Functions — github.com/IBM/AssetOpsBench

hub / github.com/IBM/AssetOpsBench / functions

Functions947 in github.com/IBM/AssetOpsBench

⨍ Functions 947◇ Types & classes 171↳ Endpoints 13

↓ 286 callersMethodget

(self, doc_id: str)

src/servers/wo/couch.py:49

↓ 23 callersMethodget

(self, data_id: str)

src/servers/vibration/data_store.py:125

↓ 20 callersFunction_step

(n: int, deps: list[int] | None = None)

src/agent/tests/test_models.py:6

↓ 20 callersMethodfind

( self, selector: Dict[str, Any], *, fields: Optional[List[str]] = None,

src/servers/wo/couch.py:71

↓ 19 callersFunctioncall_tool

Helper: call an MCP tool and return the parsed JSON response.

src/servers/tsfm/tests/conftest.py:27

↓ 19 callersMethodfrom_raw

(cls, raw: dict)

src/evaluation/models.py:29

↓ 18 callersFunctioncall_tool

Helper: call an MCP tool and return parsed JSON response.

src/servers/vibration/tests/conftest.py:47

↓ 17 callersFunctionmake_persisted_record

()

src/evaluation/tests/conftest.py:30

↓ 16 callersFunctionparse_plan

Parse an LLM-generated plan string into a Plan object.

src/agent/plan_execute/planner.py:59

↓ 15 callersFunctioncall_tool

Helper: call an MCP tool and return parsed JSON response.

src/servers/iot/tests/conftest.py:49

↓ 14 callersFunctiondb

()

src/servers/wo/main.py:53

↓ 13 callersFunctionagent_run_span

Start a root span for an agent ``run()`` call. Sets canonical attributes (runner name, GenAI system/model, question length, and any ambient `

src/observability/runspan.py:49

↓ 13 callersFunctioncall_tool

Helper: call an MCP tool and return parsed JSON response.

src/servers/fmsr/tests/conftest.py:13

↓ 12 callersFunction_make_step

( n: int, server: str = "iot", tool: str = "sites", deps: list[int] | None = None, expecte

src/agent/tests/test_runner.py:69

↓ 12 callersFunction_patch_mcp

(tool_response: str = _TOOL_RESPONSE)

src/agent/tests/test_runner.py:56

↓ 12 callersFunctionenvelope

( data: Any, *, cached: bool = False, duration_ms: int = 0, record_count: Optional[int] =

src/servers/wo/envelope.py:14

↓ 12 callersFunctionpersist_trajectory

Write a per-run evaluation record when ``AGENT_TRAJECTORY_DIR`` is set. Reads ``run_id`` / ``scenario_id`` from the same contextvars used by

src/observability/persistence.py:39

↓ 12 callersMethodrun

( self, dataset_path: str, dataset_config_dictionary: dict, tsmodel_prediction

src/servers/tsfm/anomaly.py:448

↓ 12 callersFunctiont_ms

(timer: Timer)

src/servers/wo/workorders.py:539

↓ 12 callersMethodupdate

(self, scores, X=None, timestamps=None)

src/servers/tsfm/anomaly.py:276

↓ 11 callersFunction_avg

(values: list[float])

src/evaluation/report.py:25

↓ 11 callersFunction_public

Strip CouchDB bookkeeping that agents shouldn't see.

src/servers/wo/workorders.py:41

↓ 11 callersFunctionclassify_tool

Label a Stirrup tool call ``"domain"`` / ``"code"`` / ``"other"``. MCP tools arrive as ``{server}__{tool}``; ``code_exec`` is code execution;

src/agent/stirrup_agent/trajectory.py:34

↓ 10 callersFunctionassess_iso10816

Classify vibration severity per ISO 10816. Args: rms_velocity_mm_s: Overall RMS velocity in mm/s (10-1000 Hz band). machine_

src/servers/vibration/dsp/fault_detection.py:40

↓ 10 callersFunctionerror

(message: str, code: str = "API_ERROR")

src/servers/wo/envelope.py:27

↓ 10 callersFunctionresolve_router_creds

Resolve endpoint + key for *model_id*, or ``None`` if not proxied. Args: strict: when ``True`` (default) a matched prefix with missing en

src/llm/routers.py:68

↓ 10 callersMethodrun

Run the deep-agents loop for *question*. Args: question: Natural-language question to answer. Returns: :clas

src/agent/deep_agent/runner.py:175

↓ 9 callersFunction_failed

Return an ErrorResult if the envelope reports failure, else None.

