Overview
Complete API reference documentation for the fiddler-evals package.Components
Connection
Connection management and initializationDataset Items
Build dataset items from DataFrames, CSV files, and JSON Lines filesEntities
Core entity classes for interacting with the platformEvaluators
Built-in and custom evaluators for LLM assessment- AnswerRelevance
- Coherence
- Conciseness
- ContextRelevance
- CustomJudge
- CustomJudgeSpec
- EvalFn
- Evaluator
- FTLPromptSafety
- FTLResponseFaithfulness
- InputFieldSpec
- Message
- OutputField
- OutputFieldTransform
- RAGFaithfulness
- RegexMatch
- RegexSearch
- Sentiment
- TopicClassification
Exceptions
Errors the SDK raises, and the import path to catch them- ApiError
- Conflict
- DatasetBusy
- DatasetWriteUnavailable
- IncompatibleClient
- ItemIdCollision
- NotFound
- SkipEval
- StaleDatasetVersion
- TransientError
- Unsupported
Pydantic Models
Pydantic data models and validation schemas- AddItemsFromSpansResponse
- DatasetItem
- DatasetSchemaField
- DatasetSchemaResponse
- DatasetVersionCompact
- EvaluatorOutputCoverage
- FieldMapping
- NewDatasetItem
- OperatorType
- QueryCondition
- QueryConditionType
- QueryRule
- Score
- ScoreStatus
- SearchFilter
- SearchScope
- SpanAttributeCoverage
- SpanFieldsResponse
- SpanReference