app.py

"""
AI Resume Studio – Hugging Face Space
Author: Oluwafemi Idiakhoa
Last update: 2025‑06‑27

Features
────────
1. Generate resume → Word & PDF downloads
2. Score resume against job description
3. AI Section Co‑Pilot (rewrite, quantify, bulletize…)
4. Cover‑letter generator
5. Job‑description scraper by URL
6. Multilingual export via DeepL (free & pro keys supported)
"""

# ──────────────────────────────────────────────────────────────────────────────
# Imports & setup
# ──────────────────────────────────────────────────────────────────────────────
import os, tempfile
import requests
import gradio as gr
import google.generativeai as genai
from dotenv import load_dotenv
from bs4 import BeautifulSoup
from docx import Document                 # python‑docx
from reportlab.lib.pagesizes import LETTER
from reportlab.pdfgen import canvas
import deepl                              # DeepL translation

load_dotenv()

# Gemini model (make sure key has access to this version)
genai.configure(api_key=os.getenv("API_KEY"))
GEMINI_MODEL = genai.GenerativeModel("gemini-1.5-pro-latest")

# DeepL translator
DEEPL_KEY = os.getenv("DEEPL_API_KEY")
DEEPL_TRANSLATOR = deepl.Translator(DEEPL_KEY) if DEEPL_KEY else None

# Supported DeepL target languages (code → label)
LANGS = {
    "EN": "English",
    "DE": "German",
    "FR": "French",
    "ES": "Spanish",
    "IT": "Italian",
    "NL": "Dutch",
    "PT-PT": "Portuguese",
    "PT-BR": "Portuguese (BR)",
    "PL": "Polish",
    "JA": "Japanese",
    "ZH": "Chinese",
}

# ──────────────────────────────────────────────────────────────────────────────
# Helpers
# ──────────────────────────────────────────────────────────────────────────────
def ask_gemini(prompt: str, temperature: float = 0.6) -> str:
    try:
        rsp = GEMINI_MODEL.generate_content(
            prompt, generation_config={"temperature": temperature}
        )
        return rsp.text.strip()
    except Exception as err:
        return f"[Gemini Error] {err}"

def translate_if_needed(text: str, target_code: str) -> str:
    if target_code == "EN" or not DEEPL_TRANSLATOR:
        return text  # already English or DeepL key missing
    try:
        res = DEEPL_TRANSLATOR.translate_text(text, target_lang=target_code)
        return res.text
    except Exception as err:
        return f"[DeepL Error] {err}\n\n{text}"

def save_docx(text: str) -> str:
    tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".docx")
    doc = Document()
    for line in text.splitlines():
        doc.add_paragraph(line)
    doc.save(tmp.name)
    return tmp.name

def save_pdf(text: str) -> str:
    tmp = tempfile.NamedTemporaryFile(delete=False, suffix=".pdf")
    c = canvas.Canvas(tmp.name, pagesize=LETTER)
    width, height = LETTER
    y = height - 72
    for line in text.splitlines():
        c.drawString(72, y, line)
        y -= 14
        if y < 72:
            c.showPage()
            y = height - 72
    c.save()
    return tmp.name

# ──────────────────────────────────────────────────────────────────────────────
# Core AI functions
# ──────────────────────────────────────────────────────────────────────────────
def generate_resume(name, email, phone, summary, experience, education, skills, lang):
    prompt = f"""
Create a professional resume in Markdown. **Do not use first‑person pronouns.**

Target language: {LANGS[lang]}

Name: {name}
Email: {email}
Phone: {phone}

Professional Summary:
{summary}

Experience:
{experience}

Education:
{education}

Skills:
{skills}
"""
    resume_md = ask_gemini(prompt)
    return translate_if_needed(resume_md, lang)

def score_resume(resume_md, job_desc):
    prompt = f"""
Evaluate the RESUME against the JOB DESCRIPTION. Return *only* compact Markdown:
### Match Score
<integer 0‑100>

### Suggestions
- <bullet 1>
- <bullet 2>
- <bullet 3>
"""
    return ask_gemini(prompt.format(resume_md=resume_md, job_desc=job_desc), temperature=0.4)

def refine_section(section_text, instruction, lang):
    prompt = f"""
Perform the following instruction on the resume section. Respond in {LANGS[lang]}.

Instruction: {instruction}
Text:
{section_text}
"""
    refined = ask_gemini(prompt)
    return translate_if_needed(refined, lang)

def generate_cover_letter(resume_md, job_desc, tone, lang):
    prompt = f"""
Draft a one‑page cover letter (≤ 300 words) in {tone} tone, aligning the RESUME
to the JOB DESCRIPTION. Use {LANGS[lang]} throughout. Salutation: "Dear Hiring Manager,".

