Update app.py

app.py

@@ -1,72 +1,135 @@
-from sklearn.feature_extraction.text import TfidfVectorizer
-from sklearn.metrics.pairwise import cosine_similarity
+import fitz
+from io import BytesIO
 import streamlit as st
-import nltk
-from nltk.tokenize import sent_tokenize
-
-# Download NLTK tokenizer data
-nltk.download("punkt", quiet=True)
-
-# Helper Function: Calculate ATS Score for Multiple Resumes
-def calculateATSscores(resumes, job_description):
-    vectorizer = TfidfVectorizer(stop_words="english")
-    all_text = [job_description] + resumes
-    tfidf_matrix = vectorizer.fit_transform(all_text)
-    similarity_scores = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:]).flatten()
-    return similarity_scores
-
-# Helper Function: Summarize Resume
-def summarizeResume(resume, job_description, n_sentences=3):
-    sentences = sent_tokenize(resume)
-    all_text = [job_description] + sentences
-    vectorizer = TfidfVectorizer(stop_words="english")
-    tfidf_matrix = vectorizer.fit_transform(all_text)
-    similarity_scores = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:]).flatten()
-    top_indices = similarity_scores.argsort()[-n_sentences:][::-1]
-    top_sentences = [sentences[i] for i in top_indices]
-    summary = " ".join(top_sentences)
-    return summary
-
-# Streamlit App
-st.title("AI-Powered ATS Screening Tool")
-
-# Step 1: Upload Multiple Resumes
-st.header("Step 1: Upload Resumes")
-uploaded_files = st.file_uploader("Upload your resumes (PDFs only):", type="pdf", accept_multiple_files=True)
-
-resumes = []
-if uploaded_files:
-    from convert import ExtractPDFText  # Assuming this module extracts text from PDFs
-    for file in uploaded_files:
-        extracted_text = ExtractPDFText(file)
-        resumes.append(extracted_text)
-    st.success(f"{len(resumes)} resumes uploaded successfully.")
-
-# Step 2: Input Job Description
-st.header("Step 2: Input Job Description")
-job_description = st.text_area("Paste the job description below:")
-st.info("You can copy-paste directly from the job portal.")
-
-# Step 3: Analyze and Display Results
-if resumes and job_description and st.button("Analyze Resumes"):
-    with st.spinner("Analyzing resumes..."):
-        # Calculate similarity scores
-        ats_scores = calculateATSscores(resumes, job_description)
+
+from sklearn.feature_extraction.text import TfidfVectorizer 
+from sklearn.metrics.pairwise import cosine_similarity 
+from sklearn.feature_extraction import _stop_words 
+from convert import ExtractPDFText
+
+def ExtractPDFText(pdf):
+    content = ""
+    pdf_bytes = pdf.read()
+
+    try:
+        pdf_document = fitz.open("dummy.pdf", pdf_bytes)
         
-        # Rank resumes by similarity
-        ranked_resumes = sorted(zip(resumes, ats_scores), key=lambda x: x[1], reverse=True)
+        # Iterate through pages and extract text
+        for page_number in range(pdf_document.page_count):
+            page = pdf_document[page_number]
+            text = page.get_text()
+            content += text
         
-        st.subheader("Results: Resumes Ranked by Similarity")
+    except Exception as e:
+        st.error(f"Error extracting text from PDF: {e}")
         
-        for rank, (resume_text, score) in enumerate(ranked_resumes, start=1):
-            st.markdown(f"### Rank {rank}: Similarity Score = {score:.2f}")
-            
-            # Summarize each resume
-            summary = summarizeResume(resume_text, job_description)
-            
-            st.write("*Summary of Resume:*")
-            st.write(summary)
-            st.write("---")
+    finally:
+        if "pdf_document" in locals():
+            pdf_document.close()
+
+    return content
+
+#find ats score 
+def calculateATSscore(resume_data, job_description):
+    stopwords = list(_stop_words.ENGLISH_STOP_WORDS) 
+    vectorizer = TfidfVectorizer(stop_words=stopwords) 
+    vectors = vectorizer.fit_transform([job_description, resume_data])
+    similarity_value = cosine_similarity(vectors)
+    print(similarity_value)
+    # return similarity_value[0,1]
+    return similarity_value[0,1]
+
+
+import google.generativeai as genai 
+import os
+from dotenv import load_dotenv
+load_dotenv()
+
+genai.configure(api_key = os.environ["GOOGLE_API_KEY"])
+
+model = genai.GenerativeModel("gemini-pro") 
+
+def modelFeedback(ats_score,resume_data):
+
+    input_prompt = f"""
+        You are now an ATS Score analyzer and given ATS Score is {int(ats_score*100)}%. 
+        Your task is to provide feedback to the user based on the ATS score.
+        print ATS score first. mention where resume is good and where resume lacks. 
+        talk about each section of user's resume and talk good and bad points of it. 
+    """
+    response = model.generate_content([input_prompt,resume_data],stream=True)
+    response.resolve()
+
+    return response
+
+
+# Import necessary libraries
+import time
+
+
+if "page_number" not in st.session_state:
+    st.session_state.page_number = 1
+
+if "resume_data" not in st.session_state:
+    st.session_state.resume_data = ""
+
+if "jobdescription" not in st.session_state:
+    st.session_state.jobdescription = ""
+    
+def set_page_number_and_reset_data():
+    st.session_state.page_number = 1
+    st.session_state.resume_data = ""
+
+
+def page1():
+    st.title("AI-Powered ATS Screening")
+    if not st.session_state.resume_data:
+        pdf = st.file_uploader(label="Upload your resume", type="pdf")
+        st.write("No Resume Yet? Create one [here](https://www.overleaf.com/latex/templates/tagged/cv)")
+
+        if pdf:
+            st.success("Resume uploaded successfully.")
+            st.session_state.resume_data = ExtractPDFText(pdf)
+
+def page2():
+    st.title("AI-Powered ATS Screening: Job Description")
+    st.session_state.jobdescription = st.text_area("Job Description: ")
+    st.info("You can just copy paste from the job portal")
+    submit = st.button("Submit")
+
+    if submit:
+        start()
+
+def page3():
+    st.title("Your Resume data: ")
+    if st.session_state.resume_data:
+        st.write(st.session_state.resume_data)
+    else:
+        st.error("Please upload your resume to view the extracted data")
+
+def start():
+    if st.session_state.resume_data and st.session_state.jobdescription:
+        with st.spinner("Hold on, we're calculating your ATS Score..."):
+            ATS_score = calculateATSscore(st.session_state.resume_data, st.session_state.jobdescription)
+            model_feedback = modelFeedback(ATS_score, st.session_state.resume_data)
+            # time.sleep(5)
+
+
+        st.subheader("AI FEEDBACK:")
+        st.write(model_feedback.text)
+        
+    else:
+        st.info("Please, upload Resume and Provide the Job Description")
+
+if st.session_state.page_number == 1:
+    page1()
+elif st.session_state.page_number == 2:
+    page2()
+elif st.session_state.page_number == 3:
+    page3()
 
+if st.session_state.page_number == 1:
+    st.button("View your Extracted Resume data", on_click = lambda: setattr(st.session_state,"page_number", 3))
+    st.button("Go to Job Description Page", on_click=lambda: setattr(st.session_state, "page_number", 2))
 else:
-    st.warning("Please upload resumes and input a job description to proceed.")
+    st.button("Go to PDF Upload Page", on_click=lambda: set_page_number_and_reset_data())