deerflow2/tests/unit
Willem Jiang 2e010a4619
feat: add analysis step type for non-code reasoning tasks (#677) (#723)
Add a new "analysis" step type to handle reasoning and synthesis tasks
that don't require code execution, addressing the concern that routing
all non-search tasks to the coder agent was inappropriate.

Changes:
- Add ANALYSIS enum value to StepType in planner_model.py
- Create analyst_node for pure LLM reasoning without tools
- Update graph routing to route analysis steps to analyst agent
- Add analyst agent to AGENT_LLM_MAP configuration
- Create analyst prompts (English and Chinese)
- Update planner prompts with guidance on choosing between
  analysis (reasoning/synthesis) and processing (code execution)
- Change default step_type inference from "processing" to "analysis"
  when need_search=false

Co-authored-by: Willem Jiang <143703838+willem-bd@users.noreply.github.com>
2025-11-29 09:46:55 +08:00
..
agents security: add log injection attack prevention with input sanitization (#667) 2025-10-27 20:57:23 +08:00
checkpoint feat: support dify in rag module (#550) 2025-09-16 20:30:45 +08:00
config feat: 1. replace black with ruff for fomatting and sort import (#489) 2025-08-17 22:57:23 +08:00
crawler fix: the crawling error when encountering PDF URLs (#707) 2025-11-25 09:24:52 +08:00
graph feat: add analysis step type for non-code reasoning tasks (#677) (#723) 2025-11-29 09:46:55 +08:00
llms feat: support dify in rag module (#550) 2025-09-16 20:30:45 +08:00
prompt_enhancer Upgrade langchain version to 1.x (#720) 2025-11-28 22:09:13 +08:00
rag feat: Qdrant Vector Search Support (#684) 2025-11-11 19:35:00 +08:00
server security: add log injection attack prevention with input sanitization (#667) 2025-10-27 20:57:23 +08:00
tools Upgrade langchain version to 1.x (#720) 2025-11-28 22:09:13 +08:00
utils security: add log injection attack prevention with input sanitization (#667) 2025-10-27 20:57:23 +08:00