Spaces:
Sleeping
Sleeping
File size: 3,481 Bytes
b997da5 b2647bd b997da5 97a5f0e b997da5 | 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 | 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))
|