tech.transparencia.news.enrichment

{
  "$type": "com.atproto.lexicon.schema",
  "defs": {
    "main": {
      "description": "AI enrichment of a news article with structured analysis.",
      "key": "tid",
      "record": {
        "properties": {
          "article": {
            "description": "Strong reference to the tech.transparencia.news.article record being enriched.",
            "ref": "com.atproto.repo.strongRef",
            "type": "ref"
          },
          "claims": {
            "description": "Key factual claims or assertions extracted from the article.",
            "items": {
              "ref": "tech.transparencia.defs#claim",
              "type": "ref"
            },
            "maxLength": 10,
            "type": "array"
          },
          "clickbaitReasoning": {
            "description": "Explanation of the clickbait score.",
            "maxGraphemes": 1000,
            "maxLength": 5000,
            "type": "string"
          },
          "clickbaitScore": {
            "description": "How accurately the headline represents the actual content. 0=pure clickbait (misleading/sensationalist title), 10=headline perfectly represents the article.",
            "maximum": 10,
            "minimum": 0,
            "type": "integer"
          },
          "contentDomain": {
            "description": "Primary domain/vertical of the article content, aligned with IPTC Media Topics top-level categories.",
            "knownValues": [
              "arts-culture-entertainment",
              "crime-law-justice",
              "disaster-accident",
              "economy-business-finance",
              "education",
              "environment",
              "health",
              "human-interest",
              "labour",
              "lifestyle-leisure",
              "politics",
              "religion",
              "science-technology",
              "society",
              "sport",
              "conflict-war-peace",
              "weather"
            ],
            "maxLength": 64,
            "type": "string"
          },
          "costUsd": {
            "description": "Estimated cost of this enrichment in USD as a decimal string (e.g., '0.001234'). Allows consumers to audit AI spend transparently.",
            "maxLength": 32,
            "type": "string"
          },
          "createdAt": {
            "description": "When this enrichment was generated.",
            "format": "datetime",
            "type": "string"
          },
          "economicIndicators": {
            "description": "Economic data points mentioned in the article.",
            "items": {
              "ref": "tech.transparencia.defs#economicIndicator",
              "type": "ref"
            },
            "maxLength": 10,
            "type": "array"
          },
          "emotionalTone": {
            "description": "Emotional framing of the article, independent of political orientation.",
            "knownValues": [
              "alarmist",
              "critical",
              "neutral",
              "optimistic",
              "celebratory",
              "melancholic",
              "informative",
              "satirical",
              "inspirational",
              "sensationalist",
              "technical",
              "investigative"
            ],
            "maxLength": 32,
            "type": "string"
          },
          "eventType": {
            "description": "Primary category of the news event.",
            "knownValues": [
              "announcement",
              "policy-legislation",
              "election-vote",
              "appointment-resignation",
              "scandal-controversy",
              "investigation-audit",
              "court-ruling-legal",
              "protest-demonstration",
              "armed-conflict",
              "natural-disaster",
              "accident",
              "crime-incident",
              "market-movement",
              "earnings-financial-results",
              "merger-acquisition",
              "product-launch",
              "product-recall",
              "scientific-discovery",
              "medical-breakthrough",
              "epidemic-health-crisis",
              "infrastructure-project",
              "cultural-event",
              "awards-ceremony",
              "match-result",
              "transfer-signing",
              "tournament-competition",
              "record-achievement",
              "death-obituary",
              "interview-profile",
              "opinion-editorial",
              "data-report-study",
              "forecast-prediction",
              "regulatory-action",
              "environmental-incident",
              "diplomatic-relations",
              "humanitarian-crisis",
              "trend-analysis",
              "review-critique",
              "other"
            ],
            "maxLength": 64,
            "type": "string"
          },
          "factCheckability": {
            "description": "How verifiable are the article's claims? 1=pure opinion, 5=hard data with checkable sources.",
            "maximum": 5,
            "minimum": 1,
            "type": "integer"
          },
          "impactLevel": {
            "description": "Estimated relevance and significance. 1=minor local note, 5=nationally or industry-wide relevant, 10=global crisis or historic event.",
            "maximum": 10,
            "minimum": 1,
            "type": "integer"
          },
          "impactReasoning": {
            "description": "Explanation of the impact score.",
