The Python client

reliafy-client is the Python client for the API, and it installs the reliafy command line too. Pure standard library, no dependencies; you bring SurPyval (docs) for your own fits.

Install and sign in

pip install reliafy-client        # installed as reliafy-client, imported as `reliafy`

from reliafy import Client
rf = Client(token="rlf_...")      # or RELIAFY_TOKEN, or the token saved by `reliafy login`

• Version 0.3 needs a Reliafy server with the tools API: reliafy.com has it from this release; a self-hosted server needs a build where GET /api/v1/tools answers. Against an older server the new calls stop with reliafy.ServerTooOld, and everything from 0.2 keeps working.
• Self-hosted: Client(base_url="https://reliafy.example.com"). A local server with sign-in off needs no token.

Every tool: Client.call

{ … }   # the tool's result, the same JSON an AI assistant gets

rf.call("list", "models")
rf.call("life_data_analysis", "fit", distribution="weibull",
        data=[120, 340, 510, 700, 980], censored=[0, 0, 1, 0, 0], unit="hours")
rf.call("guide", topic="life_data")                       # a tool without actions
rf.call("rbd_analysis", "analyze", rbd_id="…", simulate=True, wait=True)

rf.tools()     # the tools and actions you may call (GET /api/v1/tools)
rf.spec()      # the OpenAPI spec

• Shortcuts: rf.list(kind), rf.get(kind, id), rf.fit(data, distribution=…, censored=…), rf.reliability(model_id, times), rf.guide(topic), rf.wait(job_id).
• reliafy.call(tool, action, **args) does the same with the configured token.
• A refused call raises reliafy.ReliafyError with status (the HTTP status) and detail (what Reliafy said).

Files, photos and data

rf.upload_dataset("bearings.xlsx", name="Bearings")     # CSV or .xlsx, as a dataset
rf.import_rbd("plant.xml")                              # BlockSim, Open-PSA, Galileo, RePyability, the Excel template
rf.import_excel("fmea.xlsx", target="rcm")              # dataset, rcm or rbd
rf.add_photo("nameplate.jpg", target_id="…")            # Reliafy Maintenance users

rf.push_usage(fleet_id, [{"name": "Truck 01", "current_use": 5230}])
rf.push_measurements(tracked_fleet_id, file="readings.csv")
rf.push_lives(dataset_id, file="new_failures.csv")      # appends, then refits its models

Files are streamed from disk. The ingest helpers take rows, csv= text or file= a path, and are also on the module (reliafy.push_usage(…)). Reliafy Maintenance users also have push_readings for meter and condition readings.

The 0.2 functions (still supported)

Everything from 0.2 works unchanged, on any server: reliafy.configure(token=…), then the functions below. Each maps to one of the original routes; open one for its parameters and what it returns.

reliafy.models

"https://reliafy.com/modelling/m/a1b2…"   # str — open in the app

"https://reliafy.com/modelling/m/…"   # str

[ { "id": "a1b2…", "name": "Bearing life", "distribution": "Weibull",
    "kind": "distribution", "n": 30, "unit": "hours", "url": "/modelling/m/…" }, … ]

{ "id": …, "distribution": "Weibull", "params": [{name, value, se, ci}],
  "coefficients": [], "metrics": null, "gof": [{id, label, value}] }

{ "at": { "reliability": 0.666, "failure": 0.334, "hazard": …,
          "cumulative_hazard": …, "density": … } }   # or {"curves": {...}}

{ "id": …, "name": "Bearing life", "distribution": "Weibull", "url": "/modelling/m/…" }

reliafy.data

{ "id": "d3…", "name": "Bearings", "n_rows": 8, "columns": ["hours","failed"], "url": "/modelling/data/…" }

reliafy.strategy

{ "optimal_time": 580.5, "optimal_cost_rate": 0.51, "beneficial": true, "mttf": 1272.4 }

{ "interval": 176.4, "method": "…", "mttf": 8760, "target_availability": 0.99 }

{ "units": 128, "failures": 8, "consumer_risk": 0.097, "producer_risk": 0.192, "summary": "Test 128 units…", "oc_curve": {…}, "tradeoff": {…} }

reliafy.fleet

{ "fleet": {id, name, model_id}, "forecast": {status, method, periods, expected_failures, …} }

• Errors raise reliafy.ReliafyError; reads/writes are scoped to your own data.
• Full request and response field tables are on Original routes and ingest.