backend/agent · 4b7fa72d5fa10b7bb351355e1016525dc6d0a2c2 · Vũ Hoàng Anh / chatbot-canifa-feedback

refactor(lead-stage): clean 2-agent pipeline - no hardcoded heuristics · 4b7fa72d

Vũ Hoàng Anh authored Apr 13, 2026

- Classifier (1 LLM call): classify stage via STAGE_JSON + call tools
- Stylist (1 LLM call): format response with optional stage context
- Removed _TOOL_STAGE_MAP hardcoded heuristic (stage is optional)
- Added re import for STAGE_JSON regex parsing
- Cleaned up debug logging

4b7fa72d

Name	Last commit	Last update
..
diagram_agent		Loading commit data...
image_search_agent		Loading commit data...
lead_stage_agent		Loading commit data...
nodes		Loading commit data...
prompt_module		Loading commit data...
reaction_agent		Loading commit data...
report_agent		Loading commit data...
simulate_agent		Loading commit data...
sku_search_agent		Loading commit data...
sql_agent		Loading commit data...
store_search_agent		Loading commit data...
tag_search_agent		Loading commit data...
tool_prompts		Loading commit data...
tools		Loading commit data...
__init__.py		Loading commit data...
controller.py		Loading commit data...
controller_helpers.py		Loading commit data...
graph.py		Loading commit data...
helper.py		Loading commit data...
mock_controller.py		Loading commit data...
mock_graph.py		Loading commit data...
models.py		Loading commit data...
persona_generator.py		Loading commit data...
prompt_utils.py		Loading commit data...
streaming_callback.py		Loading commit data...
structured_models.py		Loading commit data...
system_prompt.txt		Loading commit data...
text_to_sql.py		Loading commit data...