from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from openai import AzureOpenAI as moa
from serpapi import GoogleSearch
import settings

import os
import re
import faiss
from llama_index.core import Document, VectorStoreIndex, StorageContext
from llama_index.vector_stores.faiss import FaissVectorStore

from dotenv import load_dotenv

load_dotenv()

app = FastAPI(title="IntelliHire")

client = moa(
    api_version=os.getenv('AZURE_VERSION'),
    azure_endpoint=os.getenv('AZURE_OPENAI_ENDPOINT'),
    api_key=os.getenv('AZURE_API_KEY'),
)

class QueryRequest(BaseModel):
    query: str

def generate_search_query(user_query: str) -> str:
    response = client.chat.completions.create(
        model='agile4',
        messages=[
            {
                "role": "system",
                "content": "You are a helpful assistant.",
            },
            {
                "role": "user",
                "content": f"""Convert the given query to google search for linked site ex: 'Find devops engineer in Berlin' → site:linkedin.com/in/ 'devops engineer' Berlin". Query: {user_query}, Response: <query>...</query>""",
            }
        ],
        max_completion_tokens=800,
        temperature=0.7,
    )
    text = response.choices[0].message.content
    pattern = r'<query>(.*?)</query>'
    matches = re.findall(pattern, text)
    if not matches:
        raise ValueError("Failed to parse search query from model response")
    return matches[0]

def search_google(query: str) -> list[Document]:
    params = {
        "api_key": "cf0f20c18d0e9ed6f60e66765ca6a582bf5a6297261a22e177b5cbed254c0f54",
        "engine": "google",
        "q": query,
        "google_domain": "google.co.in",
        "gl": "in",
        "hl": "en",
        "location": "India",
        "num": "50"
    }
    search = GoogleSearch(params)
    results = search.get_dict()
    docs = []
    for data in results.get('organic_results', []):
        docs.append(Document(
            text=data.get('snippet', ''),
            metadata={
                'name': ' '.join(data.get('title', '').split()[:2]),
                'link': data.get('link')
            }
        ))
    return docs

def create_index(docs: list[Document]) -> VectorStoreIndex:
    d = 1536
    faiss_index = faiss.IndexFlatL2(d)
    vector_store = FaissVectorStore(faiss_index=faiss_index)
    storage_context = StorageContext.from_defaults(vector_store=vector_store)
    return VectorStoreIndex.from_documents(
        docs,
        storage_context=storage_context,
        show_progress=True,
    )

@app.post("/search")
def search(query_request: QueryRequest):
    try:
        search_query = generate_search_query(query_request.query)
        docs = search_google(search_query)
        if not docs:
            return {"message": "No documents found"}
        index = create_index(docs)
        query_engine = index.as_query_engine(similarity_top_k=10)
        response = query_engine.query(query_request.query)
        # results=[]
        # for node in response.source_nodes:
        #     results.append({
        #         "metadata": node.metadata,
        #         "text": node.text
        #     })
        results = [
            {"metadata": node.metadata, "text": node.text}
            for node in response.source_nodes
        ]
        print(results)
        return {"results": results}
    except Exception as e:
        raise HTTPException(status_code=500, detail=str(e))