AIclinicalresearch/backend/scripts/analyze_missing_equery_context.ts

import { PrismaClient } from '@prisma/client';

const prisma = new PrismaClient();

async function main() {
  const projectId = process.argv[2];
  if (!projectId) {
    throw new Error('Usage: npx tsx scripts/analyze_missing_equery_context.ts <projectId>');
  }

  const reasonStats = await prisma.$queryRawUnsafe<Array<{ reason: string; cnt: bigint }>>(`
    WITH missing AS (
      SELECT
        e.id,
        e.record_id,
        e.category,
        e.field_name,
        e.event_id,
        e.form_name,
        e.created_at
      FROM iit_schema.equery e
      WHERE e.project_id = $1
        AND (COALESCE(e.event_id, '') = '' OR COALESCE(e.form_name, '') = '')
    ),
    flags AS (
      SELECT
        m.*,
        EXISTS (
          SELECT 1
          FROM iit_schema.qc_field_status fs
          WHERE fs.project_id = $1
            AND fs.record_id = m.record_id
        ) AS has_record_match,
        EXISTS (
          SELECT 1
          FROM iit_schema.qc_field_status fs
          WHERE fs.project_id = $1
            AND fs.record_id = m.record_id
            AND COALESCE(fs.rule_name, '') = COALESCE(m.category, '')
        ) AS has_rule_match,
        EXISTS (
          SELECT 1
          FROM iit_schema.qc_field_status fs
          WHERE fs.project_id = $1
            AND fs.record_id = m.record_id
            AND COALESCE(fs.rule_name, '') = COALESCE(m.category, '')
            AND COALESCE(fs.field_name, '') = COALESCE(m.field_name, '')
        ) AS has_rule_field_match
      FROM missing m
    )
    SELECT
      CASE
        WHEN has_rule_field_match THEN 'A_RULE_FIELD_MATCH_BUT_EVENT_FORM_EMPTY'
        WHEN has_rule_match THEN 'B_RULE_MATCH_FIELD_MISMATCH'
        WHEN has_record_match THEN 'C_RECORD_MATCH_ONLY'
        ELSE 'D_NO_RECORD_MATCH'
      END AS reason,
      COUNT(*)::bigint AS cnt
    FROM flags
    GROUP BY 1
    ORDER BY 2 DESC
  `, projectId);

  const sample = await prisma.$queryRawUnsafe<Array<{
    id: string;
    record_id: string;
    category: string | null;
    field_name: string | null;
    event_id: string | null;
    form_name: string | null;
    created_at: Date;
  }>>(`
    SELECT
      id, record_id, category, field_name, event_id, form_name, created_at
    FROM iit_schema.equery
    WHERE project_id = $1
      AND (COALESCE(event_id, '') = '' OR COALESCE(form_name, '') = '')
    ORDER BY created_at DESC
    LIMIT 20
  `, projectId);

  console.log(JSON.stringify({
    projectId,
    reasonStats: reasonStats.map((r) => ({ reason: r.reason, cnt: Number(r.cnt) })),
    sample,
  }, null, 2));
}

main()
  .catch((e) => {
    console.error(e);
    process.exit(1);
  })
  .finally(async () => {
    await prisma.$disconnect();
  });