Need to bring the latest change from main to trajectory_refactor#222
Merged
nianjunz merged 50 commits intotrajectory_refactorfrom Mar 19, 2026
Merged
Need to bring the latest change from main to trajectory_refactor#222nianjunz merged 50 commits intotrajectory_refactorfrom
nianjunz merged 50 commits intotrajectory_refactorfrom
Conversation
issue #133 Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
issue #133 Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Adding a pointer to the utterance guideline and ground truth guideline
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
…ation - Add src/servers/wo/main.py: FastMCP server with 3 tools (list_equipment, get_work_orders, summarize_work_orders) backed by the sample CSV dataset - Add src/servers/wo/tests/: conftest with mock_df fixture and 12 unit tests covering all tools; integration tests gated by requires_wo_data marker - Add src/tmp/agent_hive/tools/wo.py: agent_hive tool registration module (mirrors fmsr.py / tsfm.py pattern) referencing external woagent package - Add src/tmp/agent_hive/agents/wo_agent.py: WorderOrderAgent class backed by ReactReflectXenAgent, matching the interface expected by run_track tests - Register wo-mcp-server CLI entry point in pyproject.toml Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…on_cbm Port all @validate_inputs functions from src/tmp/workorder_agent/event_forecasting/apis/api_implementation_cbm.py to FastMCP tools, replacing the earlier placeholder stub. Tools (8 total): - get_work_orders — all WOs for equipment in date range - get_preventive_work_orders — preventive WOs only - get_corrective_work_orders — corrective WOs only - get_events — all events (WO / alert / anomaly) - get_failure_codes — full failure code catalogue - get_work_order_distribution — counts per (primary, secondary) code pair - predict_next_work_order — Markov-chain transition probability - analyze_alert_to_failure — alert-rule → maintenance transition stats Key structural changes vs. original: - Equipment/DateRange/Date objects → flat string parameters (MCP-friendly) - importlib.resources.path() → configurable WO_DATA_DIR env var (defaults to src/tmp/assetopsbench/sample_data/) - All data loaded lazily, missing files degrade gracefully - Pickle-file return values → typed Pydantic BaseModel responses - analyze_alert_to_failure now returns the computed transition result instead of None - validate_inputs decorator removed (MCP enforces types) - Tests updated: 23 unit tests + 8 integration tests (requires_wo_data) Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
- Add src/couchdb/init_wo.py: loads all WO CSVs into a CouchDB `workorder` database with _dataset discriminator field and Mango indexes - Rewrite src/servers/wo/data.py: lazy couchdb3 connection, load() queries via Mango find instead of reading CSVs - Fix src/servers/wo/tools.py: use rule_id column (was rule) for alert_events - Update tests: mock_data patches servers.wo.tools.load directly; replace requires_wo_data marker with requires_couchdb; all 23 unit tests pass Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…ompose - Add src/couchdb/init_asset_data.py: loads IoT sensor JSON into CouchDB (decouples data loading from couchdb_setup.sh) - Simplify couchdb_setup.sh: only writes local.ini and starts CouchDB - Add db-init service to docker-compose.yaml: python:3.12-slim container that runs after couchdb is healthy, calling both init_asset_data.py and init_wo.py Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…ontainer Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…derAgent Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…check connectivity, fix history date validation order Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…ver and tests Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
CouchDB rejects document fields starting with '_' (reserved for _id, _rev, etc.). Rename the collection discriminator field from _dataset to dataset in both init_wo.py and data.py. Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Add wo-mcp-server to DEFAULT_SERVER_PATHS so WorkOrderAgent tools are available to plan-execute without requiring --server overrides. Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
…UCTIONS.md - Add WorkOrder query examples (work order count, PM vs CM split, alert-to-failure probability, multi-step distribution + prediction) - Clarify that plan-execute spawns MCP servers on-demand; no manual server startup required - Fix .env.public: replace COUCHDB_DBNAME with IOT_DBNAME + WO_DBNAME - Add src/tmp/ to .gitignore Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
Signed-off-by: ChathurangiShyalika <chathurangijks@gmail.com>
Signed-off-by: ChathurangiShyalika <chathurangijks@gmail.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
issue #133 Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Can be set by env variable `SCENARIO_CLIENT_TIMEOUT` Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
updates to scenario-client
issue #193 Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
Issue193 - replace mlflow with mlflow.MLflowClient in scenario server
Removing special characters
issue #201 Signed-off-by: John D Sheehan <john.d.sheehan@ie.ibm.com>
add build-date endpoint to scenario-server
adding templates
Update couchdb Docker tag to v3.5
Debian 12 (Bookworm) enforces PEP 668 and blocks system-wide pip installs without an explicit opt-in flag, causing ModuleNotFoundError for requests and pandas during CouchDB container initialisation. Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
refactor: implement Work Order Agent MCP server and agent_hive integration
Replace all uses of "agent" terminology with "server" in models, planner, executor, runner, CLI, and tests to match MCP architecture where servers expose tools and the plan-execute runner is the actual agent. Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
IoTAgent → iot, FMSRAgent → fmsr, TSFMAgent → tsfm, WorkOrderAgent → wo, Utilities → utilities. Updates FastMCP() names in all server main.py files, DEFAULT_SERVER_PATHS keys in executor.py, all test fixtures and plan strings, and INSTRUCTIONS.md. Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
When execute_step encounters a tool=none step that has prior-step
dependencies, it was returning the static expected_output description
instead of deriving the actual value from the dependency context.
Downstream steps depending on that result would receive the description
string (e.g. "The name and id of a chiller asset...") and fail to
resolve their {step_N} placeholders correctly.
Fix: for tool=none steps with dependencies present in context, call the
LLM with a short extraction prompt to derive the concrete value. Steps
with no dependencies retain the previous behaviour (return expected_output).
Signed-off-by: Shuxin Lin <linshuhsin@gmail.com>
A LLM Notebook revising the dependnecy for tutorial
…erver refactor: rename "agent" to "server" throughout workflow/ (#170)
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
6 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.