import gradio as gr
import numpy as np
from text_extractor import extract_text_from_file
from embedder import get_embeddings
from vector_store import create_faiss_index, search_similar_cvs
from groq_api import summarize_match

# Global storage
cv_texts = []
cv_names = []
cv_vectors = []
faiss_index = None

def upload_cvs(files):
    global cv_texts, cv_names, cv_vectors, faiss_index

    try:
        if len(files) > 10:
            return "❌ Limit exceeded: Upload a maximum of 10 CVs."

        # Remove duplicates based on filename
        unique_files = []
        seen = set()
        for f in files:
            if f.name not in seen:
                seen.add(f.name)
                unique_files.append(f)
        files = unique_files

        cv_texts = [extract_text_from_file(f) for f in files]
        cv_names = [f.name for f in files]
        cv_vectors = get_embeddings(cv_texts)

        if cv_vectors is None or np.array(cv_vectors).size == 0:
            return "❌ No valid CVs."

        faiss_index = create_faiss_index(cv_vectors)
        return f"✅ Uploaded and indexed {len(files)} CV(s)."

    except Exception as e:
        return f"❌ Error during upload: {e}"

def match_jd(jd_text, match_mode):
    if faiss_index is None:
        return "❌ Please upload CVs first."
    if not jd_text.strip():
        return "⚠️ Job description is empty."

    try:
        jd_vector = get_embeddings([jd_text])[0]

        # Select CVs based on match mode
        if match_mode == "Top 3 Matches":
            indices = search_similar_cvs(jd_vector, faiss_index, k=3)
        else:  # All uploaded CVs
            indices = list(range(len(cv_names)))

        # Filter duplicates by name
        seen = set()
        unique_indices = []
        for i in indices:
            if cv_names[i] not in seen:
                seen.add(cv_names[i])
                unique_indices.append(i)

        matched = [cv_names[i] for i in unique_indices]
        texts = [cv_texts[i] for i in unique_indices]

        summary = summarize_match(jd_text, matched, texts)

        title = "✅ Match Result:" if len(matched) == 1 else f"✅ Matching {len(matched)} CVs:"
        return f"{title}\n\n" + "\n".join(matched) + f"\n\n📝 Summary:\n{summary}"

    except Exception as e:
        return f"❌ Error during matching: {e}"

def clear_data():
    global cv_texts, cv_names, cv_vectors, faiss_index
    cv_texts, cv_names, cv_vectors, faiss_index = [], [], [], None
    return "🧹 Cleared."

with gr.Blocks() as app:
    gr.Markdown("## 📄 CV Matcher with Groq API (Dynamic Matching)")

    # Upload
    file_input = gr.File(file_types=[".pdf", ".docx"], file_count="multiple", label="📤 Upload CVs (Max 10)")
    upload_button = gr.Button("📁 Upload & Index")
    upload_status = gr.Textbox(label="Upload Status")

    # Job Description Matching
    jd_input = gr.Textbox(label="📋 Paste Job Description", lines=8, placeholder="Paste job description here...")

    match_mode = gr.Radio(["Top 3 Matches", "All Uploaded CVs"], value="Top 3 Matches", label="Matching Mode")

    match_button = gr.Button("🔍 Match CVs")
    result_output = gr.Textbox(label="Match Results", lines=25)

    # Clear Session
    clear_button = gr.Button("🧹 Clear All")
    clear_status = gr.Textbox(label="Clear Status")

    # Actions
    upload_button.click(upload_cvs, inputs=[file_input], outputs=[upload_status])
    match_button.click(match_jd, inputs=[jd_input, match_mode], outputs=[result_output])
    clear_button.click(clear_data, inputs=[], outputs=[clear_status])

app.launch()