Update app.py

app.py

@@ -1,64 +1,102 @@
 import streamlit as st
-from sentence_transformers import SentenceTransformer, util
-import PyPDF2
-from docx import Document
+from transformers import AutoTokenizer, AutoModelForSequenceClassification
+from fpdf import FPDF
+from utils import extract_text, render_template
 import os
+from sentence_transformers import SentenceTransformer
+import pdfplumber
+import PyPDF2
+import docx
+from io import StringIO
 
-# Load model
-model = SentenceTransformer('sentence-transformers/all-MiniLM-L6-v2')
 
-# Helper functions
-def extract_text_from_pdf(file):
-    reader = PyPDF2.PdfReader(file)
-    text = ""
-    for page in reader.pages:
-        text += page.extract_text()
-    return text
+# Load pre-trained model for text classification or matching (you can change this to the relevant task model)
+model_name = 'bert-base-uncased'  # Replace with an appropriate model for your task
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+model = AutoModelForSequenceClassification.from_pretrained(model_name)
 
-def extract_text_from_docx(file):
-    doc = Document(file)
-    text = "\n".join([para.text for para in doc.paragraphs])
-    return text
+# Load SentenceTransformer model for semantic matching
+embedder = SentenceTransformer('all-MiniLM-L6-v2')
 
-def extract_text(file):
-    filename = file.name.lower()
-    if filename.endswith(".pdf"):
-        return extract_text_from_pdf(file)
-    elif filename.endswith(".docx"):
-        return extract_text_from_docx(file)
-    elif filename.endswith(".txt"):
-        return file.read().decode("utf-8")
-    else:
-        return "Unsupported file format."
+# Set up the streamlit interface
+st.title("Resume Tailoring & Job Fit App")
 
-def calculate_similarity(text1, text2):
-    embedding1 = model.encode(text1, convert_to_tensor=True)
-    embedding2 = model.encode(text2, convert_to_tensor=True)
-    score = util.pytorch_cos_sim(embedding1, embedding2)
-    return float(score[0][0]) * 100  # Convert to percentage
+# File upload for Resume (accepting multiple formats)
+resume_file = st.file_uploader("Upload your Resume (PDF, DOCX, TXT, HTML)", type=["pdf", "docx", "txt", "html"])
 
-# Streamlit UI
-st.set_page_config(page_title="Resume Fit Evaluator", layout="centered")
-st.title("🎯 Resume vs Job Description Fit Score")
+# Text area for Job Description or Scholarship details (paste text)
+job_description = st.text_area("Paste the Job Description or Scholarship details", height=200)
 
-with st.sidebar:
-    st.header("Upload Files")
-    resume_file = st.file_uploader("Upload your resume", type=["pdf", "docx", "txt"])
-    jd_file = st.file_uploader("Upload job description", type=["pdf", "docx", "txt"])
+# Button to submit the job description and resume for processing
+submit_button = st.button("Submit")
 
-if resume_file and jd_file:
-    with st.spinner("Extracting text and comparing..."):
-        resume_text = extract_text(resume_file)
-        jd_text = extract_text(jd_file)
+if submit_button:
+    if resume_file is not None and job_description:
+        # Extract text from resume file
+        if resume_file.type == "application/pdf":
+            resume_text = extract_text(resume_file, "pdf")
+        elif resume_file.type == "application/vnd.openxmlformats-officedocument.wordprocessingml.document":
+            resume_text = extract_text(resume_file, "docx")
+        elif resume_file.type == "text/plain":
+            resume_text = extract_text(resume_file, "txt")
+        elif resume_file.type == "text/html":
+            resume_text = extract_text(resume_file, "html")
+        
+        # Analyze the fit between the job description and resume
+        # Here, we use a semantic match between resume and job description
+        job_embedding = embedder.encode(job_description)
+        resume_embedding = embedder.encode(resume_text)
+        
+        # Calculate cosine similarity (fit score)
+        from sklearn.metrics.pairwise import cosine_similarity
+        fit_score = cosine_similarity([job_embedding], [resume_embedding])[0][0]
+        
+        # Display the fit score
+        st.write(f"Fit Score: {fit_score*100:.2f}%")
+        
+        # Extract required skills from job description (simple placeholder method)
+        required_skills = ["Python", "Machine Learning", "Data Analysis"]  # Replace with actual extraction logic
+        missing_skills = [skill for skill in required_skills if skill.lower() not in resume_text.lower()]
+        
+        if missing_skills:
+            st.write("Skills you might need to improve for better fit:")
+            st.write(", ".join(missing_skills))
+        else:
+            st.write("You have all the required skills for this job!")
 
-        similarity = calculate_similarity(resume_text, jd_text)
+        # Resume tailoring based on job description (mock implementation)
+        tailored_resume = resume_text + "\n\n[Tailored to fit job description]\n" + job_description
 
-    st.success(f"✅ Fit Score: **{similarity:.2f}%**")
+        # Template selection for preview
+        template_choice = st.selectbox("Choose a template for preview", ["Template 1", "Template 2", "Template 3"])
+        
+        # Generate resume PDF from tailored content
+        if template_choice == "Template 1":
+            rendered_resume = render_template(tailored_resume, template="template1")
+        elif template_choice == "Template 2":
+            rendered_resume = render_template(tailored_resume, template="template2")
+        else:
+            rendered_resume = render_template(tailored_resume, template="template3")
+        
+        st.text_area("Preview of your tailored resume:", rendered_resume, height=300)
 
-    with st.expander("🔍 View Resume Text"):
-        st.text_area("Resume", resume_text, height=250)
+        # Allow the user to download the generated resume
+        pdf = FPDF()
+        pdf.add_page()
+        pdf.set_auto_page_break(auto=True, margin=15)
+        pdf.set_font("Arial", size=12)
+        pdf.multi_cell(0, 10, rendered_resume)
+        pdf_output = pdf.output(dest='S', name='resume.pdf')
+        
+        st.download_button("Download Tailored Resume", pdf_output, file_name="tailored_resume.pdf")
 
-    with st.expander("📄 View Job Description Text"):
-        st.text_area("Job Description", jd_text, height=250)
-else:
-    st.info("Please upload both a resume and a job description to get started.")
+        # Application form questions (example)
+        st.subheader("Application Form")
+        application_question = st.text_input("Please answer this question:")
+        
+        if application_question:
+            # Human-like response generation using a basic chatbot approach (simple placeholder)
+            response = f"Your response to '{application_question}' could be structured as follows: 'I am excited about this opportunity because...'"
+            st.write(response)
+    else:
+        st.warning("Please upload a resume and paste the job description.")