📈 StockIQ – Indian Market Intelligence Platform

StockIQ is a high-performance, asynchronous market intelligence and analytics engine designed for the Indian equity markets (NSE/BSE). Built using Python 3.11+ and FastAPI, the platform processes real-time ticketing feeds, runs complex analytical screening workflows, and broadcasts streaming updates to consumers via WebSockets.

---

🚀 Core Architecture Features

• Asynchronous Lifespan Management: Managed database connection pools (PostgreSQL/TimescaleDB) and Redis layers without blocking the event loop.
• Dual Ingestion Pipelines: Background asyncio tasks concurrently manage streaming market tick data (Kite Connect API) and live news feeds with integrated sentiment engine hooks.
• WebSocket Streaming: Leverages Redis Pub/Sub backends to orchestrate multi-client, low-latency live price broadcasting.
• Data Layer Optimization: Implements time-series optimized schema structures via TimescaleDB alongside Redis data structures for sub-millisecond lookups.

---

📂 File-by-File Technical Directory Breakdown

Here is how the modules map directly into this architecture:

🎮 Entry Point & Routing

• main.py – The core conductor of the app. It initializes FastAPI, attaches performance middlewares (GZip compression and CORS rules), registers specific API/WebSocket routers, and handles the asynchronous safe startup/shutdown cycle of backend services.
• config.py – Centralized application configurations managed using environmental variables (e.g., database connection strings, API secrets, allowed CORS origins).

🔄 Ingestion & Streaming Layer

• market_data_ingestion.py – Connects to upstream market feeds (like Zerodha Kite Connect), processes raw ticker data, and writes the high-frequency stream directly to Redis.
• news_ingestion.py – Periodically fetches live Indian financial market news feeds and passes payload elements to the internal processing pipelines.
• ws.py – Handles full-duplex WebSocket connection lifecycles, pulling streaming tick events from Redis Pub/Sub to push instant updates directly to client dashboards.

🧠 Analytics & Logic Engines

• classifcation_engine.py – Houses machine learning or pattern-matching models (likely processing sentiment classification for incoming text feeds from news_ingestion.py).
• indicator_engine.py – Computes mathematical technical indicators (e.g., RSI, MACD, Bollinger Bands) dynamically on high-frequency market series data.
• analytics.py & analytics.R – A polyglot analytical bridge. Handles structural statistical quantitative analytics—utilizing R's powerful statistical libraries alongside Python to run portfolio optimization models or historical backtests.
• screener.py – Executes high-throughput filter workflows across thousands of instruments based on technical indicators and classification logic.

🌐 Endpoints & Entities

• market.py & news.py – FastAPI endpoints serving standardized REST API routes for historical market analytics, current system status, and categorized news elements.
• schemas.py – Pydantic validation models that enforce payload constraints, serialization, and explicit typing across data endpoints.
• schema.sql – The database blueprint establishing relational models and hyper-tables (TimescaleDB) optimized for massive financial time-series logging.