feat: update hostel and backlog handling in forms and backend logic

index.html

@@ -187,7 +187,7 @@
                                     </span>
                                     <span class="checkbox-text">
                                         <span class="checkbox-title">Hostel Accommodation</span>
-                                        <span class="checkbox-desc">Currently living in a hostel</span>
+                                        <span class="checkbox-desc">Checked = Yes (1), Unchecked = No (0)</span>
                                     </span>
                                 </label>
                             </div>
@@ -205,7 +205,7 @@
                                     </span>
                                     <span class="checkbox-text">
                                         <span class="checkbox-title">Academic Backlogs</span>
-                                        <span class="checkbox-desc">Any history of backlogs</span>
+                                        <span class="checkbox-desc">Checked = Yes (1), Unchecked = No (0)</span>
                                     </span>
                                 </label>
                             </div>

llm.py

@@ -38,6 +38,20 @@ _chat_db = SqliteDb(db_file="tmp/placement_chat.db")
 _sessions: dict = {}
 
 
+def _binary_to_yes_no(value) -> str:
+    """Map binary-encoded values to dataset semantics: 0 = No, 1 = Yes."""
+    if isinstance(value, str):
+        normalized = value.strip().lower()
+        if normalized in {"1", "true", "yes"}:
+            return "Yes"
+        if normalized in {"0", "false", "no", ""}:
+            return "No"
+    try:
+        return "Yes" if int(value) == 1 else "No"
+    except Exception:
+        return "No"
+
+
 def build_system_context(
     student_data: dict,
     prediction: dict,
@@ -59,8 +73,8 @@ def build_system_context(
         f"Stream: {student_data.get('Stream', 'N/A')}",
         f"Internships: {student_data.get('Internships', 'N/A')}",
         f"CGPA: {student_data.get('CGPA', 'N/A')}",
-        f"Hostel: {'Yes' if student_data.get('Hostel') else 'No'}",
-        f"History of Backlogs: {'Yes' if student_data.get('HistoryOfBacklogs') else 'No'}",
+        f"Hostel: {_binary_to_yes_no(student_data.get('Hostel'))}",
+        f"History of Backlogs: {_binary_to_yes_no(student_data.get('HistoryOfBacklogs'))}",
         f"Skills: {', '.join(student_data.get('skills', []))}",
         f"Desired Role: {student_data.get('desired_role', 'Not Specified')}"
     ]

main.py

@@ -157,8 +157,8 @@ class StudentData(BaseModel):
     Stream: str = Field(..., description="Stream of study")
     Internships: int = Field(..., ge=0)
     CGPA: float = Field(..., ge=0, le=10)
-    Hostel: int = Field(..., description="1 if hostel, 0 if not")
-    HistoryOfBacklogs: int = Field(..., description="1 if backlogs, 0 if not")
+    Hostel: int = Field(..., ge=0, le=1, description="0 = day scholar (not in hostel), 1 = lives in hostel")
+    HistoryOfBacklogs: int = Field(..., ge=0, le=1, description="0 = no history of backlogs, 1 = history of backlogs")
     skills: List[str] = Field(default=[], description="List of user skills for routing")
     desired_role: Optional[str] = Field(default=None, description="User's desired job role")
 
@@ -191,6 +191,16 @@ class OptionsResponse(BaseModel):
     skills: List[str]
     jobs: List[str]
 
+
+def normalize_backlogs_value(value: int) -> int:
+    """Normalize backlog input to strict dataset encoding: 0 = No, 1 = Yes."""
+    return 1 if int(value) == 1 else 0
+
+
+def normalize_hostel_value(value: int) -> int:
+    """Normalize hostel input to strict dataset encoding: 0 = No, 1 = Yes."""
+    return 1 if int(value) == 1 else 0
+
 def prepare_input(data: StudentData) -> pd.DataFrame:
     try:
         g = le_gender.transform([data.Gender])[0]
@@ -205,8 +215,8 @@ def prepare_input(data: StudentData) -> pd.DataFrame:
         'Stream': s,
         'Internships': data.Internships,
         'CGPA': data.CGPA,
-        'Hostel': data.Hostel,
-        'HistoryOfBacklogs': data.HistoryOfBacklogs
+        'Hostel': normalize_hostel_value(data.Hostel),
+        'HistoryOfBacklogs': normalize_backlogs_value(data.HistoryOfBacklogs)
     }])
     # Extract columns naturally based on dict order which matches train features order
     return df
@@ -359,6 +369,10 @@ async def whatif_analysis(data: StudentData):
         return float(model.predict_proba(df_processed)[0][1]) * 100
 
