# Citation Verification Pipeline Schema
# Based on REF-059 LitLLM Citation Processing
# Issue: #236

$schema: "https://json-schema.org/draft/2020-12/schema"
$id: "https://aiwg.io/schemas/citation-verification/v1"
title: "Citation Verification Pipeline Schema"
description: |
  Automated citation verification pipeline for corpus maintenance implementing
  DOI resolution, URL accessibility, metadata consistency, and format compliance
  per REF-059 LitLLM.

type: object
required:
  - version
  - verification_pipeline
  - verification_checks
  - ci_integration

properties:
  version:
    type: string
    pattern: "^\\d+\\.\\d+\\.\\d+$"
    default: "1.0.0"

  verification_pipeline:
    $ref: "#/$defs/VerificationPipeline"

  verification_checks:
    $ref: "#/$defs/VerificationChecks"

  ci_integration:
    $ref: "#/$defs/CIIntegration"

$defs:
  VerificationPipeline:
    type: object
    description: "Citation verification pipeline architecture"
    properties:
      enabled:
        type: boolean
        default: true

      triggers:
        type: array
        items: { type: string }
        default:
          - "pr_touching_corpus"
          - "manual_audit"
          - "scheduled_health_check"

      pipeline_flow:
        type: string
        default: |
          Trigger: PR touching .aiwg/research/corpus/
            ↓
          Extract citations from changed files
            ↓
          Run verification checks:
            1. DOI resolution
            2. URL accessibility
            3. Metadata consistency
            4. Format compliance
            5. Page number validation
            ↓
          Generate report:
            ✓ PASS: All checks passed
            ✗ FAIL: Errors found, PR blocked
            ⚠ WARN: Non-critical issues
            ↓
          Comment on PR with results

      status_codes:
        type: object
        properties:
          pass:
            type: object
            properties:
              symbol: { type: string, default: "✓" }
              description: { type: string, default: "All checks passed" }
              action: { type: string, default: "allow_merge" }

          fail:
            type: object
            properties:
              symbol: { type: string, default: "✗" }
              description: { type: string, default: "Errors found" }
              action: { type: string, default: "block_merge" }

          warn:
            type: object
            properties:
              symbol: { type: string, default: "⚠" }
              description: { type: string, default: "Non-critical issues" }
              action: { type: string, default: "allow_with_warning" }

  VerificationChecks:
    type: object
    description: "Individual verification checks"
    properties:
      doi_resolution:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify DOI exists and resolves"

          endpoint:
            type: string
            default: "https://doi.org/{doi}"

          timeout_ms:
            type: integer
            default: 5000

          accept_header:
            type: string
            default: "application/json"

          success_criteria:
            type: string
            default: "HTTP 200 response"

          on_failure:
            type: object
            properties:
              severity: { type: string, default: "error" }
              message: { type: string, default: "DOI does not resolve" }

          validation_pattern:
            type: string
            default: "^10\\.\\d{4,}/[^\\s]+$"
            description: "Valid DOI format pattern"

      url_accessibility:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify URL is accessible"

          method:
            type: string
            default: "HEAD"

          timeout_ms:
            type: integer
            default: 5000

          success_criteria:
            type: string
            default: "HTTP status < 400"

          on_failure:
            type: object
            properties:
              severity: { type: string, default: "error" }
              message: { type: string, default: "URL not accessible" }

          retry:
            type: object
            properties:
              max_attempts: { type: integer, default: 2 }
              delay_ms: { type: integer, default: 1000 }

      metadata_consistency:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify metadata matches DOI record"

          checks:
            type: array
            items: { type: string }
            default:
              - "title_match"
              - "year_match"
              - "authors_match"

          normalization:
            type: object
            properties:
              title:
                type: string
                default: "lowercase, remove punctuation"
              authors:
                type: string
                default: "last name only comparison"

          similarity_threshold:
            type: number
            default: 0.9
            description: "Minimum similarity for title match"

          on_mismatch:
            type: object
            properties:
              severity: { type: string, default: "error" }
              message: { type: string, default: "Metadata does not match DOI record" }

      format_compliance:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify required fields present"

