This repository has been archived on 2026-01-12. You can view files and clone it, but cannot push or open issues or pull requests.
Files
MeanRevisionEA/oi_scraper/README.md
Kunthawat Greethong 28a4546cd8 feat(oi): add open interest scraper module
add new oi_scraper directory for collecting open interest data
and update the main EA to integrate with the scraper functionality
2026-01-04 17:35:14 +07:00

178 lines
4.7 KiB
Markdown

# CME OI Scraper
Python scraper to pull Open Interest data from CME Group QuikStrike and current gold price from investing.com.
## What It Extracts
1. **OI Levels (from CME QuikStrike):**
- Top 3 CALL strikes by OI volume
- Top 3 PUT strikes by OI volume
2. **Gold Price (from investing.com):**
- Current gold futures price (e.g., 4345.50)
## Prerequisites
- Python 3.9+
- CME Group QuikStrike account with login credentials
## Installation
1. Copy environment variables:
```bash
cp .env.example .env
```
2. Edit `.env` and add your CME credentials:
```bash
CME_USERNAME=your_username
CME_PASSWORD=your_password
```
3. Install dependencies:
```bash
pip install -r requirements.txt
playwright install chromium
```
## Usage
### Basic Scraping
```bash
python main.py
```
This will:
- Login to CME QuikStrike
- Navigate to OI Heatmap
- Extract top 3 CALL and PUT strikes by OI volume
- Scrape current gold price from investing.com
- Export to `oi_data.csv`
### Session Persistence
The scraper automatically saves your login session to `cookies.json`. This means:
- **First run**: Logs in with your credentials, saves cookies
- **Subsequent runs**: Uses saved cookies if session is still valid
- **Session expired**: Automatically logs in again and saves new cookies
Benefits for scheduled runs:
- Faster execution (skips login when session is valid)
- Reduces login attempts to CME servers
- CME sessions typically last several days/weeks
To force a fresh login, delete `cookies.json`:
```bash
rm cookies.json
```
### Output Format
The CSV output is compatible with the EA's `LoadOIFromCSV()` and `LoadFuturePriceFromCSV()` functions:
```csv
Type,Strike,OI
CALL,4345,155398
CALL,4350,229137
CALL,4360,90649
PUT,4300,227936
PUT,4290,270135
PUT,4280,65839
[Price]
FuturePrice,4345.50
```
**Note:** The `[Price]` section contains the current gold futures price scraped from investing.com. The EA reads this value for Delta calculation.
## Configuration
Edit `.env` to customize:
- `PRODUCT_URL` - QuikStrike product page URL (requires login)
- `TOP_N_STRIKES` - Number of top strikes to export (default: 3)
- `HEADLESS` - Run browser in headless mode (default: false for debugging)
- `CSV_OUTPUT_PATH` - Output CSV file path
- `TIMEOUT_SECONDS` - Page load timeout
### Available Products
**Gold (XAUUSD/COMEX Gold - OG|GC):**
```
PRODUCT_URL=https://cmegroup.quikstrike.net/User/QuikStrikeView.aspx?pid=40&viewitemid=IntegratedOpenInterestTool
```
**Silver:**
```
PRODUCT_URL=https://cmegroup.quikstrike.net/User/QuikStrikeView.aspx?pid=41&viewitemid=IntegratedOpenInterestTool
```
**SOFR (3M SOFR):**
```
PRODUCT_URL=https://cmegroup.quikstrike.net/User/QuikStrikeView.aspx?pid=476&viewitemid=IntegratedOpenInterestTool
```
**Note:** You must be logged in to access QuikStrike data. The scraper will automatically login using credentials from `.env`.
## Integration with EA
The EA reads OI data from CSV when `InpOISource = OI_SOURCE_CSV_FILE`.
Place the generated `oi_data.csv` in MetaTrader's `MQL5/Files` directory.
## Scheduling
Use cron or Windows Task Scheduler to run periodically:
```bash
# Run every hour
0 * * * * cd /path/to/oi_scraper && python main.py
```
## Troubleshooting
**Login fails:**
- Verify credentials in `.env`
- Check if CME requires 2FA
- Set `HEADLESS=false` to see what's happening
- Check screenshots: `login_failed.png`, `login_error.png`, `login_success.png`
**No data extracted:**
- Check if table structure changed
- Increase `TIMEOUT_SECONDS`
- Check logs for detailed errors
- Screenshot saved as `login_debug.png` or `login_failed.png`
**Login page selectors changed:**
- If the scraper can't find username/password inputs, CME may have updated their login page
- Update the selectors in `login_to_cme()` function in `main.py`:
```python
# Example: update to match current CME login form
page.fill('input[id="username"]', CME_USERNAME)
page.fill('input[id="password"]', CME_PASSWORD)
page.click('button[type="submit"]')
```
**Browser issues:**
- Install Chromium dependencies: `playwright install chromium`
- Try different browser: Change `p.chromium.launch()` to `p.firefox.launch()`
## Notes
- The scraper targets the OI Heatmap table structure
- Only exports top N strikes by OI volume
- Login session is not persisted (login each run)
- Cookies could be saved for faster subsequent runs
### Finding Product IDs
To find product IDs for other instruments:
1. Visit https://www.cmegroup.com/tools-information/quikstrike/open-interest-heatmap.html
2. Login to your CME account
3. Select a product from the "Products" menu
4. The URL will update with the `pid` parameter
5. Copy that URL to your `.env` file
Example: `https://www.cmegroup.com/tools-information/quikstrike/open-interest-heatmap.html?pid=40` (Gold)