app.py

import gradio as gr
from docx import Document
from docx.shared import Pt, RGBColor
import tempfile
import re
import logging
import os

logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)

STOP_PHRASES_LOWER = {
    # (add your full STOP_PHRASES_LOWER set here)
}

def extract_keywords_from_text(text):
    if not text:
        return []
    extracted_terms = set()
    tech_keyword_regex = r"""
        \b(?:
            [A-Z]{2,}[A-Z0-9./+#-]*[A-Z0-9]? |
            [A-Za-z][A-Za-z0-9]*\.(?:NET|js|io|ai|com|org) |
            [A-Za-z][A-Za-z0-9]*\+\+ |
            [A-Z][A-Za-z0-9#+-]+ |
            [A-Z][a-z]+[A-Z][A-Za-z0-9]+
        )\b
    """
    for match in re.finditer(tech_keyword_regex, text, re.VERBOSE):
        term = match.group(0).strip()
        if term:
            extracted_terms.add(term)

    main_word_pattern = r"(?:[A-Z][A-Za-z0-9'-]*|[A-Z]{2,}[A-Z0-9'-]*)"
    connector_pattern = r"(?:\s+(?:of|and|for|in|the|to|with|&|on|as|at|by|per)\s+)?"
    for num_main_words in range(2, 6):
        current_phrase_regex_str = r"\b" + main_word_pattern + \
                                   (connector_pattern + main_word_pattern) * (num_main_words - 1) + \
                                   r"\b"
        try:
            for match in re.finditer(current_phrase_regex_str, text):
                term = match.group(0).strip()
                term = re.sub(r'\s+', ' ', term)
                if term:
                    extracted_terms.add(term)
        except re.error as e:
            logger.warning(f"Regex error for multi-word phrases (N={num_main_words}): {e}.")
            continue

    final_cleaned_terms = set()
    for term in extracted_terms:
        term_normalized = term.strip()
        if not term_normalized or len(term_normalized) < 1 or len(term_normalized) > 75:
            continue
        if term_normalized.lower() in STOP_PHRASES_LOWER:
            continue
        if not re.search(r'[A-Za-z0-9]', term_normalized):
            continue
        final_cleaned_terms.add(term_normalized)

    return sorted(final_cleaned_terms, key=lambda x: (-len(x), x.lower()))

def extract_ats_keywords(job_description_text):
    logger.info("Extracting keywords from Job Description...")
    return extract_keywords_from_text(job_description_text)

def extract_resume_keywords(resume_text):
    logger.info("Extracting keywords from Resume...")
    return extract_keywords_from_text(resume_text)

def analyze_resume_for_suggestions(docx_file_obj, job_description_text):
    if docx_file_obj is None or not job_description_text.strip():
        return "⚠️ Error: Please upload your resume and paste the job description.", gr.update(choices=[], value=[], visible=False), gr.update(visible=False), gr.update(visible=False), None

    temp_resume_path = docx_file_obj.name
    resume_doc = Document(temp_resume_path)
    resume_text = "\n".join([p.text for p in resume_doc.paragraphs])
    ats_keywords = extract_ats_keywords(job_description_text)
    resume_keywords = extract_resume_keywords(resume_text)

    ats_keywords_lower_set = {kw.lower() for kw in ats_keywords}
    resume_keywords_lower_set = {kw.lower() for kw in resume_keywords}

    matching_keywords = [kw for kw in ats_keywords if kw.lower() in resume_keywords_lower_set]
    missing_keywords = [kw for kw in ats_keywords if kw.lower() not in resume_keywords_lower_set]

    coverage = len(matching_keywords) / len(ats_keywords_lower_set) * 100 if ats_keywords_lower_set else 0

    results_str = f"📊 **Resume Scan Results** 📊\n\n"
    results_str += f"🔑 **Job Description Keywords:**\n_{', '.join(ats_keywords)}_\n\n"
    results_str += f"📄 **Keywords Found in Your Resume:**\n_{', '.join(resume_keywords)}_\n\n"
    results_str += f"✅ **Matching Keywords:**\n_{', '.join(matching_keywords)}_\n\n"
    results_str += f"🎯 **Keyword Coverage: {coverage:.2f}%**\n\n"

