Yijia-Xiao cf960d6382 refactor: one name per decision in the state and its log ...

- the Research Manager's plan and the Portfolio Manager's decision live only in investment_plan and final_trade_decision; the debate states no longer copy them as judge_decision
- the saved state log writes the Trader's plan as trader_investment_plan, its state key
- the report tree, the final display and the live CLI read the decisions from those fields

2026-09-24 19:38:41 +00:00

..

agents

refactor: one name per decision in the state and its log

2026-09-24 19:38:41 +00:00

dataflows

chore: remove code nothing reads

2026-09-24 19:38:41 +00:00

graph

refactor: one name per decision in the state and its log

2026-09-24 19:38:41 +00:00

llm_clients

fix(llm): treat deepseek-flash as a thinking model (#1389)

2026-09-24 11:45:29 -07:00

memory

docs: call it the memory log throughout

2026-09-24 19:38:41 +00:00

__init__.py

chore: remove code nothing uses

2026-09-24 04:31:05 +00:00

backtest.py

docs: call it the memory log throughout

2026-09-24 19:38:41 +00:00

default_config.py

feat(config): benchmark Taiwan, Korea, Singapore and the main European exchanges (#1392)

2026-09-24 18:45:44 +00:00

portfolio.py

feat(graph): accept the caller's portfolio as run input (#1304, #1166)

2026-09-16 20:19:49 +00:00

reporting.py

refactor: one name per decision in the state and its log

2026-09-24 19:38:41 +00:00