shoko/randebu
1
0
Fork 0
Code Issues 6 Pull Requests 1 Actions Packages Projects Releases Wiki Activity

fix: flatten strategy config schema to match engine expectations

Browse Source 
LLM was outputting nested params structure but engines expect flat fields.
This caused backtesting and simulation to never trigger any trades.

Changes:
- llm_connector.py: Update prompt to output flat condition structure
- crew.py: Update StrategyValidator to validate flat structure
- crew.py: Update StrategyExplainer to read flat structure

Fixes #25
This commit is contained in:
shokollm
2026-04-09 07:31:09 +00:00
parent 55b008d4e8
commit d81464b869
2 changed files with 27 additions and 34 deletions
Download Patch File Download Diff File Expand all files Collapse all files

21
src/backend/app/services/ai_agent/llm_connector.py
View File

@@ -61,9 +61,9 @@ class MiniMaxConnector:
system_prompt = """You are a trading strategy designer. Parse the user's natural language request into a JSON strategy_config object.
Supported conditions (MVP):
- price_drop: Token price drops by X% (requires: token, threshold_percent)
- price_rise: Token price rises by X% (requires: token, threshold_percent)
- volume_spike: Trading volume increases X% (requires: token, threshold_percent)
- price_drop: Token price drops by X% (requires: token, threshold)
- price_rise: Token price rises by X% (requires: token, threshold)
- volume_spike: Trading volume increases X% (requires: token, threshold)
- price_level: Price crosses above/below X (requires: token, price, direction)
Output ONLY valid JSON with this schema:
@@ -71,18 +71,17 @@ Output ONLY valid JSON with this schema:
"conditions": [
{
"type": "price_drop|price_rise|volume_spike|price_level",
"params": {
"token": "TOKEN_SYMBOL",
"threshold_percent": number, // for price_drop, price_rise, volume_spike
"price": number, // for price_level
"direction": "above|below" // for price_level
}
"token": "TOKEN_SYMBOL",
"chain": "bsc",
"threshold": number, // for price_drop, price_rise, volume_spike
"price": number, // for price_level
"direction": "above|below", // for price_level
"timeframe": "1h"
}
],
"actions": [
{
"type": "buy|sell|notify",
"params": {}
"type": "buy|sell|notify"
}
]
}