feat: expand integration tests with new input types and mode transitions by 0xbyt4 · Pull Request #2274 · OpenMind/OM1

0xbyt4 · 2026-02-15T19:40:10Z

Summary

Add integration tests for 5 new input types: ASR (greeting, question), battery (low), odometry (moving), GPS (outdoor location)
Add multi-input combination test: VLM + ASR simultaneous processing
Add 4 mode transition tests: input-triggered, negative (no transition), time-based timeout, and cooldown prevention
Add mock implementations: MockGoogleASR, MockUnitreeGo2Battery, MockUnitreeGo2Odom, MockGps with data providers (MockTextProvider, MockStateProvider)
Fix upstream test runner issues: typo in fixture name, DRY violations (normalize_expected_value, _extract_emotion, _detect_input_type), dead code removal, mutable default argument fix, rating_description deduplication

New Test Cases

Test Case	Input Type	What It Tests
asr_greeting_test	ASR	Robot responds to voice greeting
asr_question_test	ASR	Robot responds to voice question
battery_low_test	Battery + VLM	Low battery triggers warning behavior
odometry_awareness_test	Odometry + VLM	Robot aware of movement state
gps_location_test	GPS + VLM	Robot aware of geographic location
multi_input_vlm_asr_test	VLM + ASR	Combined visual and audio processing
mode_transition_test	ASR (mode)	Input-triggered calm -> alert transition
mode_no_transition_test	ASR (mode)	No transition when keywords absent
mode_time_based_test	ASR (mode)	Time-based patrol -> idle transition
mode_cooldown_test	ASR (mode)	Cooldown prevents repeated transitions

File Structure

tests/integration/
├── test_case_runner.py                          # MODIFIED - new runners + refactoring
├── data/
│   ├── images/
│   │   ├── indoor_1.jpg                         # existing
│   │   └── indoor_2.jpg                         # existing
│   ├── lidar/
│   │   └── sample_scan.json                     # existing
│   ├── asr/                                     # NEW directory
│   │   ├── emergency.json                       # new
│   │   ├── greeting.json                        # new
│   │   └── weather_question.json                # new
│   ├── gps/                                     # NEW directory
│   │   └── outdoor_location.json                # new
│   ├── state/                                   # NEW directory
│   │   ├── battery_low.json                     # new
│   │   └── odometry_moving.json                 # new
│   └── test_cases/
│       ├── coco_indoor_test.json5               # existing
│       ├── gemini_indoor_test.json5             # existing
│       ├── open_ai_indoor_test.json5            # existing
│       ├── rplidar_test.json5                   # existing
│       ├── vila_indoor_test.json5               # existing
│       ├── asr_greeting_test.json5              # new
│       ├── asr_question_test.json5              # new
│       ├── battery_low_test.json5               # new
│       ├── gps_location_test.json5              # new
│       ├── odometry_awareness_test.json5        # new
│       ├── multi_input_vlm_asr_test.json5       # new
│       ├── mode_transition_test.json5           # new
│       ├── mode_no_transition_test.json5        # new
│       ├── mode_time_based_test.json5           # new
│       └── mode_cooldown_test.json5             # new
└── mock_inputs/
    ├── __init__.py                              # existing
    ├── input_registry.py                        # MODIFIED - dict-based restoration
    ├── mock_vlm_coco.py                         # existing
    ├── mock_vlm_gemini.py                       # existing
    ├── mock_vlm_openai.py                       # existing
    ├── mock_vlm_vila.py                         # existing
    ├── mock_unitree_go2_rplidar.py              # existing
    ├── mock_google_asr.py                       # new
    ├── mock_battery.py                          # new
    ├── mock_gps.py                              # new
    ├── mock_odometry.py                         # new
    └── data_providers/
        ├── __init__.py                          # new
        ├── mock_image_provider.py               # existing
        ├── mock_lidar_scan_provider.py          # existing
        ├── mock_text_provider.py                # new
        └── mock_state_provider.py               # new

Add ASR, battery, odometry, GPS mock inputs and test cases. Add mode transition integration test verifying input-triggered transitions through the full cortex runtime pipeline.

Replace no-op transition callback with real _on_mode_transition. Only mock _start_orchestrators to prevent infinite loops. This tests the actual transition: stop orchestrators, reinitialize mode components, and verify new config is loaded.