    TARGET_COVERAGE = 75.0
    if coverage >= TARGET_COVERAGE:
        results_str += f"🎉 **Great! Your resume meets {TARGET_COVERAGE}% coverage!** 🎉"
        return results_str, gr.update(choices=[], value=[], visible=False), gr.update(visible=False), gr.update(visible=False), temp_resume_path

    needed = max(0, int(len(ats_keywords_lower_set) * (TARGET_COVERAGE/100.0)) - len(matching_keywords))
    results_str += f"💡 **Suggested Keywords to Add (select below):** Add approx **{needed}** more keywords."

    return results_str, gr.update(choices=missing_keywords, value=[], visible=True), gr.update(visible=True), gr.update(visible=True), temp_resume_path

def add_selected_keywords_to_docx(selected_keywords, resume_path_from_state):
    yield "⏳ Processing... Adding selected keywords to your resume.", gr.update(value=None, visible=False, interactive=False)

    if not selected_keywords:
        yield "ℹ️ No keywords selected; nothing to do.", gr.update(value=None, visible=False, interactive=False)
        return

    if not resume_path_from_state or not os.path.exists(resume_path_from_state):
        yield "⚠️ Error: Original resume path not found. Please re‑analyze.", gr.update(value=None, visible=False, interactive=False)
        return

    resume_doc = Document(resume_path_from_state)
    keywords_string = " ".join(selected_keywords)
    p = resume_doc.add_paragraph()
    run = p.add_run(keywords_string)

    run.font.size = Pt(1)
    run.font.color.rgb = RGBColor(255, 255, 255)

    with tempfile.NamedTemporaryFile(delete=False, suffix=".docx") as tmp_file:
        modified_resume_path = tmp_file.name
    resume_doc.save(modified_resume_path)

    yield f"✅ Resume updated: {len(selected_keywords)} keyword(s) appended as white 1 pt text.", gr.update(value=modified_resume_path, visible=True, interactive=True)

with gr.Blocks(theme=gr.themes.Soft()) as demo:
    gr.Markdown("# 🚀 ATS Resume Optimizer")
    gr.Markdown("Upload .docx, paste JD, analyze, select keywords to add.")

    resume_path_state = gr.State(None)

    with gr.Row():
        with gr.Column(scale=1):
            docx_file_input = gr.File(label="📄 Upload Resume", file_types=[".docx"])
            job_desc_input = gr.Textbox(lines=12, label="📝 Paste Job Description")
            analyze_button = gr.Button("🔍 Analyze", variant="primary")
        with gr.Column(scale=2):
            analysis_results_markdown = gr.Markdown()
            with gr.Group(visible=False) as keyword_selection_group:
                missing_keywords_checkboxes = gr.CheckboxGroup(label="💡 Select Keywords", choices=[])
                add_selected_button = gr.Button("✍️ Add Keywords", variant="secondary", visible=False)
            modification_status_markdown = gr.Markdown()
            download_optimized_resume_button = gr.File(label="⬇️ Download Resume", visible=False, interactive=False)

    analyze_button.click(
        fn=analyze_resume_for_suggestions,
        inputs=[docx_file_input, job_desc_input],
        outputs=[analysis_results_markdown, missing_keywords_checkboxes, keyword_selection_group, add_selected_button, resume_path_state]
    )
    add_selected_button.click(
        fn=add_selected_keywords_to_docx,
        inputs=[missing_keywords_checkboxes, resume_path_state],
        outputs=[modification_status_markdown, download_optimized_resume_button]
    )

    def clear_on_new_input():
        return "", gr.update(choices=[], value=[], visible=False), gr.update(visible=False), gr.update(visible=False), "", gr.update(visible=False, value=None, interactive=False), None

    docx_file_input.change(fn=clear_on_new_input, inputs=None, outputs=[
        analysis_results_markdown, missing_keywords_checkboxes, keyword_selection_group,
        add_selected_button, modification_status_markdown, download_optimized_resume_button,
        resume_path_state
    ])
    job_desc_input.change(fn=clear_on_new_input, inputs=None, outputs=[
        analysis_results_markdown, missing_keywords_checkboxes, keyword_selection_group,
        add_selected_button, modification_status_markdown, download_optimized_resume_button,
        resume_path_state
    ])

demo.launch()