URL-To-Answer / main_api.py
MohamedFahim's picture
Add application file
ee16852
from fastapi import FastAPI, HTTPException
from pydantic import BaseModel
from typing import List
import requests
from bs4 import BeautifulSoup
import time
import os
import json
import random
import logging
import groq
import numpy as np
from sklearn.metrics.pairwise import cosine_similarity
import uvicorn
from supabase import create_client, Client
from urllib.parse import urljoin, urlparse
# Initialize FastAPI app
app = FastAPI(
title="Web RAG System API",
description="Extract content from web pages and perform RAG operations",
version="1.0.0"
)
# Configure logging
logging.basicConfig(level=logging.INFO)
logger = logging.getLogger(__name__)
# Initialize Supabase client with environment variables
try:
url = os.environ.get('SUPABASE_URL')
key = os.environ.get('SUPABASE_SERVICE_ROLE_KEY')
if not url or not key:
logger.warning("Supabase credentials not found in environment variables")
supabase = None
else:
supabase: Client = create_client(url, key)
logger.info("Supabase client initialized successfully")
except Exception as e:
logger.error(f"Failed to initialize Supabase client: {e}")
supabase = None
# User agents for web scraping
user_agents = [
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/114.0.0.0 Safari/537.36",
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Firefox/102.0",
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Safari/605.1.15",
"Mozilla/5.0 (X11; Ubuntu; Linux x86_64; rv:102.0) Gecko/20100101 Firefox/102.0",
"Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101 Firefox/102.0",
"Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Safari/537.36",
"Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Safari/537.36",
"Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Edge/103.0.1264.49",
"Mozilla/5.0 (iPhone; CPU iPhone OS 15_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Mobile/15E148 Safari/604.1",
"Mozilla/5.0 (iPad; CPU OS 15_5 like Mac OS X) AppleWebKit/605.1.15 (KHTML, like Gecko) Version/15.5 Mobile/15E148 Safari/604.1",
"Mozilla/5.0 (Linux; Android 12; SM-G991B) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Mobile Safari/537.36",
"Mozilla/5.0 (Linux; Android 11; Pixel 5) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Mobile Safari/537.36",
"Mozilla/5.0 (Linux; Android 11; SM-A217F) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Mobile Safari/537.36",
"Mozilla/5.0 (Linux; Android 10; SM-G975F) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.0.0 Mobile Safari/537.36"
]
# Pydantic models
class RAGRequest(BaseModel):
file_path: str
prompt: str
class URL(BaseModel):
url: str
@app.get("/")
async def root():
"""Health check endpoint"""
return {"message": "Web RAG System API is running", "status": "healthy"}
@app.get("/health")
async def health_check():
"""Detailed health check"""
health_status = {
"api": "healthy",
"supabase": "connected" if supabase else "not configured",
"hf_token": "configured" if os.environ.get('hf_token') else "not configured",
"groq_token": "configured" if os.environ.get('groq_token') else "not configured"
}
return health_status
@app.post("/rag")
async def rag(request: RAGRequest):
"""Perform RAG operations on extracted text"""
try:
# Check required environment variables
hf_token = os.environ.get('hf_token')
groq_token = os.environ.get('groq_token')
if not hf_token:
raise HTTPException(status_code=500, detail="HuggingFace token not configured")
if not groq_token:
raise HTTPException(status_code=500, detail="Groq token not configured")
if not supabase:
raise HTTPException(status_code=500, detail="Supabase not configured")
logger.info(f"Processing RAG request for file: {request.file_path}")
# HuggingFace Inference API for embeddings
API_URL = "https://router.huggingface.co/hf-inference/models/BAAI/bge-large-en-v1.5/pipeline/feature-extraction"
headers = {
"Authorization": hf_token,
}
def query(payload):
response = requests.post(API_URL, headers=headers, json=payload)
if response.status_code != 200:
logger.error(f"HuggingFace API error: {response.status_code} - {response.text}")
raise HTTPException(status_code=500, detail="Failed to get embeddings from HuggingFace")
return response.json()
# Create a Groq client
groq_client = groq.Client(api_key=groq_token)
def process_with_groq(query_text, context):
prompt = f"""
Context information:
{context}
Based on the context information above, please answer the following question:
{query_text}
Answer:
"""
try:
response = groq_client.chat.completions.create(
messages=[{"role": "user", "content": prompt}],
model="llama-3.3-70b-versatile",
temperature=0.4,
max_tokens=512
)
return response.choices[0].message.content
except Exception as e:
logger.error(f"Groq API error: {e}")
raise HTTPException(status_code=500, detail="Failed to process with Groq")
def get_file_from_supabase(bucket_name, file_path):
try:
response = supabase.storage.from_(bucket_name).download(file_path)
content = response.decode('utf-8')
return content
except Exception as e:
logger.error(f"Error downloading file from Supabase: {e}")
raise HTTPException(
status_code=404,
detail=f"File not found in Supabase bucket: {file_path}"
)
# Get file content from Supabase
bucket_name = "url-2-ans-bucket"
file_path = request.file_path
content = get_file_from_supabase(bucket_name, file_path)
logger.info(f"Successfully downloaded file from Supabase: {file_path}")
# Simple text chunking
chunk_size = 1000
overlap = 200
chunks = []
for i in range(0, len(content), chunk_size - overlap):