Add three new integration test scenarios for mode transitions: - Negative case: verify no transition when input lacks trigger keywords - Time-based: verify timeout-triggered mode transition - Cooldown: verify cooldown prevents repeated transitions Also fix missing timeout_seconds in build_multi_mode_config.

Filter cooldown and time-based configs in discover_mode_transition_test_cases() instead of skipping them inside test_mode_transition(). Discovery should only return configs that will actually be run.

Keywords are checked against the fused prompt (raw_response), not the LLM output. Original keywords included words not present in the prompt (e.g. "forecast", "temperature", "power", "location"). Replaced with keywords verified to exist in system_prompt or mock input output.

- Move `import json` to top-level (maintainer convention) - Remove unused MockTextProvider.text_count, remaining_texts properties - Remove unused MockStateProvider.reset_all() method - Remove unused load_test_texts() wrapper function - Remove redundant section comments in mock_state_provider.py

The -k flag filters by test ID, not file name. Parametrized test IDs use numeric indices (test_case_path0, test_case_path1) so -k "asr" or -k "battery" would match nothing.

- Fix typo: mock_confige_provider -> mock_config_provider - Extract normalize_expected_value, _detect_input_type, _extract_emotion to module-level to eliminate duplicate definitions - Remove dead code: pass inside try/except that can never raise - Collapse 3 identical rating_description branches into one - Add GPS-only case to get_movement_types_for_config - Fix mutable default argument (= [] -> = None) in mock_llm_ask - Replace if/elif chain with dict-based class restoration in input_registry

Match upstream convention where empty config objects are omitted from agent_inputs in JSON5 config files.

… conversation provider mock - Fix load_test_state_data to collect all entries before loading, preventing data overwrite when multiple files are specified - Extract _setup_mode_transition_mocks and _cleanup_mode_transition_test helpers to eliminate repeated setup/teardown code - Add _MockConversationProvider to MockGoogleASR to match upstream formatted_latest_buffer behavior (store_user_message call) - Remove unused battery_full.json (no test uses it, battery mock only reports for low levels)

codecov · 2026-02-15T19:44:45Z

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ All tests successful. No failed tests found.

📢 Thoughts on this report? Let us know!

openminddev · 2026-02-15T22:09:04Z

Great PR! Thank you.

0xbyt4 added 15 commits February 15, 2026 17:28

feat: expand integration tests with new inputs and mode transition

906905c

Add ASR, battery, odometry, GPS mock inputs and test cases. Add mode transition integration test verifying input-triggered transitions through the full cortex runtime pipeline.

docs: add mode transition test documentation to integration README

a4e8113

fix: move skip logic from test to discovery function

49b8d7e

Filter cooldown and time-based configs in discover_mode_transition_test_cases() instead of skipping them inside test_mode_transition(). Discovery should only return configs that will actually be run.

fix: remove dead GPS base_dir branch in load_test_state_data

5fd03c4

style: apply black formatting

2431415

fix: remove incorrect test filter commands from README

d6d45ed

The -k flag filters by test ID, not file name. Parametrized test IDs use numeric indices (test_case_path0, test_case_path1) so -k "asr" or -k "battery" would match nothing.

style: remove unnecessary config: {} from test case configs

64848fe

Match upstream convention where empty config objects are omitted from agent_inputs in JSON5 config files.

style: remove config: {} from README mode transition example

89e2e1c

chore: revert README to upstream, remove redundant comments

02c44e8

0xbyt4 requested a review from a team as a code owner February 15, 2026 19:40

github-actions bot added python Python code tests Test files config Configuration files labels Feb 15, 2026

openminddev merged commit 4003f32 into OpenMind:main Feb 15, 2026
6 checks passed

0xbyt4 mentioned this pull request Feb 16, 2026

docs: update integration test README to reflect current test coverage #2290

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

feat: expand integration tests with new input types and mode transitions#2274

feat: expand integration tests with new input types and mode transitions#2274
openminddev merged 15 commits intoOpenMind:mainfrom
0xbyt4:feature/integration-test-expansion

0xbyt4 commented Feb 15, 2026

Uh oh!

codecov bot commented Feb 15, 2026 •

edited

Loading

Uh oh!

openminddev commented Feb 15, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

0xbyt4 commented Feb 15, 2026

Summary

New Test Cases

File Structure

Uh oh!

codecov bot commented Feb 15, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

openminddev commented Feb 15, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

codecov bot commented Feb 15, 2026 •

edited

Loading