RESUME:
{resume_md}

JOB DESCRIPTION:
{job_desc}
"""
    letter = ask_gemini(prompt)
    return translate_if_needed(letter, lang)

# ──────────────────────────────────────────────────────────────────────────────
# JD scraper
# ──────────────────────────────────────────────────────────────────────────────
def scrape_job_description(url):
    try:
        headers = {"User-Agent": "Mozilla/5.0"}
        page = requests.get(url, headers=headers, timeout=10)
        soup = BeautifulSoup(page.text, "html.parser")

        # Heuristics for popular boards
        selectors = [
            "div.jobsearch-jobDescriptionText",          # Indeed
            "section.description",                       # generic
            "div.jobs-description__content",             # LinkedIn
            "div#job-details"                            # Greenhouse
        ]
        for sel in selectors:
            block = soup.select_one(sel)
            if block:
                return block.get_text(" ", strip=True)
        return soup.get_text(" ", strip=True)[:3000]
    except Exception as e:
        return f"[Error] {e}"

# ──────────────────────────────────────────────────────────────────────────────
# Wrapper to export resume + files
# ──────────────────────────────────────────────────────────────────────────────
def generate_resume_and_files(name, email, phone, summary,
                              experience, education, skills, lang):
    resume_md = generate_resume(name, email, phone, summary,
                                experience, education, skills, lang)
    return resume_md, save_docx(resume_md), save_pdf(resume_md)

# ──────────────────────────────────────────────────────────────────────────────
# Gradio UI
# ──────────────────────────────────────────────────────────────────────────────
with gr.Blocks(title="AI Resume Studio") as demo:
    gr.Markdown("## 🧠 Interactive AI Resume Studio (Gemini × DeepL)")

    LANG_CHOICES = [(v, k) for k, v in LANGS.items()]

    # ── 1 · Resume generation ────────────────────────────────────────────────
    with gr.Tab("📄 Generate Resume"):
        with gr.Row():
            name  = gr.Textbox(label="Name")
            email = gr.Textbox(label="Email")
            phone = gr.Textbox(label="Phone")
        summary    = gr.Textbox(label="Professional Summary")
        experience = gr.Textbox(label="Experience")
        education  = gr.Textbox(label="Education")
        skills     = gr.Textbox(label="Skills")
        lang_sel   = gr.Dropdown(LANG_CHOICES, value="EN", label="Output Language")

        resume_md  = gr.Markdown(label="Generated Resume")
        docx_file  = gr.File(label="⬇ Word (.docx)")
        pdf_file   = gr.File(label="⬇ PDF (.pdf)")
        gen_btn    = gr.Button("Generate Resume")

        gen_btn.click(
            generate_resume_and_files,
            inputs=[name, email, phone, summary, experience, education, skills, lang_sel],
            outputs=[resume_md, docx_file, pdf_file],
        )

    # ── 2 · Match score ──────────────────────────────────────────────────────
    with gr.Tab("🧮 Score Resume Against Job"):
        resume_in = gr.Textbox(label="Resume (Markdown)", lines=12)
        jd_in     = gr.Textbox(label="Job Description",  lines=8)
        score_out = gr.Markdown(label="Score & Suggestions")
        score_btn = gr.Button("Evaluate")

        score_btn.click(score_resume, inputs=[resume_in, jd_in], outputs=score_out)

    # ── 3 · AI Section Co‑Pilot ──────────────────────────────────────────────
    with gr.Tab("✏️ AI Section Co‑Pilot"):
        sec_text = gr.Textbox(label="Section Text", lines=6)
        action   = gr.Radio(
            ["Rewrite", "Make More Concise", "Quantify Achievements", "Convert to Bullet Points"],
            label="Action"
        )
        sec_lang = gr.Dropdown(LANG_CHOICES, value="EN", label="Language")
        sec_out  = gr.Textbox(label="AI Output", lines=6)
        sec_btn  = gr.Button("Apply")

        sec_btn.click(
            refine_section, inputs=[sec_text, action, sec_lang], outputs=sec_out
        )

    # ── 4 · Cover‑letter generator ───────────────────────────────────────────
    with gr.Tab("📧 Cover‑Letter Generator"):
        cv_resume = gr.Textbox(label="Resume (Markdown)", lines=12)
        cv_jd     = gr.Textbox(label="Job Description",  lines=8)
        cv_tone   = gr.Radio(["Professional", "Friendly", "Enthusiastic"], value="Professional", label="Tone")
        cv_lang   = gr.Dropdown(LANG_CHOICES, value="EN", label="Language")
        cv_out    = gr.Markdown(label="Cover Letter")
        cv_btn    = gr.Button("Generate Cover Letter")

        cv_btn.click(
            generate_cover_letter,
            inputs=[cv_resume, cv_jd, cv_tone, cv_lang],
            outputs=cv_out,
        )

    # ── 5 · Job‑description scraper ──────────────────────────────────────────
    with gr.Tab("🌐 Job Description Scraper"):
        url_in  = gr.Textbox(label="Job URL")
        jd_out  = gr.Textbox(label="Extracted Description", lines=12)
        scrape_btn = gr.Button("Fetch Description")

        scrape_btn.click(scrape_job_description, inputs=url_in, outputs=jd_out)

demo.launch(share=False)  # HF Spaces already publishes the app