          required_fields:
            type: array
            items: { type: string }
            default:
              - "Title"
              - "Authors"
              - "Year"
              - "Venue"
              - "URL"
              - "Summary"
              - "Key Quotes"
              - "AIWG Relevance"

          on_missing:
            type: object
            properties:
              severity: { type: string, default: "error" }
              message: { type: string, default: "Missing required fields" }

      page_validation:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify page numbers are valid"

          checks:
            type: array
            items: { type: string }
            default:
              - "page_within_range"
              - "section_consistency"

          on_invalid:
            type: object
            properties:
              severity: { type: string, default: "warning" }
              message: { type: string, default: "Page number may be invalid" }

      cross_reference:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          description:
            type: string
            default: "Verify cross-references resolve"

          pattern:
            type: string
            default: "REF-\\d{3}"

          on_unresolved:
            type: object
            properties:
              severity: { type: string, default: "error" }
              message: { type: string, default: "Cross-reference does not resolve" }

  CIIntegration:
    type: object
    description: "CI/CD integration configuration"
    properties:
      github_actions:
        type: object
        properties:
          workflow_name:
            type: string
            default: "Citation Verification"

          trigger:
            type: object
            properties:
              event: { type: string, default: "pull_request" }
              paths:
                type: array
                items: { type: string }
                default:
                  - ".aiwg/research/corpus/**"

          workflow_template:
            type: string
            default: |
              name: Citation Verification

              on:
                pull_request:
                  paths:
                    - ".aiwg/research/corpus/**"

              jobs:
                verify-citations:
                  runs-on: ubuntu-latest
                  steps:
                    - uses: actions/checkout@v4

                    - name: Setup Node.js
                      uses: actions/setup-node@v4
                      with:
                        node-version: "20"

                    - name: Install AIWG
                      run: npm install -g aiwg

                    - name: Run citation verification
                      id: verify
                      run: |
                        aiwg verify-citations --corpus .aiwg/research/corpus/ \
                          --format github \
                          --output verification-report.md

                    - name: Comment on PR
                      uses: actions/github-script@v7
                      with:
                        script: |
                          const fs = require("fs");
                          const report = fs.readFileSync("verification-report.md", "utf8");

                          github.rest.issues.createComment({
                            issue_number: context.issue.number,
                            owner: context.repo.owner,
                            repo: context.repo.repo,
                            body: report
                          });

                    - name: Fail if errors found
                      if: steps.verify.outputs.errors > 0
                      run: exit 1

      pre_commit_hook:
        type: object
        properties:
          enabled:
            type: boolean
            default: false

          script:
            type: string
            default: |
              #!/bin/bash
              # Verify citations in staged corpus files
              STAGED=$(git diff --cached --name-only | grep -E "\.aiwg/research/corpus/")

              if [ -n "$STAGED" ]; then
                for FILE in $STAGED; do
                  if ! aiwg verify-citation "$FILE" --quiet; then
                    echo "ERROR: Citation verification failed for $FILE"
                    exit 1
                  fi
                done
              fi

      scheduled_audit:
        type: object
        properties:
          enabled:
            type: boolean
            default: true

          cron:
            type: string
            default: "0 0 * * 0"
            description: "Weekly on Sunday"

          workflow_template:
            type: string
            default: |
              name: Corpus Health Check

              on:
                schedule:
                  - cron: "0 0 * * 0"
                workflow_dispatch:

              jobs:
                health-check:
                  runs-on: ubuntu-latest
                  steps:
                    - uses: actions/checkout@v4

                    - name: Run full corpus verification
                      run: |
                        aiwg verify-citations --corpus .aiwg/research/corpus/ \
                          --all \
                          --output corpus-health-report.md

                    - name: Upload report
                      uses: actions/upload-artifact@v4
                      with:
                        name: corpus-health-report
                        path: corpus-health-report.md

