{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://raw.githubusercontent.com/shruggietech/insonic/v0.0.0/schemas/v0.0.0/speaker-dataset.schema.json",
  "title": "Speaker training dataset snapshot",
  "description": "Portable immutable training corpus selection. Source, speaker, attribution, transcript and preparation revisions are frozen; later training preparation binds final bytes in a separate manifest.",
  "type": "object",
  "properties": {
    "schema_version": {
      "const": "0.0.0",
      "description": "Version of this portable document contract."
    },
    "kind": {
      "const": "speaker-dataset",
      "description": "Discriminator identifying an immutable speaker training dataset snapshot."
    },
    "workspace_id": {
      "$ref": "common.schema.json#/$defs/uuid",
      "description": "Workspace owning the snapshot and every referenced member."
    },
    "dataset_snapshot_id": {
      "$ref": "common.schema.json#/$defs/uuid",
      "description": "Stable immutable training_dataset_snapshot identity."
    },
    "speaker_id": {
      "$ref": "common.schema.json#/$defs/uuid",
      "description": "Originating catalog speaker for this selected corpus."
    },
    "speaker_identity_revision": {
      "$ref": "common.schema.json#/$defs/revision",
      "description": "Exact identity revision used when selecting this speaker's corpus."
    },
    "catalog_revision": {
      "$ref": "common.schema.json#/$defs/revision",
      "description": "Catalog revision against which selection was resolved."
    },
    "created_at": {
      "$ref": "common.schema.json#/$defs/utcInstant",
      "description": "UTC instant when this snapshot was frozen."
    },
    "manifest_sha256": {
      "$ref": "common.schema.json#/$defs/sha256",
      "description": "Digest of the separately published canonical snapshot manifest, not a hash asserted over this transport envelope including itself."
    },
    "manifest_artifact": {
      "$ref": "common.schema.json#/$defs/artifactReference",
      "description": "Optional published artifact containing the canonical immutable snapshot manifest."
    },
    "selection": {
      "$ref": "#/$defs/selection",
      "description": "Complete frozen selection recipe and effective filters."
    },
    "preparation": {
      "$ref": "#/$defs/preparation",
      "description": "Requested preparation contract frozen before final training inputs are materialized."
    },
    "members": {
      "type": "array",
      "description": "Ordered included segment memberships; empty snapshots represent no matching corpus and do not imply runnable training.",
      "items": {
        "$ref": "#/$defs/member"
      }
    },
    "exclusions": {
      "type": "array",
      "description": "Frozen excluded-segment decisions and their technical or user-supplied reasons.",
      "items": {
        "$ref": "#/$defs/exclusion"
      }
    },
    "summary": {
      "$ref": "#/$defs/summary",
      "description": "Frozen corpus totals and diagnostic summary."
    },
    "extensions": {
      "$ref": "common.schema.json#/$defs/extensions",
      "description": "Namespaced dataset metadata preserved without altering immutable membership."
    }
  },
  "required": [
    "schema_version",
    "kind",
    "workspace_id",
    "dataset_snapshot_id",
    "speaker_id",
    "speaker_identity_revision",
    "catalog_revision",
    "created_at",
    "manifest_sha256",
    "selection",
    "preparation",
    "members",
    "exclusions",
    "summary"
  ],
  "additionalProperties": false,
  "$defs": {
    "selection": {
      "type": "object",
      "description": "Exact effective selection recipe, allowing automatic corpus selection without segment-by-segment approval.",
      "properties": {
        "recipe_id": {
          "$ref": "common.schema.json#/$defs/uuid",
          "description": "Stable selection recipe identity."
        },
        "recipe_revision": {
          "$ref": "common.schema.json#/$defs/revision",
          "description": "Frozen recipe revision."
        },
        "preset": {
          "type": "string",
          "minLength": 1,
          "description": "Readable preparation/diagnostic preset identity."
        },
        "preset_version": {
          "type": "string",
          "minLength": 1,
          "description": "Version defining the chosen preset's behavior."
