app.gainforest.evaluator.evaluation
A single evaluation produced by an evaluator service. Exactly one of 'subject' (single target) or 'subjects' (batch) must be provided.
Details · record, key tid
| Kind | record |
|---|---|
| Record key | tid |
| Authority | evaluator.gainforest.app |
| Indexed |
Record
record · key tidProperties
confidence
integerOverall confidence in this evaluation (0-1000, where 1000 = 100.0%).
minimum: 0maximum: 1000
Timestamp of when this evaluation was produced.
dynamicProperties
stringAdditional structured data as a JSON string. Escape hatch for experimental result types before they are formalized into the union.
maxGraphemes: 10000
Identifier for the type of evaluation (must match one declared in the evaluator's service record).
maxGraphemes: 64
method
refMethod/model provenance for this specific evaluation (overrides service-level method if set).
neg
booleanIf true, this is a negation/withdrawal of a previous evaluation (like label negation).
result
unionThe typed evaluation result. The $type field determines which result schema is used.
- app.gainforest.evaluator.defs#speciesIdResult
- app.gainforest.evaluator.defs#dataQualityResult
- app.gainforest.evaluator.defs#verificationResult
- app.gainforest.evaluator.defs#classificationResult
- app.gainforest.evaluator.defs#measurementResult
- app.gainforest.evaluator.defs#bioacousticsResult
- app.gainforest.evaluator.defs#deforestationResult
- app.gainforest.evaluator.defs#carbonEstimationResult
subject
refSingle target record being evaluated. Use this OR subjects, not both.
subjects
arrayBatch evaluation: multiple target records sharing the same result. Use this OR subject, not both.
maxLength: 100
supersedes
string · at-uriAT-URI of a previous evaluation record that this one supersedes (e.g., model re-run with improved version).
Similar lexicons
semantic matchLexicons whose schemas describe something close to this one. Check them before publishing a new one that overlaps.
-
Shared type definitions for decentralized evaluator services. Evaluators attach structured, typed evaluation data to records (a more sophisticated evolution of the labeler pattern).
-
Declaration record published at rkey 'self' to register an account as an evaluator service. Analogous to app.bsky.labeler.service for labelers.
-
An evaluation of a hypercert record (e.g. an activity and its impact).
-
A subscription record published by a user in their own repo to request evaluations from a specific evaluator service. The evaluator detects subscriptions via Jetstream and processes matching records. Deleting this record unsubscribes.