Update app.py

app.py

@@ -8,11 +8,9 @@ import docx
 
 db = GraphDB()
 
-# Radio button options
 EDUCATION_OPTIONS = ["High School", "Undergraduate", "Graduate"]
 STATUS_OPTIONS = ["Student", "Employed", "Continuous Learning"]
 
-# ---------------- Helper: read PDF or DOCX ----------------
 def read_file(file):
     if file.name.endswith(".pdf"):
         reader = PdfReader(file.name)
@@ -22,41 +20,35 @@ def read_file(file):
         return "\n".join(p.text for p in doc.paragraphs)
     return ""
 
-# ---------------- Core pipeline ----------------
 def run_pipeline(education, status, role, file):
     user_id = str(uuid.uuid4())
-
-    # Store user profile
     db.create_or_update_user(user_id, education, status, role)
     db.log_interaction(user_id, type_="PROFILE_CREATED", details=role, role=role)
 
-    # Read resume text
     text = read_file(file)
+    extracted = extract_skills(text, threshold=0.50)
 
-    # Extract skills
-    skills = extract_skills(text, threshold=0.50)
-
-    # Store skills in Neo4j
-    for skill, conf in skills:
+    # Store extracted skills + evidence trail
+    for skill, conf, evidence in extracted:
         db.add_skill(user_id, skill, conf)
-        db.log_interaction(user_id, type_="SKILL_ADDED", skill=skill)
-
-    # Recommendations
-    recommendations = generate_recommendations(db, user_id)
+        db.log_interaction(user_id, type_="SKILL_ADDED", skill=skill, details=evidence)
 
-    # Normalize recommendation display
-    recommended_skills_df = []
-    for r in recommendations:
-        if isinstance(r, tuple):
-            recommended_skills_df.append([r[0], r[1]])
-        else:
-            recommended_skills_df.append([r, ""])
+    # Format extracted skills into collapsible HTML
+    extracted_html = ""
+    for skill, conf, evidence in extracted:
+        extracted_html += f"""
+        <details>
+            <summary><b>{skill}</b> — confidence {conf}</summary>
+            <p style="margin-left:10px;">Evidence: {evidence}</p>
+        </details>
+        """
 
-    extracted_skills_df = [[skill, conf] for skill, conf in skills]
+    # Generate recommendations (no evidence)
+    recommendations = generate_recommendations(db, user_id)
+    recommended_df = [[r[0], r[1]] for r in recommendations]
 
-    return user_id, extracted_skills_df, recommended_skills_df
+    return user_id, extracted_html, recommended_df
 
-# ---------------- Gradio UI ----------------
 interface = gr.Interface(
     fn=run_pipeline,
     inputs=[
@@ -67,12 +59,13 @@ interface = gr.Interface(
     ],
     outputs=[
         gr.Textbox(label="User ID"),
-        gr.Dataframe(headers=["Skill", "Confidence"], label="Extracted Skills"),
+        gr.HTML(label="Extracted Skills + Evidence Trails"),
         gr.Dataframe(headers=["Skill", "Confidence"], label="Recommended Skills")
     ],
     title="Skill Grapher",
-    description="Upload your CV to extract your skill profile and receive personalized skill recommendations."
+    description="Upload your CV to extract skills with evidence trails and receive personalized skill recommendations."
 )
 
 interface.launch()
 
+