import os
from io import BytesIO
import gradio as gr
import fitz  # PyMuPDF
import docx  # python-docx
import google.generativeai as genai

# --- Configure Google Gemini API key from Space Secret ---
GOOGLE_API_KEY = os.getenv("GOOGLE_API_KEY", "").strip()
if GOOGLE_API_KEY:
    genai.configure(api_key=GOOGLE_API_KEY)

# Create the model lazily so we can show a friendly error if the key is missing
def get_model():
    if not GOOGLE_API_KEY:
        raise RuntimeError(
            "Missing GOOGLE_API_KEY secret. Go to Settings → Variables and secrets and add it."
        )

    # List available models and print them
    available_models = genai.list_models()
    print(available_models)  # List the models to check what's available

    # Use a valid model if gemini-1.5-flash is not available
    return genai.GenerativeModel("gemini-2.5-flash")  # Or another valid model from available models


# --- Helpers to read resume files ---
def extract_text_from_resume(file_path: str) -> str:
    if not file_path:
        return ""

    file_path_lower = file_path.lower()
    try:
        if file_path_lower.endswith(".pdf"):
            text = ""
            with open(file_path, "rb") as f:
                pdf_bytes = f.read()
            doc = fitz.open(stream=pdf_bytes, filetype="pdf")
            for page in doc:
                text += page.get_text() or ""
            doc.close()
            return text

        elif file_path_lower.endswith(".docx"):
            # python-docx can open a path directly
            d = docx.Document(file_path)
            return "\n".join(p.text for p in d.paragraphs)

        elif file_path_lower.endswith(".txt"):
            with open(file_path, "r", encoding="utf-8", errors="ignore") as f:
                return f.read()

        else:
            raise ValueError("Unsupported file type. Please upload .pdf, .docx, or .txt.")
    except Exception as e:
        raise RuntimeError(f"Failed to read resume: {e}")


# --- Main analysis function ---
def analyze_resume(job_description: str, resume_file_path: str):
    try:
        model = get_model()
    except Exception as e:
        return (f"❌ {e}", "", "")

    if not job_description.strip():
        return ("❌ Please paste a job description.", "", "")

    if not resume_file_path:
        return ("❌ Please upload a resume file.", "", "")

    try:
        resume_text = extract_text_from_resume(resume_file_path)
    except Exception as e:
        return (f"⚠️ {e}", "", "")

    if not resume_text.strip():
        return ("⚠️ Could not extract text from the resume.", "", "")

    # ---------- Gemini prompts (your Colab logic, adapted) ----------
    try:
        # 1) Analysis of missing items
        analysis_prompt = f"""
Analyze the following resume based on the provided job description.
Identify missing keywords, skills, or experience mentioned in the job description
that are not present in the resume. Return a clear, readable list.
Job Description:
{job_description}
Resume:
{resume_text}
"""
        analysis_resp = model.generate_content(analysis_prompt)
        analysis_text = (getattr(analysis_resp, "text", "") or "").strip()

        # 2) ATS-optimized resume
        resume_prompt = f"""
Based on the original resume, job description, and the missing items analysis,
rewrite the resume to be ATS-optimized. Incorporate missing keywords/skills naturally.
Use concise bullet points and clear section headings. Do not use * and give ATS score too pelase
Original Resume:
{resume_text}
Job Description:
{job_description}
Missing Items Analysis:
{analysis_text}
Generate the new ATS-optimized resume:
"""
        resume_resp = model.generate_content(resume_prompt)
        optimized_resume = (getattr(resume_resp, "text", "") or "").strip()

        # 3) Tailored cover letter
        cover_letter_prompt = f"""
Write a tailored cover letter that aligns the candidate's experience to the job.
Keep it 200–350 words, professional, specific, and with a strong call to action.
Job Description:
{job_description}
ATS-Optimized Resume:
{optimized_resume}
"""
        cover_letter_resp = model.generate_content(cover_letter_prompt)
        cover_letter_text = (getattr(cover_letter_resp, "text", "") or "").strip()

        return (analysis_text, optimized_resume, cover_letter_text)

    except Exception as e:
        return (f"⚠️ Error analyzing with Gemini: {e}", "", "")


# --- Gradio UI ---
with gr.Blocks() as demo:
    gr.Markdown("## 📊 HireReady — Resume & Job Match Analyzer (Google Gemini)")

    with gr.Row():
        job_desc = gr.Textbox(
            label="Job Description",
            placeholder="Paste the job description here...",
            lines=10
        )

    with gr.Row():
        resume_file = gr.File(
            label="Upload Resume (.pdf, .docx, .txt)",
            file_count="single",
            file_types=[".pdf", ".docx", ".txt"],
            type="filepath"  # returns a temp file path (string)
        )

    analyze_btn = gr.Button("🔍 Analyze")
    analysis_output = gr.Textbox(label="Analysis of Missing Items", lines=14)
    resume_output = gr.Textbox(label="ATS-Optimized Resume", lines=18)
    cover_letter_output = gr.Textbox(label="Tailored Cover Letter", lines=16)

    analyze_btn.click(
        fn=analyze_resume,
        inputs=[job_desc, resume_file],
        outputs=[analysis_output, resume_output, cover_letter_output]
    )

# For Spaces
if __name__ == "__main__":
    demo.launch()