import re
import spacy
from sklearn.feature_extraction.text import TfidfVectorizer

nlp = spacy.load('en_core_web_sm')

def clean_text(text):
    # Remove special characters
    text = re.sub(r'[^a-zA-Z\s]', '', text)
    text = text.lower()
    return text

def extract_keywords(text):
    doc = nlp(text)
    # Extract nouns and proper nouns
    keywords = [token.text for token in doc if token.pos_ in ['NOUN', 'PROPN']]
    return keywords

# Example
resume_text = "Data analyst with 5 years of experience in Python, SQL, and Tableau..."
cleaned = clean_text(resume_text)
keywords = extract_keywords(cleaned)
print("Keywords:", keywords[:10])

# TF-IDF keyword extraction
def extract_keywords_tfidf(texts):
    vectorizer = TfidfVectorizer(max_features=50, stop_words='english')
    tfidf_matrix = vectorizer.fit_transform(texts)
    feature_names = vectorizer.get_feature_names_out()
    return feature_names

# Compare resume to job description
def match_resume(resume_text, job_description):
    # Clean both texts
    clean_resume = clean_text(resume_text)
    clean_job = clean_text(job_description)

    # Extract keywords
    resume_keywords = set(extract_keywords(clean_resume))
    job_keywords = set(extract_keywords(clean_job))

    # Calculate match score
    match_score = len(resume_keywords.intersection(job_keywords)) / len(job_keywords) * 100
    return match_score

from sklearn.metrics.pairwise import cosine_similarity
from sklearn.feature_extraction.text import TfidfVectorizer

def score_resume(resume_text, job_description):
    # Create TF-IDF vectors
    vectorizer = TfidfVectorizer(stop_words='english')
    docs = [resume_text, job_description]
    tfidf_matrix = vectorizer.fit_transform(docs)

    # Calculate cosine similarity
    similarity = cosine_similarity(tfidf_matrix[0:1], tfidf_matrix[1:2])[0][0]

    # Also check keyword overlap
    resume_keywords = set(extract_keywords(resume_text))
    job_keywords = set(extract_keywords(job_description))

    overlap_score = len(resume_keywords.intersection(job_keywords)) / len(job_keywords)

    # Combined score (weighted)
    final_score = 0.6 * similarity + 0.4 * overlap_score
    return final_score * 100

# Rank multiple resumes
def rank_resumes(resumes, job_description):
    scores = []
    for resume in resumes:
        score = score_resume(resume, job_description)
        scores.append(score)

    # Sort resumes by score
    ranked = sorted(zip(resumes, scores), key=lambda x: x[1], reverse=True)
    return ranked

# Example usage
resume_list = [resume1_text, resume2_text, resume3_text]
ranked_resumes = rank_resumes(resume_list, job_description)

for i, (resume, score) in enumerate(ranked_resumes):
    print(f"Rank {i+1}: Score {score:.2f}%")

import streamlit as st

st.title("AI Resume Analyzer")

# Job description input
job_description = st.text_area("Paste Job Description", height=150)

# Resume upload
uploaded_file = st.file_uploader("Upload Resume (PDF or TXT)", type=["txt", "pdf"])

if job_description and uploaded_file:
    # Read resume text
    resume_text = uploaded_file.read().decode("utf-8")

    # Score resume
    score = score_resume(resume_text, job_description)

    # Display results
    st.subheader("Resume Match Score")
    st.metric("Match Score", f"{score:.2f}%")

    # Show keywords
    st.subheader("Keywords Found")
    keywords = extract_keywords(resume_text)
    st.write(keywords[:20])

    # Recommendations
    st.subheader("Recommendations")
    st.write("Add more keywords from the job description to improve your score.")

# Deploy on Streamlit Cloud
# 1. Create requirements.txt
# 2. Push to GitHub
# 3. Deploy on streamlit.io

# requirements.txt
# streamlit
# spacy
# scikit-learn
# pandas

# Run locally:
# streamlit run app.py