chunk = content[i:i + chunk_size]
if len(chunk) > 100:
chunks.append({"text": chunk, "position": i})
logger.info(f"Created {len(chunks)} chunks from document")
# Get embeddings for all chunks
chunk_embeddings = []
for chunk in chunks:
embedding = query({"inputs": chunk["text"]})
chunk_embeddings.append(embedding)
# Get embedding for the query
query_embedding = query({"inputs": request.prompt})
# Calculate similarity between query and all chunks
similarities = []
for chunk_embedding in chunk_embeddings:
query_np = np.array(query_embedding)
chunk_np = np.array(chunk_embedding)
if len(query_np.shape) == 1:
query_np = query_np.reshape(1, -1)
if len(chunk_np.shape) == 1:
chunk_np = chunk_np.reshape(1, -1)
similarity = cosine_similarity(query_np, chunk_np)[0][0]
similarities.append(similarity)
# Get top 3 most similar chunks
top_k = 3
top_indices = np.argsort(similarities)[-top_k:][::-1]
relevant_chunks = [chunks[i]["text"] for i in top_indices]
context_text = "\n\n".join(relevant_chunks)
# Process with Groq
answer = process_with_groq(request.prompt, context_text)
# Prepare sources
sources = [{"text": chunks[i]["text"][:200] + "...", "position": chunks[i]["position"]}
for i in top_indices]
return {
"sources": sources,
"user_query": request.prompt,
"assistant_response": answer,
"file_source": f"supabase://{bucket_name}/{file_path}"
}
except HTTPException:
raise
except Exception as e:
logger.exception("Error occurred in RAG process")
raise HTTPException(status_code=500, detail=f"An error occurred: {str(e)}")
@app.post("/extract_links")
async def extract_links(url: URL):
"""Extract unique links from a given URL"""
def extract_unique_links(url_string, max_retries=3, timeout=30):
for attempt in range(max_retries):
try:
headers = {'User-Agent': random.choice(user_agents)}
response = requests.get(url_string, headers=headers, timeout=timeout)
response.raise_for_status()
soup = BeautifulSoup(response.text, 'html.parser')
base_url = urlparse(url_string)
base_url = f"{base_url.scheme}://{base_url.netloc}"
a_tags = soup.find_all('a', href=True)
links = []
for a in a_tags:
href = a.get('href')
full_url = urljoin(base_url, href)
links.append(full_url)
unique_links = list(dict.fromkeys(links))
unique_links.insert(0, url_string)
return unique_links
except requests.RequestException as e:
logger.warning(f"Attempt {attempt + 1} failed: {e}")
if attempt < max_retries - 1:
wait_time = 5 * (attempt + 1)
time.sleep(wait_time)
else:
logger.error(f"Failed to retrieve {url_string} after {max_retries} attempts.")
raise HTTPException(status_code=500, detail=f"Failed to retrieve {url_string} after {max_retries} attempts.")
return []
try:
unique_links = extract_unique_links(url.url)
return {"unique_links": unique_links}
except Exception as e:
logger.exception("Error in extract_links")
raise HTTPException(status_code=500, detail=f"Failed to extract links: {str(e)}")
@app.post("/extract_text")
async def extract_text(urls: List[str]):
"""Extract text content from multiple URLs"""
if not supabase:
raise HTTPException(status_code=500, detail="Supabase not configured")
output_file = "extracted_text.txt"
def upload_text_content(filename, content, bucket_name):
try:
file_content = content.encode('utf-8')
# Try to upload first
try:
response = supabase.storage.from_(bucket_name).upload(
path=filename,
file=file_content,
file_options={"content-type": "text/plain"}
)
logger.info(f"Text file uploaded successfully: {filename}")
return response
except Exception as upload_error:
# If upload fails (file exists), try to update
try:
response = supabase.storage.from_(bucket_name).update(
path=filename,
file=file_content,
file_options={"content-type": "text/plain"}
)
logger.info(f"Text file updated successfully: {filename}")
return response
except Exception as update_error:
logger.error(f"Error updating text content: {update_error}")
raise HTTPException(status_code=500, detail="Failed to save file to storage")
except Exception as e:
logger.error(f"Error with file operations: {e}")
raise HTTPException(status_code=500, detail="Failed to save file to storage")
def text_data_extractor(links):
extracted_texts = []
for link in links:
parsed_url = urlparse(link)
if not parsed_url.scheme:
logger.warning(f"Invalid URL: {link}")
continue
retries = 3
while retries > 0:
try:
headers = {'User-Agent': random.choice(user_agents)}
response = requests.get(link, headers=headers, timeout=30)
response.raise_for_status()
soup = BeautifulSoup(response.text, 'html.parser')
text = soup.get_text()
clean_text = ' '.join(text.split())
extracted_texts.append({"url": link, "text": clean_text})
break
except requests.RequestException as e:
retries -= 1
logger.warning(f"Retry {3 - retries} for {link} failed: {e}")
if retries > 0:
wait_time = 5 * (3 - retries)
time.sleep(wait_time)
if retries == 0:
extracted_texts.append({
"url": link,
"text": "Failed to retrieve text after multiple attempts."
})
return extracted_texts
try:
extracted_data = text_data_extractor(urls)
string_output = json.dumps(extracted_data, ensure_ascii=False, indent=2)
# Upload to Supabase
upload_text_content(output_file, string_output, "url-2-ans-bucket")
return {"extracted_data": extracted_data, "file_saved": output_file}
except Exception as e:
logger.exception("Error in extract_text")
raise HTTPException(status_code=500, detail=f"Failed to extract text: {str(e)}")
# Main execution
if __name__ == "__main__":
# Run the FastAPI app
uvicorn.run(
"main_api:app",
host="0.0.0.0",
port=8000,
reload=False, # Disable reload for production
access_log=True
)