src/servers/wo/main.py:70

↓ 9 callersFunction_make_run_result

(items, raw_responses=None)

src/agent/openai_agent/tests/test_runner.py:134

↓ 9 callersFunction_resolve_args_with_llm

Generate tool arguments from the task description and prior step results.

src/agent/plan_execute/executor.py:194

↓ 9 callersFunction_result

(stype: str, passed: bool, run_id: str = "", **ops_kwargs)

src/evaluation/tests/test_report.py:21

↓ 9 callersFunction_text_part

(part_id, text, *, message_id=None, part_type="text")

src/agent/opencode_agent/tests/test_runner.py:242

↓ 9 callersFunctioncall_tool

Helper: call an MCP tool and return parsed JSON response.

src/servers/utilities/tests/conftest.py:8

↓ 9 callersFunctionmake_scenario

()

src/evaluation/tests/conftest.py:14

↓ 8 callersFunction_build_trajectory_from_events

Convert OpenCode events into the shared SDK-style trajectory shape. Text/tool parts are merged by part id, turns are split on OpenCode step-f

src/agent/opencode_agent/runner.py:423

↓ 8 callersFunction_db_url

(db, *parts)

src/couchdb/loader.py:199

↓ 8 callersFunctionbuild_report

(results: list[ScenarioResult])

src/evaluation/report.py:106

↓ 8 callersFunctioncompute_fft

Compute single-sided FFT magnitude spectrum. Args: data: 1D time-domain signal fs: Sampling frequency in Hz window:

src/servers/vibration/dsp/fft_analysis.py:20

↓ 8 callersMethodevaluate

( self, *, trajectories_path: Path, scenarios_paths: list[Path], )

src/evaluation/evaluator.py:41

↓ 8 callersFunctionis_openai_compat

True if *model_id* targets an OpenAI-compatible router (e.g. TokenRouter).

src/llm/routers.py:63

↓ 8 callersFunctionmake_backend

Return the right backend for *model_id*. OpenAI-compatible routers (e.g. ``tokenrouter/...``) use the native ``openai`` SDK via :class:`OpenA

src/llm/__init__.py:18

↓ 8 callersFunctionmock_llm

Factory fixture: MockLLM(response='').

src/agent/tests/conftest.py:29

↓ 8 callersFunctionresolve_model

Strip any known router prefix from *model_id* if present. ``"tokenrouter/MiniMax-M3"`` -> ``"MiniMax-M3"``; ``"anthropic/claude-sonnet-4-6"``

src/llm/routers.py:53

↓ 8 callersMethodresolved_order

Return steps in topological order (dependencies before dependents).

src/agent/plan_execute/models.py:32

↓ 8 callersMethodrun

Execute a tool or workflow. ToolUniverse form: run({"name": "iot.sensors", "arguments": {...}}) Shorthand: run("iot.sensors

src/mcphub/__init__.py:202

↓ 8 callersFunctionsequential_llm

Factory fixture: SequentialMockLLM(responses=[...]).

src/agent/tests/conftest.py:39

↓ 7 callersFunction_build_trajectory

Extract a Trajectory from a Runner.run result. Walks ``result.new_items`` to collect text messages, tool calls, and tool outputs. Token usag

src/agent/openai_agent/runner.py:102

↓ 7 callersFunction_build_trajectory

Extract a :class:`Trajectory` from the deep-agent message list. Walks through ``AIMessage`` / ``ToolMessage`` entries in order, grouping each

src/agent/deep_agent/runner.py:86

↓ 7 callersFunction_doc_id

(site_id: str, wonum: str)

src/servers/wo/workorders.py:33

↓ 7 callersFunction_token_count

(value: Any)

src/evaluation/metrics.py:78

↓ 7 callersFunction_token_count

(value: Any)

src/agent/opencode_agent/runner.py:294

↓ 7 callersFunctionadd_common_args

Register the args shared by every SDK CLI. Adds the positional ``question`` plus ``--model-id``, ``--show-trajectory``, ``--json``, and ``--v

src/agent/_cli_common.py:37

↓ 7 callersFunctionevaluate_static_json

Evaluate one structured gold answer against one model answer.

