Update app.py

app.py

@@ -1,57 +1,74 @@
 import gradio as gr
-from keybert import KeyBERT
-from sentence_transformers import SentenceTransformer
+from transformers import pipeline
 import re
 
-# ✅ Load Hugging Face model (no API key needed)
-model = SentenceTransformer('all-MiniLM-L6-v2')
-kw_model = KeyBERT(model)
+# ✅ Load Hugging Face summarization/chat model
+summarizer = pipeline("text-generation", model="mistralai/Mistral-7B-Instruct-v0.1")
 
-# 🔍 Helper: Clean keywords from text (split on commas, newlines)
-def clean_keywords(text):
-    # Lowercase, remove special characters, and split on space/comma
-    text = re.sub(r"[^a-zA-Z0-9\s]", " ", text.lower())
-    words = re.split(r"\s+", text)
-    return set(w for w in words if w and len(w) > 2)  # ignore very short words
+# ✅ Highlight matching keywords
+from sklearn.feature_extraction.text import ENGLISH_STOP_WORDS
 
+def highlight_keywords(resume_text, job_desc):
+    resume_words = set(re.findall(r"\b\w{3,}\b", resume_text.lower())) - ENGLISH_STOP_WORDS
+    job_words = set(re.findall(r"\b\w{3,}\b", job_desc.lower())) - ENGLISH_STOP_WORDS
+    matched = resume_words & job_words
+    highlighted = resume_text
+    for word in sorted(matched, key=len, reverse=True):
+        highlighted = re.sub(rf"\\b({re.escape(word)})\\b", r"**\1**", highlighted, flags=re.IGNORECASE)
+    return highlighted
 
-# 🔍 Main function
-def extract_keywords(job_desc, resume_text, analyze_with_jd):
+# 🔍 Prompt for dynamic section classification and feedback
+def build_dynamic_prompt(job_desc, resume_text, analyze_with_jd):
+    prompt = f"""
+You are an expert resume analyst.
+Classify the content of the resume into meaningful categories based on the text (e.g., Technical Skills, Soft Skills, Certifications, Projects, Work Experience, Education, Personal Information).
+
+You do not need a fixed list of section names—choose the most suitable ones based on the content.
+
+Then:
+- For each section, summarize its contents.
+- If a job description is provided, compare each section against it.
+- Highlight missing or weak areas relevant to the job.
+- Provide smart, actionable suggestions to improve each section.
+- Output in structured Markdown with headings for each section.
+"""
+    if analyze_with_jd and job_desc:
+        prompt += f"\nJob Description:\n{job_desc}\n"
+    prompt += f"\nResume:\n{resume_text}"
+    return prompt
+
+# 🧠 Function to call Hugging Face model and get structured resume feedback
+def analyze_resume(job_desc, resume_text, analyze_with_jd):
     if not resume_text.strip():
         return "⚠️ Please paste your resume text."
 
-    # Step 1: Combine input for keyword extraction
-    combined_text = job_desc + "\n\n" + resume_text if analyze_with_jd and job_desc else resume_text
-    extracted_keywords = kw_model.extract_keywords(combined_text, top_n=15, stop_words='english')
-    extracted_set = set([kw.lower() for kw, _ in extracted_keywords])
+    user_prompt = build_dynamic_prompt(job_desc, resume_text, analyze_with_jd)
 
-    # Step 2: Tokenize job description and resume separately
-    jd_tokens = clean_keywords(job_desc) if analyze_with_jd and job_desc else set()
-    resume_tokens = clean_keywords(resume_text)
+    try:
+        response = summarizer(user_prompt, max_new_tokens=768, do_sample=True, temperature=0.7)[0]['generated_text']
+        cleaned = re.sub(rf".*?{re.escape(resume_text)}", "", response, flags=re.DOTALL).strip()
+        if analyze_with_jd and job_desc:
+            highlighted_resume = highlight_keywords(resume_text, job_desc)
+            return f"### 🔍 Resume with Highlighted Matches\n\n{highlighted_resume}\n\n---\n{cleaned}"
+        return cleaned
+    except Exception as e:
+        return f"❌ Error: {str(e)}"
 
-    # Step 3: Match and miss
-    matched = sorted(jd_tokens & resume_tokens)
-    missing = sorted(jd_tokens - resume_tokens)
+# 🎛️ Gradio UI
+def create_ui():
+    with gr.Blocks() as demo:
+        with gr.Row():
+            with gr.Column():
+                analyze_checkbox = gr.Checkbox(label="Analyze with Job Description", value=True)
+                job_desc = gr.Textbox(label="Job Description", lines=6, placeholder="Paste job description here...")
+                resume_text = gr.Textbox(label="Resume Text", lines=16, placeholder="Paste resume content here...")
+            with gr.Column():
+                output_analysis = gr.Markdown(label="Resume Analysis and Suggestions")
 
-    # Step 4: Output
-    result = ""
-    result += f"🔍 **Extracted Keywords:** {', '.join(extracted_set)}\n\n"
-    result += f"✅ **Matched (Job & Resume):** {', '.join(matched) or 'None'}\n"
-    result += f"❌ **Missing in Resume:** {', '.join(missing) or 'None'}\n"
+        resume_text.change(fn=analyze_resume, inputs=[job_desc, resume_text, analyze_checkbox], outputs=output_analysis)
+        job_desc.change(fn=analyze_resume, inputs=[job_desc, resume_text, analyze_checkbox], outputs=output_analysis)
 
-    return result
+    return demo
 
-# 🎛️ Gradio UI
-with gr.Blocks() as demo:
-    with gr.Row():
-        with gr.Column():
-            analyze_checkbox = gr.Checkbox(label="Analyze with Job Description", value=True)
-            job_desc = gr.Textbox(label="Job Description", lines=6, placeholder="Paste job description here...")
-            resume_text = gr.Textbox(label="Resume Text", lines=12, placeholder="Paste resume content here...")
-        with gr.Column():
-            output_keywords = gr.Markdown(label="Keyword Match Result")  # Markdown for styled output
-
-    resume_text.change(fn=extract_keywords, inputs=[job_desc, resume_text, analyze_checkbox], outputs=output_keywords)
-    job_desc.change(fn=extract_keywords, inputs=[job_desc, resume_text, analyze_checkbox], outputs=output_keywords)
-
-demo.launch()
+if __name__ == '__main__':
+    create_ui().launch()