components/generators/daily_feed.py

import os
import sys
import json
import requests

sys.path.append(os.path.abspath(os.path.join(os.path.dirname(__file__), "..")))

import redis
from typing import List, Dict
from llama_index.core import VectorStoreIndex
from llama_index.core.query_engine import RetrieverQueryEngine
from llama_index.core.schema import Document
from llama_index.core.settings import Settings

# ✅ Disable OpenAI fallback
Settings.llm = None

# 🔐 Environment variables
REDIS_URL = os.environ.get("UPSTASH_REDIS_URL", "redis://localhost:6379")
REDIS_KEY = os.environ.get("UPSTASH_REDIS_TOKEN")
MISTRAL_URL = os.environ.get("MISTRAL_URL")       # Hugging Face endpoint
HF_TOKEN = os.environ.get("HF_TOKEN")             # Hugging Face token

# ✅ Redis client
redis_client = redis.Redis.from_url(REDIS_URL, decode_responses=True)

# 📰 Topics to summarize
TOPICS = ["India news", "World news", "Tech news", "Finance news", "Sports news"]

# ✍️ Build Mistral prompt
def build_prompt(content: str, topic: str) -> str:
    return (
        f"You are a news summarizer. Summarize the following content in 25-30 words. "
        f"Make it engaging and informative. Include appropriate emojis. Topic: {topic}\n\n{content}"
    )

# 🧠 Send prompt to Mistral
def call_mistral(prompt: str) -> str:
    if not prompt or len(prompt.strip()) < 50:
        print(f"⚠️ Skipping empty or invalid prompt:\n{prompt}\n")
        return None

    headers = {
        "Authorization": f"Bearer {HF_TOKEN}",
        "Content-Type": "application/json"
    }
    payload = {
        "inputs": [{"role": "user", "content": prompt}]
    }

    print(f"\n📤 Prompt sent to Mistral:\n{prompt[:300]}...\n")  # show a snippet for debugging

    try:
        response = requests.post(MISTRAL_URL, headers=headers, json=payload, timeout=20)
        response.raise_for_status()
        return response.json()["outputs"][0]["content"].strip()
    except Exception as e:
        print(f"⚠️ Mistral error: {e}")
        return None

# ✂️ Generate summaries per topic
def summarize_topic(docs: List[str], topic: str) -> List[Dict]:
    feed = []

    for i, doc in enumerate(docs[:5]):
        if not doc or len(doc.strip()) < 200:
            print(f"⚠️ Skipped short/empty doc {i+1} for '{topic}'\n")
            continue

        print(f"📄 Doc {i+1} preview:\n{doc[:300]}...\n")
        prompt = build_prompt(doc, topic)
        summary = call_mistral(prompt)

        if summary:
            feed.append({
                "summary": summary,
                "image_url": "https://source.unsplash.com/800x600/?news",
                "article_link": "https://google.com/search?q=" + topic.replace(" ", "+")
            })

    return feed

# 🔁 Full pipeline
def generate_and_cache_daily_feed(documents: List[Document]):
    index = VectorStoreIndex.from_documents(documents)
    retriever = index.as_retriever()
    query_engine = RetrieverQueryEngine(retriever=retriever)

    final_feed = []

    for topic in TOPICS:
        print(f"\n🔍 Generating for: {topic}")
        response = query_engine.query(topic)
        docs = [str(node.get_content()) for node in response.source_nodes]

        topic_feed = summarize_topic(docs, topic)
        final_feed.append({
            "topic": topic.lower().replace(" news", ""),
            "feed": topic_feed
        })

    # 💾 Cache feed to Redis
    redis_client.set(REDIS_KEY, json.dumps(final_feed, ensure_ascii=False))
    print(f"✅ Cached daily feed under key '{REDIS_KEY}'")
    return final_feed

# 📦 For API access
def get_cached_daily_feed():
    cached = redis_client.get(REDIS_KEY)
    return json.loads(cached) if cached else []