        },
        "attribution_bases": {
          "type": "array",
          "description": "Allowed assignment bases; an empty array includes every declared basis.",
          "items": {
            "type": "string",
            "minLength": 1,
            "description": "One included attribution basis."
          },
          "uniqueItems": true
        },
        "source_asset_ids": {
          "type": "array",
          "description": "Optional source filter; an empty array covers the speaker's whole workspace corpus.",
          "items": {
            "$ref": "common.schema.json#/$defs/uuid",
            "description": "One included original media asset."
          },
          "uniqueItems": true
        },
        "languages": {
          "type": "array",
          "description": "Optional language filter; an empty array is language-independent.",
          "items": {
            "type": "string",
            "minLength": 1,
            "description": "One declared language identifier."
          },
          "uniqueItems": true
        },
        "minimum_duration_us": {
          "type": "integer",
          "minimum": 0,
          "description": "Minimum original segment duration considered by this recipe."
        },
        "maximum_duration_us": {
          "anyOf": [
            {
              "type": "integer",
              "minimum": 1,
              "description": "Maximum original segment duration considered by this recipe."
            },
            {
              "type": "null"
            }
          ],
          "description": "Upper segment-duration bound, or null when no upper bound is configured."
        },
        "explicitly_excluded_segment_ids": {
          "type": "array",
          "description": "User-selected exclusions applied automatically to this snapshot.",
          "items": {
            "$ref": "common.schema.json#/$defs/uuid",
            "description": "Excluded stable segment identity."
          },
          "uniqueItems": true
        },
        "options": {
          "$ref": "common.schema.json#/$defs/nonsecretOptions",
          "description": "Adapter-specific filters and quality thresholds validated by the selected preset/adapter."
        },
        "extensions": {
          "$ref": "common.schema.json#/$defs/extensions",
          "description": "Namespaced custom selection rules with documented adapter interpretation."
        }
      },
      "required": [
        "recipe_id",
        "recipe_revision",
        "preset",
        "preset_version",
        "attribution_bases",
        "source_asset_ids",
        "languages",
        "minimum_duration_us",
        "maximum_duration_us",
        "explicitly_excluded_segment_ids",
        "options"
      ],
      "additionalProperties": false
    },
    "preparation": {
      "type": "object",
      "description": "Frozen requested preparation contract; final prepared bytes are bound by a separate immutable preparation manifest.",
      "properties": {
        "adapter_id": {
          "type": "string",
          "minLength": 1,
          "description": "Selected preparation adapter identity."
        },
        "contract_version": {
          "type": "string",
          "minLength": 1,
          "description": "Version of the preparation request/result contract."
        },
        "adapter_version": {
          "type": "string",
          "minLength": 1,
          "description": "Version of the selected adapter implementation."
        },
        "options": {
          "$ref": "common.schema.json#/$defs/nonsecretOptions",
          "description": "Exact audio preparation options without credentials or temporary paths."
        },
        "transcript_required": {
          "type": "boolean",
          "description": "Whether this training input requires selected transcript and alignment references."
        },
        "output_format": {
          "anyOf": [
            {
              "type": "string",
              "minLength": 1,
              "description": "Requested prepared audio format."
            },
            {
              "type": "null"
            }
          ],
          "description": "Requested audio format, or null when the adapter consumes source streams directly."
        },
        "sample_rate_hz": {
          "anyOf": [
            {
              "type": "integer",
              "minimum": 1,
              "description": "Requested prepared sample rate in hertz."
            },
            {
              "type": "null"
            }
          ],
          "description": "Requested sample rate, or null when unchanged or not applicable."
        },
        "channel_count": {
          "anyOf": [
            {
              "type": "integer",
              "minimum": 1,
              "description": "Requested prepared channel count."
            },
            {
              "type": "null"
            }
          ],
          "description": "Requested output channel count, or null when unchanged or not applicable."
        }
      },
      "required": [
        "adapter_id",
        "contract_version",
        "adapter_version",
        "options",
        "transcript_required",
        "output_format",
        "sample_rate_hz",
        "channel_count"
      ],
      "additionalProperties": false
    },
    "member": {
      "type": "object",
      "description": "One included, ordered, immutable dataset member; all referenced revisions and source bytes are frozen.",
      "properties": {
        "ordinal": {
          "type": "integer",
          "minimum": 0,
          "description": "Zero-based explicit order within this snapshot."
