run-llama / workflows-py / 21290367662
91%

DEFAULT BRANCH: main
Ran
Jobs 4
Files 145
Run time 1min
Badge
Embed ▾
README BADGES
x

If you need to use a raster PNG badge, change the '.svg' to '.png' in the link

Markdown

Textile

RDoc

HTML

Rst
coverage: 91.333% (+2.6%) from 88.684%
21290367662

push

github

web-flow 
add llama_index integration tests (#293)

* chore: add llama_index to gitignore

Added cloned reference repository llama_index/ to gitignore to avoid
tracking the large external repository.

* feat: add llama-index-integration-tests package

Add a new internal (not published) package for running integration tests
against llama-index-core workflow abstractions. This ensures that updates
to the workflows package don't break the agent classes in llama-index-core.

Includes tests for:
- FunctionAgent: basic execution, tools, max iterations, context handling
- ReActAgent: basic execution, retry on parse errors, memory persistence
- AgentWorkflow: multi-agent execution, handoff, memory, event streaming

Uses a custom MockFunctionCallingLLM to mock LLM responses without external
dependencies, ensuring tests run reliably.

* refactor: use llama-index-core MockFunctionCallingLLM

- Update llama-index-core dependency to >= 0.14.13
- Remove custom mock_llm.py, use built-in MockFunctionCallingLLM
- Enable test_early_stopping_method_generate (no longer needs skip)

All 12 integration tests now pass.

* fix: improve integration tests package

- Update basedpyright config to Python 3.13
- Fix ruff format/check issues (sorted imports, removed unused imports)
- Remove unused conftest.py fixture
- Add workflow feature tests:
  - wait_for_event inside tool functions (HITL pattern)
  - Context state management via ctx.store
  - Pickle serialization for workflow pause/resume
  - Multiple sequential tool calls with shared state

All 16 tests pass.

* refactor: reorganize integration tests by workflow feature

- Organize tests by workflow capability rather than agent type:
  - test_context_store.py: ctx.store.get/set operations
  - test_event_streaming.py: stream_events() and event types
  - test_human_in_the_loop.py: wait_for_event, serialization
  - test_error_handling.py: max_iterations, early stopping

- Add pytest parameterization for FunctionAgent vs ReActAgent where ... (continued)

11213 of 12277 relevant lines covered (91.33%)

1.02 hits per line

Jobs
ID Job ID Ran Files Coverage
1 llama-index-integration-tests - 21290367662.1 34
81.48
 GitHub Action Run
2 llama-index-utils-workflow - 21290367662.2 33
68.65
 GitHub Action Run
3 workflows-dev - 21290367662.3 8
78.59
 GitHub Action Run
4 llama-index-workflows - 21290367662.4 99
95.4
 GitHub Action Run
Source Files on build 21290367662