num, ward_nums, selected_wards, n_cols)
-    if errors:
-        return "❌ Mapping invalid:\n" + "\n".join(errors)
-    try:
-        cleaned = build_cleaned_from_indices(df, mapping_idx)
-    except Exception as e:
-        return f"❌ Error building cleaned data: {e}"
-    capacities = _capacities_from_df(cleaned, capacity_df)
-    return make_rank1_report(cleaned, capacities)
-
-def on_all_ranks_report(file, selected_wards, capacity_df, name_num, id_num, 
-                        med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num):
+def on_stepwise_report(file, selected_wards, capacity_df, name_num, id_num, 
+                       med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num):
+    # Requires cleaned mapping to be valid
     df, msg = read_table(file)
     if df is None:
         return "Please upload a valid file."
@@ -464,7 +413,7 @@ def on_all_ranks_report(file, selected_wards, capacity_df, name_num, id_num,
     except Exception as e:
         return f"❌ Error building cleaned data: {e}"
     capacities = _capacities_from_df(cleaned, capacity_df)
-    return make_all_ranks_report(cleaned, capacities)
+    return simulate_stepwise_report(cleaned, capacities)
 
 def on_assign(file, selected_wards, capacity_df, name_num, id_num,
               med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num):
@@ -482,7 +431,6 @@ def on_assign(file, selected_wards, capacity_df, name_num, id_num,
     }
     _errors, mapping_idx = collect_mapping_numbers(name_num, id_num, ward_nums, selected_wards, n_cols)
     cleaned = build_cleaned_from_indices(df, mapping_idx)
-
     capacities = _capacities_from_df(cleaned, capacity_df)
 
     total_capacity = sum(capacities.values())
@@ -500,9 +448,9 @@ def on_assign(file, selected_wards, capacity_df, name_num, id_num,
     not_assigned.to_csv(not_assigned_path, index=False, encoding="utf-8-sig")
 
     leftover_text = "Remaining capacity (จำนวนรับที่เหลือ):\n" + "\n".join([f"- {ward_display(k)}: {v}" for k, v in leftover.items()])
-    allocation = make_all_ranks_report(cleaned, capacities) + "\n\n---\n\n" + "## Allocation Summary (สรุปการจัดสรร)\n"  # Keep the all-ranks context too
+    stepwise_md = simulate_stepwise_report(cleaned, capacities)
 
-    return status, assigned.head(30), assigned_path, not_assigned_path, leftover_text, allocation
+    return status, assigned.head(30), assigned_path, not_assigned_path, leftover_text, stepwise_md
 
 with gr.Blocks(title=APP_TITLE) as demo:
     gr.Markdown(f"# {APP_TITLE}")
@@ -553,28 +501,20 @@ with gr.Blocks(title=APP_TITLE) as demo:
                        outputs=[status, available, name_num, id_num, 
                                 med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num])
 
-        # Reports (pre-assignment)
-        rank1_btn = gr.Button("Show Rank 1 Report (ดูสรุปอันดับ 1)")
-        rank1_report = gr.Markdown(label="Rank 1 Results (การแสดงผลอันดับที่ 1)")
-        rank1_btn.click(
-            fn=on_rank1_report,
-            inputs=[file, selected_wards, capacity_df, name_num, id_num,
-                    med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num],
-            outputs=rank1_report
-        )
+        # >>> Moved CLEAN button here (before reports) <<<
+        clean_btn = gr.Button("Clean data (ดูพรีวิว)", variant="primary")
+
+        # Reports (stepwise randomized)
+        step_btn = gr.Button("Show Stepwise Report (ดูรายงานทีละอันดับแบบสุ่ม)")
+        step_report = gr.Markdown(label="Stepwise Report (รายงานรอบต่อรอบ)")
 
-        all_ranks_btn = gr.Button("Show All Ranks Report (ดูสรุปทุกอันดับ)")
-        all_ranks_report = gr.Markdown(label="All Ranks Report (การแสดงผลทุกอันดับ)")
-        all_ranks_btn.click(
-            fn=on_all_ranks_report,
+        step_btn.click(
+            fn=on_stepwise_report,
             inputs=[file, selected_wards, capacity_df, name_num, id_num,
                     med_num, med1_num, med2_num, surg_num, ped_num, comm_num, psy_num, obs_num],
-            outputs=all_ranks_report
+            outputs=step_report
         )
 
-    with gr.Row():
-        clean_btn = gr.Button("Clean data (ดูพรีวิว)", variant="primary")
-
     preview = gr.Dataframe(label="Cleaned preview (first 30 rows)", visible=True)
     cleaned_file = gr.File(label="Download cleaned.csv")
 
@@ -590,7 +530,7 @@ with gr.Blocks(title=APP_TITLE) as demo:
     assigned_file = gr.File(label="Download assigned.csv")
     not_assigned_file = gr.File(label="Download not_assigned.csv")
     leftover_text = gr.Textbox(label="Remaining capacity summary", interactive=False)
-    allocation_report = gr.Markdown(label="All Ranks & Allocation Report")
+    allocation_report = gr.Markdown(label="Stepwise Report (ผลการสุ่มรอบต่อรอบ)")
 
     assign_btn.click(
         fn=on_assign,