Update app.py

app.py

@@ -7,12 +7,10 @@ from langchain_core.output_parsers import StrOutputParser
 from sklearn.metrics.pairwise import cosine_similarity
 from dotenv import load_dotenv
 from fpdf import FPDF
-from wordcloud import WordCloud
-import matplotlib.pyplot as plt
+from collections import Counter
 import numpy as np
 import tempfile
 import os
-import io
 
 load_dotenv()
 groq_api_key = os.getenv("GROQ_API_KEY")
@@ -33,7 +31,7 @@ def extract_skills(text):
 def generate_learning_resources(missing_skills):
     suggestions = []
     for skill in missing_skills:
-        suggestions.append(f"For {skill}: Search for '{skill} tutorial' on YouTube or Coursera.")
+        suggestions.append(f"For {skill}: Search for '{skill} tutorial' on YouTube, Coursera, or Udemy.")
     return "\n".join(suggestions)
 
 def generate_skill_gap_report(user_skills, job_skills, missing_skills, match_percent):
@@ -60,75 +58,62 @@ def generate_skill_gap_report(user_skills, job_skills, missing_skills, match_per
 
     return report
 
-def create_pdf(report_text):
+def create_pdf(full_report_text):
     pdf = FPDF()
     pdf.add_page()
     pdf.set_font("Arial", size=12)
-    pdf.multi_cell(0, 10, report_text)
-    output_path = "skill_gap_report.pdf"
+    pdf.multi_cell(0, 10, full_report_text)
+    output_path = "multi_jd_skill_gap_report.pdf"
     pdf.output(output_path)
     return output_path
 
-from PIL import Image
-
-def generate_wordcloud(missing_skills):
-    if not missing_skills:
-        return None
-    text = " ".join(missing_skills)
-    wordcloud = WordCloud(width=600, height=400, background_color='black', colormap='Set3').generate(text)
-    buffer = io.BytesIO()
-    plt.figure(figsize=(6, 4))
-    plt.imshow(wordcloud, interpolation='bilinear')
-    plt.axis('off')
-    plt.tight_layout()
-    plt.savefig(buffer, format='png')
-    plt.close()
-    buffer.seek(0)
-    return Image.open(buffer)
-
-
 def process_skill_gap(resume_pdf, jd_pdfs):
     if resume_pdf is None or jd_pdfs is None:
-        return 0, "❌ Please upload Resume and at least one Job Description PDF.", "", "", "", None, None, ""
+        return 0, "❌ Please upload Resume and Job Descriptions.", "", "", "", None, "", ""
 
     resume_text = extract_text_from_pdf(resume_pdf)
     user_skills = extract_skills(resume_text)
 
-    best_match_percent = 0
-    best_missing_skills = []
-    best_job_skills = []
-    best_similarity = 0
-    best_report = ""
+    all_missing_skills = []
+    full_report = ""
 
-    for jd_pdf in jd_pdfs:
+    for idx, jd_pdf in enumerate(jd_pdfs, start=1):
         jd_text = extract_text_from_pdf(jd_pdf)
         job_skills = extract_skills(jd_text)
 
         common = set(user_skills) & set(job_skills)
         match_percent = (len(common) / len(job_skills)) * 100 if job_skills else 0
         missing_skills = list(set(job_skills) - set(user_skills))
+        all_missing_skills.extend(missing_skills)
 
         embed_model = HuggingFaceEmbeddings(model_name="all-MiniLM-L6-v2")
         vectors = embed_model.embed_documents([resume_text, jd_text])
         similarity_score = cosine_similarity([vectors[0]], [vectors[1]])[0][0]
         similarity_percent = round(similarity_score * 100, 2)
 
-        if match_percent > best_match_percent:
-            best_match_percent = match_percent
-            best_missing_skills = missing_skills
-            best_job_skills = job_skills
-            best_similarity = similarity_percent
-            best_report = generate_skill_gap_report(user_skills, job_skills, missing_skills, match_percent)
+        ai_report = generate_skill_gap_report(user_skills, job_skills, missing_skills, match_percent)
+
+        full_report += f"\n🔎 JD {idx}:\n"
+        full_report += f"✅ Skill Match: {match_percent}%\n"
+        full_report += f"❌ Missing Skills: {', '.join(missing_skills) if missing_skills else 'None'}\n"
+        full_report += f"🔗 Similarity Score: {similarity_percent}%\n"
+        full_report += f"📋 AI Report:\n{ai_report}\n"
+        full_report += "-------------------------\n"
+
+    resources = generate_learning_resources(list(set(all_missing_skills)))
+    pdf_path = create_pdf(full_report)
 
-    resources = generate_learning_resources(best_missing_skills)
-    pdf_path = create_pdf(best_report)
-    wordcloud_img = generate_wordcloud(best_missing_skills) if best_missing_skills else None
+    most_common_skills = Counter(all_missing_skills).most_common(3)
+    top_missing_skills_text = "Top Missing Skills Across JDs: " + ", ".join(
+        [f"{skill} ({count} times)" for skill, count in most_common_skills]
+    ) if most_common_skills else "No missing skills detected."
 
