Files

cauvang32 9c180bdd89 Refactor OpenAI utilities and remove Python executor

- Removed the `analyze_data_file` function from tool definitions to streamline functionality.
- Enhanced the `execute_python_code` function description to clarify auto-installation of packages and file handling.
- Deleted the `python_executor.py` module to simplify the codebase and improve maintainability.
- Introduced a new `token_counter.py` module for efficient token counting for OpenAI API requests, including support for Discord image links and cost estimation.

2025-10-02 21:49:48 +07:00

2.0 KiB

Raw Permalink Blame History

Quick Reference: File Storage & Context Management

📁 File Storage TL;DR

Non-Images → Disk (/tmp/bot_code_interpreter/user_files/)
MongoDB → Only metadata (file_id, path, size, timestamps)
Images → Discord CDN links only
Expiration → 48 hours, auto-cleanup

🔢 Token Limits (config.py)

gpt-4o: 8000
gpt-4.1: 8000
o1/o3/o4: 4000
gpt-5: 4000
Default: 4000

🔄 Context Management

Strategy: Sliding window (like ChatGPT)

Keep: System prompt + recent messages
Group: User+Assistant pairs together
Trim: Oldest-first when over limit
No summarization: Zero extra API calls

Token Budget:

System: Always included
Conversation: 80% of available
Response: 20% reserved

📊 Key Improvements

Metric	Old	New	Improvement
DB Size (100 files)	200MB	50KB	99.97% ↓
Context Method	Fixed limits	Model-specific	Configurable
Pairing	None	User+Asst	Coherent
API Calls	Extra for summary	None	Free

💻 Code Examples

Upload File

result = await upload_discord_attachment(attachment, user_id, db)
# Returns: {"file_id": "...", "file_path": "..."}

Use in Code

df = load_file('file_id')  # Auto-loads from disk
df.to_csv('output.csv')    # Auto-captured

Generated Files

result["generated_files"] = [
    {
        "filename": "chart.png",
        "data": b"...",
        "type": "image",
        "file_id": "..."
    }
]

⚙️ Configuration

Edit src/config/config.py:

MODEL_TOKEN_LIMITS = {
    "openai/gpt-4.1": 8000,  # Adjust here
}

🔍 Monitoring

# Log output shows:
Sliding window trim: 45 → 28 messages (17 removed, ~3200/4000 tokens)
Saved file sales.csv for user 123: file_id

🚨 Common Issues

File expired: Re-upload (48h limit) Context too large: Automatic trim Disk full: Check cleanup task

📖 Full Documentation

See: docs/FILE_STORAGE_AND_CONTEXT_MANAGEMENT.md

2.0 KiB Raw Permalink Blame History