github-mirrors/tradingagents
1
0
Fork 0
mirror of https://github.com/TauricResearch/TradingAgents.git synced 2026-09-27 06:56:39 +03:00
Code Issues Packages Projects Releases Wiki Activity

docs: call it the memory log throughout

Browse Source 
- docstrings, comments, messages and the README section use "memory log", matching TradingMemoryLog and memory_log_path
- the write, read and settle paths are named for what they do, not by design phase
This commit is contained in:
Yijia-Xiao
2026-09-24 19:38:41 +00:00
parent c50420fc5c
commit 5ac5786d0d
9 changed files with 25 additions and 23 deletions
Download Patch File Download Diff File Expand all files Collapse all files
README.md
+4 -4
View File
@@ -304,9 +304,9 @@ An empty `positions` list means a flat book, which is different from passing not
TradingAgents persists two kinds of state across runs.
### Decision log
### Memory log
The decision log is always on. Each completed run appends its decision to `~/.tradingagents/memory/trading_memory.md`. On the next run for the same ticker, TradingAgents fetches the realised return (raw, and alpha against the instrument's regional benchmark), generates a one-paragraph reflection, and injects the most recent same-ticker decisions plus recent cross-ticker lessons into the Portfolio Manager prompt, so each analysis carries forward what worked and what didn't.
The memory log is always on. Each completed run appends its decision to `~/.tradingagents/memory/trading_memory.md`. On the next run for the same ticker, TradingAgents fetches the realised return (raw, and alpha against the instrument's regional benchmark), generates a one-paragraph reflection, and injects the most recent same-ticker decisions plus recent cross-ticker lessons into the Portfolio Manager prompt, so each analysis carries forward what worked and what didn't.
Override the path with `TRADINGAGENTS_MEMORY_LOG_PATH`.
@@ -330,7 +330,7 @@ _, decision = ta.propagate("NVDA", "2026-09-01")
## Evaluating decisions over time
One run gives one decision, which cannot tell you whether the system decides well. `run_backtest` runs the same pipeline over a grid of tickers and dates, writes to a decision log of its own, and scores the decisions whose holding window has since traded.
One run gives one decision, which cannot tell you whether the system decides well. `run_backtest` runs the same pipeline over a grid of tickers and dates, writes to a memory log of its own, and scores the decisions whose holding window has since traded.
```python
from tradingagents.backtest import iter_grid, run_backtest, summarize
@@ -346,7 +346,7 @@ From the CLI:
tradingagents backtest NVDA,AAPL --start 2026-06-01 --end 2026-08-01 --every 7
```
Each cell is scored on realized alpha against the instrument's regional benchmark, grouped by rating. Your own decision log is never written to, and re-running the same grid with `run_id=result.run_id` skips the cells that already ran, so an interrupted sweep continues where it stopped.
Each cell is scored on realized alpha against the instrument's regional benchmark, grouped by rating. Your own memory log is never written to, and re-running the same grid with `run_id=result.run_id` skips the cells that already ran, so an interrupted sweep continues where it stopped.
## Reproducibility