-    skill_text = f"✅ Skill Match: {best_match_percent}%"
-    missing_text = f"❌ Missing Skills: {', '.join(best_missing_skills) if best_missing_skills else 'None'}"
-    similarity_text = f"🔎 Best JD Similarity Score: {best_similarity}%"
+    overall_match = round(
+        (sum([len(set(user_skills) & set(extract_skills(extract_text_from_pdf(jd)))) for jd in jd_pdfs]) / (len(user_skills) * len(jd_pdfs))) * 100,
+        2
+    ) if user_skills else 0
 
-    return best_match_percent, skill_text, missing_text, similarity_text, best_report, pdf_path, wordcloud_img, resources
+    return overall_match, "✅ Analysis done across all JDs", ", ".join(set(all_missing_skills)), "Multi-JD Comparison Completed", full_report, pdf_path, top_missing_skills_text, resources
 
 with gr.Blocks() as demo:
     gr.HTML("""
@@ -144,28 +129,31 @@ with gr.Blocks() as demo:
     </style>
     """)
 
-    gr.Markdown("# 🧠 AI Skill Gap Checker - Advanced Version")
-    gr.Markdown("Upload your **Resume PDF** and **Multiple Job Description PDFs**. The AI will compare and suggest learning resources.")
+    gr.Markdown("# 🧠 AI Skill Gap Checker - Multi-JD Comparison (No Word Cloud)")
+    gr.Markdown("Upload your **Resume PDF** and **Multiple Job Descriptions (PDFs)**. AI will compare and generate detailed reports with skill gap and recommendations.")
 
     with gr.Row():
         resume_file = gr.File(label="📄 Upload Resume (PDF)", type="binary")
-        jd_files = gr.File(label="📄 Upload Job Description PDFs (Multiple Allowed)", type="binary", file_types=[".pdf"], file_count="multiple")
+        jd_files = gr.File(label="📄 Upload Multiple Job Descriptions (PDFs)", type="binary", file_types=[".pdf"], file_count="multiple")
 
     submit_btn = gr.Button("🚀 Analyze Skill Gap")
 
-    match_slider = gr.Slider(minimum=0, maximum=100, step=1, label="Job Match Percentage (%)", interactive=False)
-    skill_match_text = gr.Textbox(label="Skill Match Details", interactive=False)
-    missing_skills_text = gr.Textbox(label="Missing Skills", interactive=False)
-    similarity_text = gr.Textbox(label="Best JD Similarity Score", interactive=False)
-    report_output = gr.Textbox(label="AI-Generated Skill Gap Report", lines=12, interactive=False)
-    download_pdf = gr.File(label="📥 Download Skill Gap Report (PDF)")
-    wordcloud_output = gr.Image(label="📊 Missing Skills Word Cloud")
+    match_slider = gr.Slider(minimum=0, maximum=100, step=1, label="Overall Skill Match (%)", interactive=False)
+    skill_match_text = gr.Textbox(label="Status", interactive=False)
+    missing_skills_text = gr.Textbox(label="All Missing Skills", interactive=False)
+    similarity_text = gr.Textbox(label="Status Message", interactive=False)
+    report_output = gr.Textbox(label="AI-Generated Multi-JD Skill Gap Report", lines=20, interactive=False)
+    download_pdf = gr.File(label="📥 Download Full Report as PDF")
+    top_skills_output = gr.Textbox(label="Top Missing Skills Across JDs", interactive=False)
     learning_resources = gr.Textbox(label="📚 AI Learning Resource Recommendations", interactive=False)
 
     submit_btn.click(
         fn=process_skill_gap,
         inputs=[resume_file, jd_files],
-        outputs=[match_slider, skill_match_text, missing_skills_text, similarity_text, report_output, download_pdf, wordcloud_output, learning_resources]
+        outputs=[
+            match_slider, skill_match_text, missing_skills_text, similarity_text,
+            report_output, download_pdf, top_skills_output, learning_resources
+        ]
     )
 
 demo.launch()