qlm-measure

Open measurement SDK for education AI — evidence events, record verification, and dataset clients.

Apache-2.0 · Developer Portal

What this is

A typed evidence-event vocabulary, a hash-chained record format, and a verifier that checks bookkeeping integrity without containing any estimation mathematics. The verifier runs locally, makes no network calls, and records never leave your machine.

What this is not

Not the estimation engine (that runs server-side). Not a learning analytics system. Not a replacement for the measurement service.

Install

Python:

pip install qlm-measure

JavaScript / TypeScript:

npm install qlm-measure

Produce a record from your own tool

See docs/recorder-quickstart.md — ten lines, no QLM account, no network.

Verify a record

qlm-measure verify record.json
qlm-measure verify records.jsonl --format json --report report.json
qlm-measure explain HASH.ENTRY_RECOMPUTE
qlm-measure version

Exit codes: 0 all valid, 1 at least one invalid, 2 usage or parse error.

Verify a product hash chain

Per-event chains sealed by QLM products (TeachProof clinical evidence, Play clinical events) are defined once, by name and version, in qlm-measure/schemes — see docs/chain-schemes.md.

qlm-measure verify-chain events.json --family tpc/clinical
qlm-measure verify-chain --list

Packages

• Python — pip install qlm-measure
• JavaScript/TypeScript — npm install qlm-measure

Both packages ship the same 10 checks, the same report schema, and produce byte-identical reports (minus the tool block). The check catalog, canonicalization spec, and golden vectors are shared.

Limitations (read first)

• The verifier checks bookkeeping integrity, not estimation correctness. It can tell you if a record was tampered with. It cannot tell you if the posteriors are accurate.
• Evidence event schemas are v0.2 and may change in minor versions. Pin your dependency.
• The estimation service requires the QLM API. The verifier does not.

Assistance capture and native exports

The prepared QLM Transformer and Airflow workflows add assistance definitions, capture declarations, explicit unknowns and native research-export mappings. See the release plan and demo source. These tools are separate from the measurement SDK and hosted estimation engine.

License

The SDK is Apache-2.0. See LICENSE. Transformer and Airflow subtrees retain their MIT licenses; data and specifications retain separate terms. See license scopes.