src/evaluation/scorers/static_json.py:292

↓ 7 callersFunctionparse_structured_answer

Parse JSON/Python-like structured output into a Python object.

src/evaluation/scorers/static_json.py:177

↓ 7 callersMethodput

(self, doc: Dict[str, Any])

src/servers/wo/couch.py:56

↓ 7 callersFunctionrun_sdk_cli

Run the standard SDK-CLI lifecycle. Loads ``.env``, parses args with the caller's parser factory, configures stderr logging, initialises OTEL

src/agent/_cli_common.py:130

↓ 7 callersFunctionset_run_context

Set the ambient run/scenario IDs read by the next :func:`agent_run_span`. CLIs call this once before invoking ``runner.run(...)``; the runner its

src/observability/runspan.py:23

↓ 7 callersMethodto_dict

(self)

src/servers/vibration/dsp/fault_detection.py:172

↓ 6 callersFunction_iso

(dt: datetime)

src/servers/wo/workorders.py:37

↓ 6 callersFunction_make_message_item

Create a fake MessageOutputItem.

src/agent/openai_agent/tests/test_runner.py:112

↓ 6 callersFunction_mutation

( res: Dict[str, Any], verb: str )

src/servers/wo/main.py:77

↓ 6 callersFunction_parse_json

Extract a JSON object from an LLM response, with markdown fence handling. Returns the parsed dict, or None if no JSON object could be extracted.

src/agent/plan_execute/executor.py:225

↓ 6 callersFunction_resolve_run_dir

Return OpenCode's working directory for this run. The safe default is the repo root with local file/bash/edit tools denied. Any filesystem/co

src/agent/opencode_agent/runner.py:192

↓ 6 callersFunctioncompute_bearing_frequencies

Compute bearing characteristic frequencies. Args: rpm: Shaft rotational speed in RPM n_balls: Number of rolling elements (ba

src/servers/vibration/dsp/bearing_freqs.py:65

↓ 6 callersMethodexecute_plan

Execute all plan steps in dependency order.

src/agent/plan_execute/executor.py:78

↓ 6 callersFunctionprint_result

Common post-run output: JSON dump OR trajectory + answer sections.

src/agent/_cli_common.py:120

↓ 6 callersMethodrun

Run the claude-agent-sdk loop for *question*. Args: question: Natural-language question to answer. Returns:

src/agent/claude_agent/runner.py:112

↓ 6 callersFunctionvar

(est, act)

src/servers/wo/workorders.py:195

↓ 5 callersMethod_all_pass_response

(self)

src/evaluation/tests/test_scorers.py:66

↓ 5 callersFunction_build_dataset_config

( timestamp_column: str, target_columns: List[str], conditional_columns: Optional[List[str]],

src/servers/tsfm/main.py:85

↓ 5 callersFunction_is_known_site

(site_name: str)

src/servers/iot/main.py:245

↓ 5 callersFunction_make_tool_call_item

Create a fake ToolCallItem.

src/agent/openai_agent/tests/test_runner.py:119

↓ 5 callersFunction_make_usage

(input_tokens: int, output_tokens: int)

src/agent/openai_agent/tests/test_runner.py:130

↓ 5 callersFunction_normalize_model

(model: str)

src/evaluation/metrics.py:159

↓ 5 callersFunction_parse_tool_call

Parse LLM output into a {tool, args} dict (utility, not used in main path).

src/agent/plan_execute/executor.py:351

↓ 5 callersFunction_result

(passed: bool = True, ops: OpsMetrics | None = None)

src/evaluation/tests/test_metrics.py:18

↓ 5 callersFunction_set

(fut, value, error=False)

src/mcphub/__init__.py:147

↓ 5 callersFunctionbuild_trajectory

Convert a Stirrup message history into a :class:`Trajectory`.

