src/streamlit_app.py

# streamlit_app.py (moved from app.py)

import streamlit as st
import requests
from bs4 import BeautifulSoup
import pandas as pd
import pdfplumber
from docx import Document
from io import BytesIO
import os
from dotenv import load_dotenv
from groq import Groq
from urllib.parse import quote
import pytesseract
from PIL import Image

# --- Load API Key ---
load_dotenv(dotenv_path=os.path.join(os.path.dirname(__file__), '..', '.env'))
groq_api = os.getenv("GROQ_API_KEY")
try:
    client = Groq(api_key=groq_api) if groq_api else None
except Exception as e:
    st.error(f"Failed to initialize Groq client: {str(e)}")
    client = None

# --- Resume Parser ---
def extract_text(file):
    if file.name.endswith(".pdf"):
        try:
            with pdfplumber.open(file) as pdf:
                text = "\n".join(page.extract_text() or "" for page in pdf.pages)
                if not text.strip():
                    st.warning("⚠️ No text extracted from PDF. Attempting OCR...")
                    # OCR fallback
                    images = [page.to_image(resolution=300).original for page in pdf.pages]
                    ocr_text = "\n".join(pytesseract.image_to_string(Image.open(img)) for img in images)
                    return ocr_text
                return text
        except Exception as e:
            st.error(f"PDF error: {e}")
    elif file.name.endswith(".docx"):
        try:
            text = "\n".join(p.text for p in Document(file).paragraphs)
            if not text.strip():
                st.warning("⚠️ DOCX file has no readable text.")
            return text
        except Exception as e:
            st.error(f"DOCX error: {e}")
    else:
        st.warning("Unsupported file type.")
    return None

# --- Dice Scraper ---
def scrape_dice_jobs(title, loc):
    base_url = "https://www.dice.com/jobs"
    headers = {"User-Agent": "Mozilla/5.0"}
    search_url = f"{base_url}?q={quote(title)}&location={quote(loc)}"
    response = requests.get(search_url, headers=headers)
    soup = BeautifulSoup(response.text, "html.parser")

    jobs = []
    listings = soup.select("div.card")

    for card in listings:
        job_title = card.select_one("a.card-title-link")
        company = card.select_one("span.card-company")
        location = card.select_one("span.card-location")
        link = job_title['href'] if job_title else None

        if job_title and company:
            jobs.append({
                "Title": job_title.text.strip(),
                "Company": company.text.strip(),
                "Location": location.text.strip() if location else "",
                "Link": f"https://www.dice.com{link}" if link else ""
            })

    return pd.DataFrame(jobs)

# --- LinkedIn Scraper ---
def scrape_linkedin_jobs(title, loc):
    headers = {"User-Agent": "Mozilla/5.0"}
    search_url = f"https://www.linkedin.com/jobs/search/?keywords={quote(title)}&location={quote(loc)}"
    response = requests.get(search_url, headers=headers)
    soup = BeautifulSoup(response.text, "html.parser")

    jobs = []
    listings = soup.select(".base-search-card")

    for card in listings:
        job_title = card.select_one("h3")
        company = card.select_one("h4")
        location = card.select_one(".job-search-card__location")
        link = card.select_one("a")['href'] if card.select_one("a") else None

        if job_title and company:
            jobs.append({
                "Title": job_title.text.strip(),
                "Company": company.text.strip(),
                "Location": location.text.strip() if location else "",
                "Link": link or ""
            })

    return pd.DataFrame(jobs)

# --- Glassdoor Scraper ---
def scrape_glassdoor_jobs(title, loc):
    headers = {"User-Agent": "Mozilla/5.0"}
    search_url = f"https://www.glassdoor.com/Job/jobs.htm?sc.keyword={quote(title)}&locT=C&locId=0&locKeyword={quote(loc)}"
    response = requests.get(search_url, headers=headers)
    soup = BeautifulSoup(response.text, "html.parser")

    jobs = []
    listings = soup.select(".react-job-listing")

    for card in listings:
        job_title = card.select_one("a.jobLink")
        company = card.select_one("div.jobHeader")
        location = card.select_one("span.pr-xxsm")
        link = job_title['href'] if job_title else None

        if job_title and company:
            jobs.append({
                "Title": job_title.text.strip(),
                "Company": company.text.strip(),
                "Location": location.text.strip() if location else "",
                "Link": f"https://www.glassdoor.com{link}" if link else ""
            })

    return pd.DataFrame(jobs)

# --- AI Matching ---
def match_resume_with_jobs(resume_text, jobs_df):
    if not client:
        return jobs_df.assign(MatchScore="API Error")

    results = []
    for _, row in jobs_df.iterrows():
        prompt = f"""
        Compare the following resume and job title. Rate the match from 0 to 100.

        Resume:
        {resume_text[:2000]}

        Job Title: {row['Title']}
        Company: {row['Company']}

        Reply with a number only between 0 to 100.
        """
        try:
            response = client.chat.completions.create(
                model="llama3-70b-8192",
                messages=[{"role": "user", "content": prompt}],
                temperature=0.2,
                max_tokens=10
            )
            score = response.choices[0].message.content.strip()
        except Exception as e:
            score = "Error"
        results.append(score)

    jobs_df["MatchScore"] = results
    return jobs_df

# --- UI Setup ---
st.set_page_config(page_title="Job Matcher Bot", layout="wide")
st.title("🧲 Job Matcher & Cover Letter Generator")

# --- Inputs ---
st.subheader("Step 1: Upload Resume")
resume_file = st.file_uploader("Upload Resume", type=["pdf", "docx"])

st.subheader("Step 2: Search Jobs")
with st.form("job_form"):
    job_title = st.text_input("Job Title", value="Python Developer")
    location = st.text_input("Location", value="Remote")
    platform = st.selectbox("Choose Job Platform", ["Dice", "LinkedIn", "Glassdoor"])
    submit = st.form_submit_button("Search & Match")

# --- Main Execution ---
if submit and resume_file:
    with st.spinner("Extracting resume & jobs..."):
        resume_text = extract_text(resume_file)

        if platform == "Dice":
            job_results = scrape_dice_jobs(job_title, location)
        elif platform == "LinkedIn":
            job_results = scrape_linkedin_jobs(job_title, location)
        else:
            job_results = scrape_glassdoor_jobs(job_title, location)

        if job_results.empty:
            st.warning("❌ No jobs found. Try a different search.")
        elif not resume_text:
            st.error("❌ Failed to extract text from resume.")
        else:
            st.success(f"✅ Found {len(job_results)} jobs on {platform}. Matching with resume...")
            matched_jobs = match_resume_with_jobs(resume_text, job_results)
            st.dataframe(matched_jobs.sort_values("MatchScore", ascending=False), use_container_width=True)

            csv = matched_jobs.to_csv(index=False).encode('utf-8')
            st.download_button("⬇ Download Matches (CSV)", csv, "matched_jobs.csv")
else:
    st.info("📄 Please upload resume and enter search criteria to begin.")