        },
        "segment_id": {
          "$ref": "common.schema.json#/$defs/uuid",
          "description": "Stable source segment identity."
        },
        "segment_revision": {
          "$ref": "common.schema.json#/$defs/revision",
          "description": "Exact segment boundary revision."
        },
        "source": {
          "$ref": "speaker-segments.schema.json#/$defs/source",
          "description": "Original byte/stream/channel identity and source interval."
        },
        "processing_run_id": {
          "$ref": "common.schema.json#/$defs/uuid",
          "description": "Diarization or segmentation run that produced the member interval."
        },
        "voice_id": {
          "$ref": "common.schema.json#/$defs/uuid",
          "description": "Run-local acoustic voice identity used by this member."
        },
        "attribution": {
          "$ref": "speaker-segments.schema.json#/$defs/attribution",
          "description": "Exact attribution and speaker identity revisions used for membership."
        },
        "transcript": {
          "$ref": "speaker-segments.schema.json#/$defs/transcript",
          "description": "Optional exact transcript/cue evidence; required by application validation when preparation.transcript_required is true."
        },
        "prepared_audio": {
          "$ref": "speaker-segments.schema.json#/$defs/preparedAudio",
          "description": "Any audio already materialized before this snapshot was frozen."
        },
        "diagnostics": {
          "type": "array",
          "description": "Technical findings observed when this member was selected.",
          "items": {
            "$ref": "speaker-segments.schema.json#/$defs/diagnostic",
            "description": "One method-qualified selection finding."
          }
        },
        "extensions": {
          "$ref": "common.schema.json#/$defs/extensions",
          "description": "Namespaced member annotations that preserve frozen core provenance."
        }
      },
      "required": [
        "ordinal",
        "segment_id",
        "segment_revision",
        "source",
        "processing_run_id",
        "voice_id",
        "attribution",
        "diagnostics"
      ],
      "additionalProperties": false
    },
    "exclusion": {
      "type": "object",
      "description": "One explicit selection exclusion retained for explainable automatic corpus selection.",
      "properties": {
        "segment_id": {
          "$ref": "common.schema.json#/$defs/uuid",
          "description": "Stable segment excluded from this snapshot."
        },
        "segment_revision": {
          "$ref": "common.schema.json#/$defs/revision",
          "description": "Segment revision evaluated by this decision."
        },
        "reason_codes": {
          "type": "array",
          "description": "Declared reasons for exclusion, such as duplicate-source-span or configured-overlap-filter.",
          "items": {
            "type": "string",
            "minLength": 1,
            "description": "One machine-readable reason code."
          },
          "minItems": 1,
          "uniqueItems": true
        },
        "method": {
          "type": "string",
          "minLength": 1,
          "description": "Preset/adapter or user action responsible for the decision."
        },
        "method_version": {
          "type": "string",
          "minLength": 1,
          "description": "Version of the method when this exclusion was computed."
        },
        "details": {
          "$ref": "common.schema.json#/$defs/nonsecretOptions",
          "description": "Nonsecret diagnostic values and effective thresholds explaining the decision."
        }
      },
      "required": [
        "segment_id",
        "segment_revision",
        "reason_codes",
        "method",
        "method_version",
        "details"
      ],
      "additionalProperties": false
    },
    "summary": {
      "type": "object",
      "description": "Totals frozen when selection completes; the application verifies consistency against membership.",
      "properties": {
        "segment_count": {
          "type": "integer",
          "minimum": 0,
          "description": "Number of included members."
        },
        "source_asset_count": {
          "type": "integer",
          "minimum": 0,
          "description": "Distinct original source assets represented by included members."
        },
        "original_duration_us": {
          "type": "integer",
          "minimum": 0,
          "description": "Sum of selected original intervals in integer microseconds after the recorded deduplication policy."
