app.py

import gradio as gr
import pandas as pd
import requests
import re
import tempfile
import shutil
import os
from difflib import SequenceMatcher
import json
from urllib.parse import quote_plus
import zipfile
from datetime import datetime


# -----------zip_and_prepare_download--------------
def zip_and_prepare_download(file_bytes, inner_filename, zip_prefix="Download"):
    zip_filename = f"{zip_prefix}_{datetime.now().strftime('%Y%m%d_%H%M%S')}.zip"
    zip_file_path = tempfile.NamedTemporaryFile(delete=False, suffix=".zip").name

    with zipfile.ZipFile(zip_file_path, 'w') as zipf:
        zipf.writestr(inner_filename, file_bytes)

    print(f"[DEBUG] Created ZIP at {zip_file_path} with filename {zip_filename}")
    return zip_file_path


# -----------Utilities--------------
def construct_query(row):
    query = str(row['Applicant Name'])
    optional_fields = ['Job Title', 'State', 'City', 'Skills']

    for field in optional_fields:
        if field in row and pd.notna(row[field]):
            value = row[field]
            query += f" {str(value).strip()}" if str(value).strip() else ""

    query += " linkedin"
    print(f"[DEBUG] Search Query: {query}")
    return query

def get_name_from_url(link):
    match = re.search(r'linkedin\.com/in/([a-zA-Z0-9-]+)', link)
    if match:
        profile_name = match.group(1).replace('-', ' ')
        print(f"[DEBUG] Extracted profile name from URL: {profile_name}")
        return profile_name
    return None

def calculate_similarity(name1, name2):
    similarity = SequenceMatcher(None, name1.lower().strip(), name2.lower().strip()).ratio()
    print(f"[DEBUG] Similarity between '{name1}' and '{name2}' = {similarity}")
    return similarity

def fetch_linkedin_links(query, api_key, applicant_name):
    try:
        print(f"[DEBUG] Sending request to BrightData for query: {query}")
        url = "https://api.brightdata.com/request"
        google_url = f"https://www.google.com/search?q={quote_plus(query)}"

        payload = {
            "zone": "serp_api2",
            "url": google_url,
            "method": "GET",
            "country": "us",
            "format": "raw",
            "data_format": "html"
        }

        headers = {
            "Authorization": f"Bearer {api_key}",
            "Content-Type": "application/json"
        }

        response = requests.post(url, headers=headers, json=payload)
        response.raise_for_status()
        html = response.text

        linkedin_regex = r'https://(?:[a-z]{2,3}\.)?linkedin\.com/in/[a-zA-Z0-9\-_/]+'
        matches = re.findall(linkedin_regex, html)
        print(f"[DEBUG] Found {len(matches)} LinkedIn link(s) in search result")

        for link in matches:
            profile_name = get_name_from_url(link)
            if profile_name:
                similarity = calculate_similarity(applicant_name, profile_name)
                if similarity >= 0.5:
                    print(f"[DEBUG] Match found: {link}")
                    return link
        print(f"[DEBUG] No matching LinkedIn profile found for: {applicant_name}")
        return None

    except Exception as e:
        print(f"[ERROR] Error fetching LinkedIn link for query '{query}': {e}")
        return None


# ----------Process Excel---------------
def process_file_gradio(file_obj, api_key):
    try:
        df = pd.read_excel(file_obj.name)
        print(f"[DEBUG] Input file read successfully. Rows: {len(df)}")

        if 'Applicant Name' not in df.columns:
            return None, "❌ Missing required column: 'Applicant Name'"

        df = df[df['Applicant Name'].notna()]
        df = df[df['Applicant Name'].str.strip() != '']
        print(f"[DEBUG] Valid applicant rows after filtering: {len(df)}")

        df['Search Query'] = df.apply(construct_query, axis=1)
        df['LinkedIn Link'] = df.apply(
            lambda row: fetch_linkedin_links(row['Search Query'], api_key, row['Applicant Name']),
            axis=1
        )

        temp_dir = tempfile.mkdtemp()
        output_file = os.path.join(temp_dir, "updated_with_linkedin_links.csv")
        df.to_csv(output_file, index=False)
        print(f"[DEBUG] Output written to: {output_file}")

        with open(output_file, "rb") as f:
            csv_bytes = f.read()

        zip_path = zip_and_prepare_download(csv_bytes, "updated_with_linkedin_links.csv", "LinkedIn_Links")

        shutil.rmtree(temp_dir)
        return zip_path, "✅ Success! Download your file below."

    except Exception as e:
        print(f"[ERROR] Error processing file: {e}")
        return None, f"❌ Error: {str(e)}"


# ----------Gradio UI---------------
with gr.Blocks(title="LinkedIn Scraper") as demo:
    gr.Markdown("## 🔗 LinkedIn Profile Scraper")
    gr.Markdown("Upload an Excel file with applicant details to fetch best-matching LinkedIn profile links (via Google Search using BrightData API).")

    with gr.Row():
        api_key_input = gr.Textbox(label="🔐 BrightData API Key", type="password", placeholder="Enter your BrightData SERP API Key")
        file_input = gr.File(label="📤 Upload Excel File (.xlsx)", file_types=[".xlsx"])

    process_btn = gr.Button("🚀 Start Processing")
    status_output = gr.Textbox(label="📢 Status")
    download_btn = gr.File(label="📥 Download ZIP")

    def run_pipeline(api_key, file):
        if not api_key:
            return None, "❗ Please enter your API key"
        if not file:
            return None, "❗ Please upload a valid Excel file"
        return process_file_gradio(file, api_key)

    process_btn.click(
        fn=run_pipeline,
        inputs=[api_key_input, file_input],
        outputs=[download_btn, status_output]
    )

# Run app
if __name__ == "__main__":
    demo.launch()