Daily Stock Analysis Powered by Gemini AI

Welcome to the daily_stock_analysis project – a fully‑featured, open‑source solution for automated A‑Share analysis that runs for free on GitHub Actions or Docker. Built with Python 3.10+, it integrates multiple data providers, Gemini (Google AI Studio) for natural‑language market insights, and versatile notification channels.

What This Tool Offers

Tech Stack Overview

• Python 3.10+ with pyproject.toml / requirements.txt.
• Data Sources: AkShare (free), Tushare, Baostock, YFinance.
• News Search: Tavily (free key) or SerpAPI as fallback.
• AI Backend: Google Gemini (default), with optional OpenAI‑compatible APIs like DeepSeek, Claude, or locally hosted models.
• Deployment: GitHub Actions (.github/workflows/daily_analysis.yml) or Docker Compose with a single docker-compose up -d.
• Scheduler: Simple cron logic that triggers the analysis each weekday at 18:00 Beijing time.

Quick Start – GitHub Actions (Zero‑Cost)

1. Fork the Repository – Click Fork in the upper‑right corner.
2. Create Required Secrets – Navigate to Settings → Secrets → Actions and add:

   GEMINI_API_KEY          # (required)  Gemini key from Google AI Studio
   # OR
   OPENAI_API_KEY          # optional, if you prefer an OpenAI‑compatible endpoint
   OPENAI_BASE_URL         # optional – e.g., https://api.deepseek.com/v1
   OPENAI_MODEL            # optional – e.g., deepseek-chat
   WECHAT_WEBHOOK_URL      # optional – Enterprise WeChat
   FEISHU_WEBHOOK_URL      # optional – Feishu
   TELEGRAM_BOT_TOKEN      # optional – Telegram bot token
   TELEGRAM_CHAT_ID        # optional – Telegram chat id
   EMAIL_SENDER            # optional – sender email
   EMAIL_PASSWORD          # optional – app password
   EMAIL_RECEIVERS         # optional – comma‑separated list
   CUSTOM_WEBHOOK_URLS     # optional – any POST JSON webhook
   FEISHU_APP_ID           # optional – Feishu app id
   FEISHU_APP_SECRET       # optional
   FEISHU_FOLDER_TOKEN     # optional – Feishu folder key
   STOCK_LIST              # (required) list of tickers, e.g., 600519,300750,002594
   TAVILY_API_KEYS         # (recommended) for news search
   SERPAPI_API_KEYS        # optional 
   TUSHARE_TOKEN           # optional
   

3. Enable the Workflow – Go to the Actions tab, click I understand my workflows, go ahead and enable them.
4. Test Immediately – Select the “每日股票分析” workflow and click Run workflow.
5. Celebrate – The workflow will build a Docker image, fetch data, run the analysis, and post the results to your specified channels.

Quick Start – Local or Docker Deployment

Use this method if you prefer a local sandbox or want to tweak the code.

# Clone & install
git clone https://github.com/ZhuLinsen/daily_stock_analysis.git
cd daily_stock_analysis
pip install -r requirements.txt

# Copy env example
cp .env.example .env
nano .env   # fill in values

# Run a full analysis
python main.py

# Only market review
python main.py --market-review

# Run scheduled task (cron inside the container)
python main.py --schedule

For Docker:

# Create config file
cp .env.example .env
nano .env

# Launch services
docker-compose up -d

# Tail logs
docker-compose logs -f

How the Analysis Works

1. Data Fetch – Stock price, volume, MA5/MA10/MA20, volume distribution, and news headlines are sourced from the configured providers.
2. AI Inference – The analyzer.py module feeds cleaned indicators and news into Gemini (or the selected OpenAI‑compatible model) and receives a concise decision summary, target and stop‑loss prices, and a compliance checklist.
3. Market‑Review – market_analyzer.py aggregates daily index moves, sector performance, and capital flows.
4. Notification – notification.py formats the dashboard and market review into Markdown or interactive cards and posts to every enabled channel.
5. Scheduling – scheduler.py and the GitHub Actions workflow ensure the job runs once per weekday at 18:00 local time.

Customization Ideas

Roadmap Highlights

• Custom Webhook & Push Services – Support for Pushover, Discord, Slack, and proprietary APIs.
• Cross‑Market Expansion – Include HK, US, and EU market data.
• Historical Backtesting – Run simulations against past data.
• Web Dashboard – Real‑time analytics with live charts.
• Community‑Driven Enhancements – Contributions for new data sources and models are very welcome.

How to Contribute

• Fork and create a feature/bug‑fix branch.
• Run tests (pytest) and lint (ruff).
• Open a PR with a clear description and link to any relevant issue.
• Reviewers will check for documentation updates and consistency with the codebase.

Final Thoughts

daily_stock_analysis is a modern, low‑friction toolkit for anyone looking to add AI‑driven market insights to their daily workflow. With zero hosting costs and a plug‑and‑play architecture, you can start receiving market dashboards in your corporate chat or inbox in minutes. Grab the repo, set your secrets, and watch the tool auto‑run every weekday.

Happy analyzing!