        },
        "prepared_duration_us": {
          "anyOf": [
            {
              "type": "integer",
              "minimum": 0,
              "description": "Known prepared duration in integer microseconds."
            },
            {
              "type": "null"
            }
          ],
          "description": "Known prepared duration, or null until preparation is materialized."
        },
        "excluded_segment_count": {
          "type": "integer",
          "minimum": 0,
          "description": "Number of excluded segment decisions retained in this snapshot."
        },
        "diagnostic_counts": {
          "type": "object",
          "description": "Counts keyed by declared technical diagnostic code.",
          "additionalProperties": {
            "type": "integer",
            "minimum": 0
          }
        },
        "languages": {
          "type": "array",
          "description": "Language identifiers observed or selected for the included corpus.",
          "items": {
            "type": "string",
            "minLength": 1,
            "description": "One language identifier."
          },
          "uniqueItems": true
        }
      },
      "required": [
        "segment_count",
        "source_asset_count",
        "original_duration_us",
        "prepared_duration_us",
        "excluded_segment_count",
        "diagnostic_counts",
        "languages"
      ],
      "additionalProperties": false
    }
  },
  "examples": [
    {
      "schema_version": "0.0.0",
      "kind": "speaker-dataset",
      "workspace_id": "00000000-0000-4000-8000-00000000000a",
      "dataset_snapshot_id": "00000000-0000-4000-8000-00000000000b",
      "speaker_id": "00000000-0000-4000-8000-000000000006",
      "speaker_identity_revision": 1,
      "catalog_revision": 1,
      "created_at": {
        "iso": "2026-10-04T20:01:00Z",
        "unix_ns": 1791144060000000000
      },
      "manifest_sha256": "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
      "selection": {
        "recipe_id": "00000000-0000-4000-8000-00000000000c",
        "recipe_revision": 1,
        "preset": "speech-clean",
        "preset_version": "1.0.0",
        "attribution_bases": [],
        "source_asset_ids": [],
        "languages": [],
        "minimum_duration_us": 0,
        "maximum_duration_us": null,
        "explicitly_excluded_segment_ids": [],
        "options": {
          "exclude_duplicate_source_spans": true
        }
      },
      "preparation": {
        "adapter_id": "example.audio-preparation",
        "contract_version": "1",
        "adapter_version": "1.0.0",
        "options": {
          "normalize": false
        },
        "transcript_required": false,
        "output_format": "wav-pcm-s16le",
        "sample_rate_hz": 16000,
        "channel_count": 1
      },
      "members": [
        {
          "ordinal": 0,
          "segment_id": "00000000-0000-4000-8000-000000000007",
          "segment_revision": 1,
          "source": {
            "media_entry_id": "00000000-0000-4000-8000-000000000001",
            "source_asset_id": "00000000-0000-4000-8000-000000000002",
            "source_artifact": {
              "artifact_id": "00000000-0000-4000-8000-000000000003",
              "sha256": "aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa",
              "byte_length": 640000
            },
            "stream_id": "00000000-0000-4000-8000-000000000004",
            "stream_index": 0,
            "channels": [
              0
            ],
            "channel_policy": "preserve",
            "interval": {
              "start_us": 1000000,
              "end_us": 3000000
            }
          },
          "processing_run_id": "00000000-0000-4000-8000-000000000008",
          "voice_id": "00000000-0000-4000-8000-000000000009",
          "attribution": {
            "attribution_id": "00000000-0000-4000-8000-000000000005",
            "revision": 1,
            "speaker_id": "00000000-0000-4000-8000-000000000006",
            "speaker_identity_revision": 1,
            "basis": "acoustic-comparison",
            "method": "example.voice-matcher",
            "confidence": {
              "value": 0.88,
              "scale": "unit-interval",
              "method_version": "1.0.0"
            }
          },
          "diagnostics": []
        }
      ],
      "exclusions": [],
      "summary": {
        "segment_count": 1,
        "source_asset_count": 1,
        "original_duration_us": 2000000,
        "prepared_duration_us": null,
        "excluded_segment_count": 0,
        "diagnostic_counts": {},
        "languages": []
      }
    }
  ]
}