# Verification result schema
verification_result:
  type: object
  properties:
    file:
      type: string
    ref_id:
      type: string
    status:
      type: string
      enum: [pass, fail, warn]
    checks:
      type: object
      properties:
        doi_resolution:
          type: object
          properties:
            passed: { type: boolean }
            message: { type: string }
        url_accessibility:
          type: object
          properties:
            passed: { type: boolean }
            status_code: { type: integer }
            message: { type: string }
        metadata_consistency:
          type: object
          properties:
            passed: { type: boolean }
            mismatches: { type: array }
        format_compliance:
          type: object
          properties:
            passed: { type: boolean }
            missing_fields: { type: array }
        page_validation:
          type: object
          properties:
            passed: { type: boolean }
            issues: { type: array }
    error_count:
      type: integer
    warning_count:
      type: integer

# CLI commands
cli_commands:
  verify_citations:
    command: "aiwg verify-citations [path]"
    description: "Verify citations in corpus"
    options:
      - name: "--corpus"
        description: "Path to corpus directory"
      - name: "--all"
        description: "Verify entire corpus"
      - name: "--format"
        description: "Output format (text, github, json)"
      - name: "--output"
        description: "Write report to file"
      - name: "--strict"
        description: "Fail on warnings"

  verify_citation:
    command: "aiwg verify-citation <file>"
    description: "Verify single citation file"
    options:
      - name: "--quiet"
        description: "Only output on failure"

  corpus_health:
    command: "aiwg corpus-health"
    description: "Full corpus health check"
    options:
      - name: "--check-urls"
        description: "Include URL accessibility checks"
      - name: "--check-dois"
        description: "Include DOI resolution checks"

# Agent protocol
agent_protocol:
  verify_citation:
    description: "Verify single citation file"
    steps:
      - read_citation_file
      - extract_metadata
      - if_has_doi:
          - verify_doi_resolution
          - verify_metadata_consistency
      - if_has_url:
          - verify_url_accessibility
      - verify_format_compliance
      - verify_page_numbers
      - verify_cross_references
      - aggregate_results
      - return_verification_result

  verify_corpus:
    description: "Verify entire corpus"
    steps:
      - discover_corpus_files
      - for_each_file:
          - verify_citation
          - record_result
      - aggregate_results
      - generate_report
      - return_corpus_health

  generate_report:
    description: "Generate verification report"
    steps:
      - collect_all_results
      - calculate_summary
      - format_by_output_type
      - include_recommendations
      - return_report

# Storage
storage:
  verification_logs: ".aiwg/logs/citation-verification/"
  health_reports: ".aiwg/reports/corpus-health/"

# Research targets (from REF-059)
research_targets:
  doi_validation: "Verify DOIs exist and resolve correctly"
  url_health: "Ensure all URLs are accessible"
  metadata_accuracy: "Confirm metadata matches source records"
  format_compliance: "Enforce required field presence"
  page_validation: "Validate page number accuracy"

# Example verification report
example_verification_report: |
  ## Citation Verification Report

  **Status**: ✗ FAILED (2 errors, 1 warning)

  ### REF-043-voice-consistency-quality.md

  ✓ DOI resolves (10.1145/3586183.3606763)
  ✓ URL accessible (200)
  ✓ Metadata matches DOI record
  ✓ Format compliance
  ✓ Page numbers valid (24 total)

  **Status**: PASS

  ### REF-999-new-paper.md

  ✗ DOI does not resolve (10.1234/fake.doi)
  ✓ URL accessible (200)
  ✗ Metadata mismatch: Year 2024 vs 2023 in DOI record
  ⚠ Missing field: "AIWG Relevance"
  ✓ Page numbers valid (15 total)

  **Status**: FAIL (2 errors, 1 warning)

  ---

  **Summary**:
  - Total files: 2
  - Passed: 1
  - Failed: 1
  - Errors: 2
  - Warnings: 1

  **Action Required**: Fix 2 errors before merge.

# References
references:
  research:
    - "@.aiwg/research/findings/REF-059-litllm-citation-processing.md"
  implementation:
    - "#236"
  related:
    - "@agentic/code/frameworks/sdlc-complete/schemas/flows/citation-integrity.yaml"
    - "@agentic/code/frameworks/sdlc-complete/schemas/flows/fair-metadata.yaml"
    - "@.aiwg/research/corpus/"