            "maxGraphemes": 1000,
            "maxLength": 5000,
            "type": "string"
          },
          "inputTokens": {
            "description": "Number of input tokens sent to the AI model for this enrichment.",
            "type": "integer"
          },
          "language": {
            "description": "Language of this enrichment's free-text fields (BCP-47, e.g., 'es', 'en').",
            "format": "language",
            "type": "string"
          },
          "locations": {
            "description": "Geographic locations referenced in the article.",
            "items": {
              "ref": "tech.transparencia.defs#location",
              "type": "ref"
            },
            "maxLength": 20,
            "type": "array"
          },
          "modelUsed": {
            "description": "AI model identifier used for this enrichment (e.g., 'gemini-2.5-flash').",
            "maxLength": 256,
            "type": "string"
          },
          "modelVersion": {
            "description": "Specific model version or checkpoint if available.",
            "maxLength": 128,
            "type": "string"
          },
          "neutralHeadline": {
            "description": "AI-rewritten neutral headline, stripped of editorial bias.",
            "maxGraphemes": 512,
            "maxLength": 2048,
            "type": "string"
          },
          "organizationEntities": {
            "description": "Organizations mentioned in the article with structured metadata.",
            "items": {
              "ref": "tech.transparencia.defs#organization",
              "type": "ref"
            },
            "maxLength": 30,
            "type": "array"
          },
          "orientationConfidence": {
            "description": "Confidence in the orientation assessment (decimal string 0-1).",
            "maxLength": 16,
            "type": "string"
          },
          "orientationReasoning": {
            "description": "Explanation of why this orientation was assigned.",
            "maxGraphemes": 1000,
            "maxLength": 5000,
            "type": "string"
          },
          "outputTokens": {
            "description": "Number of output tokens received from the AI model for this enrichment.",
            "type": "integer"
          },
          "people": {
            "description": "People mentioned in the article with role and sentiment.",
            "items": {
              "ref": "tech.transparencia.defs#person",
              "type": "ref"
            },
            "maxLength": 50,
            "type": "array"
          },
          "politicalOrientation": {
            "description": "Detected political bias of the article's framing (not the topic itself).",
            "knownValues": [
              "left",
              "center-left",
              "center",
              "center-right",
              "right",
              "neutral"
            ],
            "maxLength": 32,
            "type": "string"
          },
          "readingLevel": {
            "description": "Estimated reading difficulty of the original article.",
            "knownValues": [
              "basic",
              "intermediate",
              "advanced"
            ],
            "maxLength": 32,
            "type": "string"
          },
          "region": {
            "description": "Geographic scope of the article's relevance.",
            "knownValues": [
              "local",
              "state",
              "national",
              "regional",
              "global"
            ],
            "maxLength": 32,
            "type": "string"
          },
          "relatedKeywords": {
            "description": "Keywords for event matching and clustering.",
            "items": {
              "maxGraphemes": 64,
              "maxLength": 256,
              "type": "string"
            },
            "maxLength": 15,
            "type": "array"
          },
          "structuredRefs": {
            "description": "Typed references to external entities (laws, stocks, products, teams, etc.).",
            "items": {
              "ref": "tech.transparencia.defs#structuredRef",
              "type": "ref"
            },
            "maxLength": 20,
            "type": "array"
          },
          "summary": {
            "description": "AI-generated summary of the article (2-3 sentences).",
            "maxGraphemes": 2000,
            "maxLength": 10000,
            "type": "string"
          },
          "timeline": {
            "description": "Chronological events mentioned in the article.",
            "items": {
              "ref": "tech.transparencia.defs#timelineEntry",
              "type": "ref"
            },
            "maxLength": 20,
            "type": "array"
          },
          "topics": {
            "description": "Main topics covered (max 10).",
            "items": {
              "maxGraphemes": 128,
              "maxLength": 512,
              "type": "string"
            },
            "maxLength": 10,
            "type": "array"
          }
        },
        "required": [
          "article",
          "summary",
          "impactLevel",
          "eventType",
          "contentDomain",
          "modelUsed",
          "createdAt"
        ],
        "type": "object"
      },
      "type": "record"
    }
  },
  "description": "AI-generated structured metadata for a news article. Uses the sidecar pattern — references the article via strongRef so multiple enrichments (from different models or analysts) can coexist.",
  "id": "tech.transparencia.news.enrichment",
  "lexicon": 1
}