{
  "$schema": "https://json-schema.org/draft/2020-12/schema",
  "$id": "https://example.com/schemas/gi.schema.json",
  "title": "Grounded Insight Layer Episode Artifact (gi.json)",
  "description": "Schema for per-episode GIL data with Insights, Quotes, and grounding relationships",
  "type": "object",
  "additionalProperties": false,
  "required": [
    "schema_version",
    "model_version",
    "prompt_version",
    "episode_id",
    "nodes",
    "edges"
  ],
  "properties": {
    "schema_version": {
      "type": "string",
      "pattern": "^1\\.0$",
      "description": "Schema version (must be 1.0)"
    },
    "model_version": {
      "type": "string",
      "minLength": 1,
      "description": "Insight text lineage: model id derived from gi_insight_source and the summarization / generate_insights stack (see podcast_scraper.gi.provenance and GROUNDED_INSIGHTS_GUIDE.md)"
    },
    "prompt_version": {
      "type": "string",
      "minLength": 1,
      "description": "Prompt version used for extraction (e.g., v2.1)"
    },
    "episode_id": {
      "type": "string",
      "minLength": 1,
      "description": "Episode identifier (e.g., episode:abc123)"
    },
    "nodes": {
      "type": "array",
      "items": { "$ref": "#/$defs/node" },
      "description": "All nodes in this episode's GIL data"
    },
    "edges": {
      "type": "array",
      "items": { "$ref": "#/$defs/edge" },
      "description": "All edges in this episode's GIL data"
    }
  },
  "$defs": {
    "node": {
      "oneOf": [
        { "$ref": "#/$defs/podcast_node" },
        { "$ref": "#/$defs/episode_node" },
        { "$ref": "#/$defs/speaker_node" },
        { "$ref": "#/$defs/topic_node" },
        { "$ref": "#/$defs/insight_node" },
        { "$ref": "#/$defs/quote_node" }
      ]
    },
    "node_base": {
      "type": "object",
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1, "description": "Unique node identifier" },
        "type": { "type": "string", "enum": ["Podcast", "Episode", "Speaker", "Topic", "Insight", "Quote"], "description": "Node type" },
        "properties": { "type": "object", "description": "Type-specific properties" },
        "confidence": { "type": "number", "minimum": 0, "maximum": 1, "description": "Extraction confidence (optional, for ML-derived nodes)" }
      }
    },
    "podcast_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Podcast" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["title", "rss_url"],
          "properties": { "title": { "type": "string", "minLength": 1 }, "rss_url": { "type": "string", "minLength": 1 }, "publisher": { "type": "string" } }
        }
      }
    },
    "episode_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Episode" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["podcast_id", "title", "publish_date"],
          "properties": {
            "podcast_id": { "type": "string", "minLength": 1 },
            "title": { "type": "string", "minLength": 1 },
            "publish_date": { "type": "string", "format": "date-time" },
            "audio_url": { "type": "string" },
            "duration_ms": { "type": "integer", "minimum": 0 }
          }
        }
      }
    },
    "speaker_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Speaker" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["name"],
          "properties": { "name": { "type": "string", "minLength": 1 }, "aliases": { "type": "array", "items": { "type": "string" } } }
        }
      }
    },
    "topic_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Topic" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["label"],
          "properties": { "label": { "type": "string", "minLength": 1 }, "aliases": { "type": "array", "items": { "type": "string" } } }
        }
      }
    },
    "insight_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Insight" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["text", "episode_id", "grounded"],
          "properties": {
            "text": { "type": "string", "minLength": 1, "description": "The insight statement (can be rephrased for clarity)" },
            "episode_id": { "type": "string", "minLength": 1, "description": "Episode where insight was extracted" },
            "grounded": { "type": "boolean", "description": "Whether insight has ≥1 supporting quote (CRITICAL: must be explicit)" }
          }
        },
        "confidence": { "type": "number", "minimum": 0, "maximum": 1, "description": "Extraction confidence (0.0-1.0)" }
      },
      "description": "A key takeaway or conclusion extracted from episode content"
    },
    "quote_node": {
      "type": "object",
      "additionalProperties": false,
      "required": ["id", "type", "properties"],
      "properties": {
        "id": { "type": "string", "minLength": 1 },
        "type": { "const": "Quote" },
        "properties": {
          "type": "object",
          "additionalProperties": false,
          "required": ["text", "episode_id", "char_start", "char_end", "timestamp_start_ms", "timestamp_end_ms", "transcript_ref"],
          "properties": {
            "text": { "type": "string", "minLength": 1, "description": "VERBATIM text from transcript (no paraphrasing!)" },
            "episode_id": { "type": "string", "minLength": 1, "description": "Episode containing the quote" },
            "speaker_id": { "type": ["string", "null"], "description": "Speaker who said the quote (nullable if no diarization)" },
            "char_start": { "type": "integer", "minimum": 0, "description": "Character start in transcript text" },
            "char_end": { "type": "integer", "minimum": 0, "description": "Character end in transcript text" },
            "timestamp_start_ms": { "type": "integer", "minimum": 0, "description": "Timestamp start (milliseconds)" },
            "timestamp_end_ms": { "type": "integer", "minimum": 0, "description": "Timestamp end (milliseconds)" },
            "transcript_ref": { "type": "string", "minLength": 1, "description": "Reference to transcript artifact (e.g., transcript.json)" }
          }
        }
      },
      "description": "A verbatim transcript span used as evidence for an Insight"
    },
    "edge": {
      "type": "object",
      "additionalProperties": false,
      "required": ["type", "from", "to"],
      "properties": {
        "type": { "type": "string", "enum": ["HAS_EPISODE", "SPOKE_IN", "HAS_INSIGHT", "SUPPORTED_BY", "SPOKEN_BY", "ABOUT", "RELATED_TO"], "description": "Edge type" },
        "from": { "type": "string", "minLength": 1, "description": "Source node ID" },
        "to": { "type": "string", "minLength": 1, "description": "Target node ID" },
        "properties": { "type": "object", "additionalProperties": true, "description": "Optional edge properties" }
      },
      "allOf": [
        {
          "if": { "properties": { "type": { "enum": ["ABOUT", "RELATED_TO"] } } },
          "then": { "properties": { "properties": { "type": "object", "additionalProperties": true, "properties": { "confidence": { "type": "number", "minimum": 0, "maximum": 1, "description": "Edge confidence (optional)" } } } } }
        }
      ]
    }
  }
}