Update app.py

app.py

@@ -409,68 +409,74 @@ def analyze_google_fit(resume_summary):
     
     # Create a more specific prompt for T5 that focuses on detailed assessment
     prompt = f"""
-Generate Google candidate assessment.
-Skills detected: {skills_text}.
-Experience: {experience_highlights}.
-Strongest area: {top_category} ({categories_sorted[0][1]["adjusted_score"]}%).
-Weakest area: {weak_category} ({categories_sorted[-1][1]["adjusted_score"]}%).
-Overall match: {match_percentage}%.
-
-Write a detailed 3-4 sentence assessment for this Google applicant.
-Start with "This candidate" and include:
-1. Specific strengths relating to Google's needs
-2. Technical skills evaluation
-3. Areas for improvement
-4. Overall Google fit assessment
-
-This candidate
-"""
+        Generate a professional expert assessment for a Google job candidate.
+        Skills detected: {skills_text}.
+        Strongest area: {top_category} ({categories_sorted[0][1]["adjusted_score"]}%).
+        Weakest area: {weak_category} ({categories_sorted[-1][1]["adjusted_score"]}%).
+        Overall match: {match_percentage}%.
+
+        Write an evaluative assessment that analyzes the candidate's fit for Google.
+        Start with "This candidate" and:
+        - Evaluate their technical skills in relation to Google's standards
+        - Assess their strengths and potential contributions
+        - Provide specific areas where improvement is needed
+        - Give an overall evaluation of their Google fit
+
+        Your assessment should be an expert evaluation, not a summary of their resume.
+        Include specific actionable advice for improvement.
+
+        This candidate
+        """
     
-    try:
-        # Generate the assessment using T5
-        assessment_results = models['evaluator'](
-            prompt, 
-            max_length=300,
-            do_sample=True,
-            temperature=0.7,
-            num_return_sequences=3
-        )
+        try:
+            # Generate the assessment using T5
+            assessment_results = models['evaluator'](
+                prompt, 
+                max_length=300,
+                do_sample=True,
+                temperature=0.75,  # Slightly higher temperature for more evaluative content
+                num_return_sequences=3
+            )
+    
+            # Find the best response
+            best_assessment = None
+            for result in assessment_results:
+                text = result['generated_text'].strip()
         
-        # Find the best response
-        best_assessment = None
-        for result in assessment_results:
-            text = result['generated_text'].strip()
-            
-            # Clean up and check if valid
-            text = re.sub(r'Generate Google candidate assessment.*?Overall match:.*?%\.', '', text, flags=re.DOTALL)
-            text = re.sub(r'Write a detailed.*?This candidate', 'This candidate', text, flags=re.DOTALL)
-            
-            # Check if it's a good response
-            if text.lower().startswith("this candidate") and len(text) > 100 and "1." not in text and "2." not in text:
-                best_assessment = text
-                break
+                # Clean up and check if valid
+                text = re.sub(r'Generate a professional expert assessment.*?Overall match:.*?%\.', '', text, flags=re.DOTALL)
+                text = re.sub(r'Write an evaluative assessment.*?This candidate', 'This candidate', text, flags=re.DOTALL)
         
-        # Use the best response or the first one if none were ideal
-        if best_assessment:
-            assessment = best_assessment
-        else:
-            # Use first response but clean it up
-            text = assessment_results[0]['generated_text']
-            text = re.sub(r'Generate Google candidate assessment.*?Overall match:.*?%\.', '', text, flags=re.DOTALL)
-            text = re.sub(r'Write a detailed.*?This candidate', 'This candidate', text, flags=re.DOTALL)
-            
-            # Remove numbering if present
-            text = re.sub(r'\d\.\s', '', text)
-            
-            if not text.lower().startswith("this candidate"):
-                text = "This candidate " + text
-                
-            assessment = text
+                # Check if it's a good response - looking for evaluative language
+                if (text.lower().startswith("this candidate") and 
+                    len(text) > 100 and 
+                    ("would" in text.lower() or "could" in text.lower() or "should" in text.lower() or 
+                     "needs" in text.lower() or "requires" in text.lower() or "lacks" in text.lower())):
+                    best_assessment = text
+                    break
+    
+            # Use the best response or the first one if none were ideal
+            if best_assessment:
+                assessment = best_assessment
+            else:
+                # Use first response but clean it up
+                text = assessment_results[0]['generated_text']
+                text = re.sub(r'Generate a professional expert assessment.*?Overall match:.*?%\.', '', text, flags=re.DOTALL)
+                text = re.sub(r'Write an evaluative assessment.*?This candidate', 'This candidate', text, flags=re.DOTALL)
+        
+                # Remove numbering or bullets if present
+                text = re.sub(r'[-•]\s', '', text)
+                text = re.sub(r'\d\.\s', '', text)
+        
+                if not text.lower().startswith("this candidate"):
+                    text = "This candidate " + text
             
-    except Exception as e:
-        # Fall back to manual assessment
-        print(f"Error in T5 assessment generation: {e}")
-        assessment = f"""This candidate demonstrates some skills relevant to Google, particularly in {top_category}. Their experience with {skills_text} could be valuable, though they would benefit from strengthening their {weak_category}. Based on the resume analysis, they appear to be a {match_percentage}% match for Google's requirements."""
+                assessment = text
+        
+        except Exception as e:
+            # Fall back to manual assessment - making this more evaluative
+            print(f"Error in T5 assessment generation: {e}")
+            assessment = f"""This candidate shows promise in {top_category} but would need significant development in {weak_category} to meet Google's standards. Their technical skills in {skills_text} align with Google's engineering needs, but they would benefit from developing stronger problem-solving capabilities and a more innovative approach to complex challenges. At {match_percentage}%, they could become a more competitive candidate with targeted improvement in these areas."""
     
     # Final cleanup
     # Remove any remaining artifacts or formatting
@@ -669,10 +675,10 @@ if uploaded_file is not None and st.button("Analyze My Google Fit"):
         breakdown_data = []
         for category, details in category_details.items():
             label = {"technical_skills": "Technical Programming Skills",
-             "advanced_tech": "Advanced Technology Knowledge",
-             "problem_solving": "Problem Solving Abilities", 
-             "innovation": "Innovation Mindset",
-             "soft_skills": "Collaboration & Leadership"}[category]
+                     "advanced_tech": "Advanced Technology Knowledge",
+                     "problem_solving": "Problem Solving Abilities", 
+                     "innovation": "Innovation Mindset",
+                     "soft_skills": "Collaboration & Leadership"}[category]
     
             # Create a visual indicator for the score
             score = details["adjusted_score"]
@@ -682,12 +688,12 @@ if uploaded_file is not None and st.button("Analyze My Google Fit"):
                 "Category": label,
                 "Score": f"{score}%",
                 "Matching Skills": ", ".join(details["matching_keywords"][:3]) if details["matching_keywords"] else "None detected"
-    })
+            })
 
         # Convert to DataFrame and display
-        import pandas as pd
         breakdown_df = pd.DataFrame(breakdown_data)
-        st.table(breakdown_df)
+        # Remove the index column entirely
+        st.table(breakdown_df.set_index('Category').reset_index())  # This removes the numerical index
 
         # Show a note about how scores are calculated
         with st.expander("How are these scores calculated?"):