     orig_dict = data.model_dump()
+    orig_dict['Hostel'] = normalize_hostel_value(orig_dict.get('Hostel', 0))
+    orig_dict['HistoryOfBacklogs'] = normalize_backlogs_value(orig_dict.get('HistoryOfBacklogs', 0))
+    current_hostel = orig_dict['Hostel']
+    current_backlogs = orig_dict['HistoryOfBacklogs']
     orig_risk = _predict(orig_dict)
     orig_level, _ = get_placement_level(orig_risk / 100)
     
@@ -421,7 +435,7 @@ async def whatif_analysis(data: StudentData):
         )
         
     # Clear Backlogs
-    if data.HistoryOfBacklogs == 1:
+    if current_backlogs == 1:
         mod = orig_dict.copy()
         mod['HistoryOfBacklogs'] = 0
         add_scenario(
@@ -443,7 +457,7 @@ async def whatif_analysis(data: StudentData):
             )
             
     # Stay in Hostel
-    if data.Hostel == 0:
+    if current_hostel == 0:
         mod = orig_dict.copy()
         mod['Hostel'] = 1
         mod_risk = _predict(mod)
@@ -460,7 +474,7 @@ async def whatif_analysis(data: StudentData):
     best_case = orig_dict.copy()
     if data.CGPA < 9.0: best_case['CGPA'] = min(data.CGPA + 1.0, 10.0)
     if data.Internships < 3: best_case['Internships'] += 1
-    if data.HistoryOfBacklogs == 1: best_case['HistoryOfBacklogs'] = 0
+    if current_backlogs == 1: best_case['HistoryOfBacklogs'] = 0
     
     combined_risk = _predict(best_case)
     combined_level, _ = get_placement_level(combined_risk / 100)

script.js

@@ -44,6 +44,18 @@ let currentFormData = null;
 let availableSkills = [];
 let selectedSkills = [];
 
+function encodeBacklogsFromCheckbox(checkboxId = 'backlogs') {
+    return document.getElementById(checkboxId).checked ? 1 : 0;
+}
+
+function decodeBacklogsToChecked(value) {
+    return Number(value) === 1;
+}
+
+function encodeHostelFromCheckbox(checkboxId = 'hostel') {
+    return document.getElementById(checkboxId).checked ? 1 : 0;
+}
+
 /**
  * Initialize the application
  */
@@ -270,8 +282,8 @@ function collectFormData() {
         Stream: document.getElementById('stream').value,
         Internships: parseInt(document.getElementById('internships').value) || 0,
         CGPA: parseFloat(document.getElementById('cgpa').value) || 0,
-        Hostel: document.getElementById('hostel').checked ? 1 : 0,
-        HistoryOfBacklogs: document.getElementById('backlogs').checked ? 1 : 0,
+        Hostel: encodeHostelFromCheckbox('hostel'),
+        HistoryOfBacklogs: encodeBacklogsFromCheckbox('backlogs'),
         skills: document.getElementById('skills').value.split(',').map(s => s.trim()).filter(s => s),
         desired_role: document.getElementById('desired_role').value || null
     };
@@ -950,7 +962,7 @@ function initSimulator() {
 
     // Toggle Setup
     const toggle = document.getElementById('simBacklogToggle');
-    toggle.checked = (currentFormData.HistoryOfBacklogs === 1);
+    toggle.checked = decodeBacklogsToChecked(currentFormData.HistoryOfBacklogs);
     toggle.onchange = () => { triggerSimulate(); };
 }
 
@@ -990,8 +1002,7 @@ async function simulateRisk() {
     simulatedData.Internships = parseInt(document.getElementById('simInternNum').value);
     simulatedData.CGPA = parseFloat(document.getElementById('simCGPANum').value);
 
-    const hasBacklog = document.getElementById('simBacklogToggle').checked;
-    simulatedData.HistoryOfBacklogs = hasBacklog ? 1 : 0;
+    simulatedData.HistoryOfBacklogs = encodeBacklogsFromCheckbox('simBacklogToggle');
 
     try {
         const response = await fetchPrediction(simulatedData);