src/agent/stirrup_agent/trajectory.py:100

↓ 5 callersFunctionenvelope_spectrum

Full envelope analysis pipeline: band-pass -> envelope -> FFT. Args: signal: Raw vibration time-domain signal. fs: Sampling

src/servers/vibration/dsp/envelope.py:69

↓ 5 callersMethodexecute_step

Execute a single plan step. 1. Resolve the MCP server assigned to this step. 2. If no tool is specified, return expected_output direc

src/agent/plan_execute/executor.py:127

↓ 5 callersFunctionfinal_answer

Final answer: last non-empty assistant text, else the finish reason.

src/agent/stirrup_agent/trajectory.py:152

↓ 5 callersMethodgenerate_plan

Generate a plan for a question given available servers and their tools. Args: question: The user question to answer.

src/agent/plan_execute/planner.py:112

↓ 5 callersMethodget

(self, doc_id)

src/servers/wo/tests/test_workorders.py:23

↓ 5 callersFunctionload_scenarios

Load scenarios from one or more files or directories. Supported inputs: 1. Existing JSON / JSONL scenario files. 2. A directory containi

src/evaluation/loader.py:45

↓ 5 callersFunctionmetrics_from_trajectory

Extract per-task ops metrics from a persisted trajectory record.

src/evaluation/metrics.py:25

↓ 5 callersMethodrun

Run the agent on *question* and return a structured result.

src/agent/runner.py:46

↓ 4 callersFunction_build_mcp_connections

Convert ``server_paths`` entries into ``MultiServerMCPClient`` specs. Entry-point names (``str``) become ``{"command": "uv", "args": ["run", name

src/agent/deep_agent/runner.py:65

↓ 4 callersFunction_build_mcp_servers

Convert server_paths entries into MCPServerStdio instances. Entry-point names (str without path separators) become ``MCPServerStdio(command="

src/agent/openai_agent/runner.py:77

↓ 4 callersFunction_build_mcp_servers

Convert server_paths entries into claude-agent-sdk mcp_servers dicts. Entry-point names (str without path separators) become ``{"command": "u

src/agent/claude_agent/runner.py:62

↓ 4 callersFunction_build_run_config

Build a RunConfig with a LiteLLM model provider when needed. When *model_id* starts with a proxy-router prefix (``litellm_proxy/`` or ``token

src/agent/openai_agent/runner.py:49

↓ 4 callersFunction_fake_agent

Return a mock with an ``ainvoke`` coroutine that yields *final_state*.

src/agent/deep_agent/tests/test_runner.py:219

↓ 4 callersFunction_make_sine

Generate a pure sine wave and store it; return (data_id, signal, sr).

src/servers/vibration/tests/test_tools.py:20

↓ 4 callersFunction_make_tool_output_item

Create a fake ToolCallOutputItem.

src/agent/openai_agent/tests/test_runner.py:125

↓ 4 callersFunction_parse_result

Extract the first JSON dict from an MCP tool call result.

src/servers/vibration/tests/test_mcp_e2e.py:146

↓ 4 callersFunction_peak_at

(f_target: float)

src/servers/vibration/dsp/fault_detection.py:116

↓ 4 callersFunction_read_ts_data

(dataset_path: str, dataset_config_dictionary=None)

src/servers/tsfm/io.py:77

↓ 4 callersFunction_resolve_args

Simple string substitution of {{step_N}} placeholders (kept for tests).

src/agent/plan_execute/executor.py:335

↓ 4 callersFunction_resolve_opencode_model_and_provider

Translate AssetOpsBench router model IDs into OpenCode config. OpenCode wants ``provider/model``. For AssetOpsBench router prefixes such as

src/agent/opencode_agent/runner.py:112

↓ 4 callersFunction_sdk_env

Build env overrides for the claude-agent-sdk subprocess. When routing through a proxy (LiteLLM proxy or TokenRouter) the SDK needs the proxy

src/agent/claude_agent/runner.py:44

↓ 4 callersMethod_submit

(self, action: str, payload)

src/mcphub/__init__.py:87

next →1–100 of 947, ranked by callers