feat: promote all indexed documents public regardless of validation status

scripts/promote-public-documents-batch.ts

@@ -4,6 +4,14 @@ loadEnvConfig(process.cwd());
 
 const BATCH_SIZE = 25;
 
+function withPublicCorpusMetadata(metadata: unknown): Record<string, unknown> {
+  const base =
+    typeof metadata === "object" && metadata !== null && !Array.isArray(metadata)
+      ? (metadata as Record<string, unknown>)
+      : {};
+  return { ...base, public_corpus: true };
+}
+
 async function loadAdminClient() {
   const { createAdminClient } = await import("@/lib/supabase/admin");
   return createAdminClient();
@@ -25,7 +33,6 @@ async function fetchBatchIds(supabase: Awaited<ReturnType<typeof loadAdminClient
     .select("id")
     .eq("status", "indexed")
     .not("owner_id", "is", null)
-    .in("metadata->>clinical_validation_status", ["locally_reviewed", "approved"])
     .limit(BATCH_SIZE);
   if (error) throw new Error(error.message);
   return (data ?? []).map((row) => row.id as string);
@@ -34,14 +41,37 @@ async function fetchBatchIds(supabase: Awaited<ReturnType<typeof loadAdminClient
 async function promoteBatch(supabase: Awaited<ReturnType<typeof loadAdminClient>>, ids: string[]) {
   if (ids.length === 0) return;
 
-  const { error } = await supabase
-    .from("documents")
-    .update({
-      owner_id: null,
-      updated_at: new Date().toISOString(),
-    })
-    .in("id", ids);
-  if (error) throw new Error(error.message);
+  const updatedAt = new Date().toISOString();
+  const { data: documents, error: fetchError } = await supabase.from("documents").select("id, metadata").in("id", ids);
+  if (fetchError) throw new Error(fetchError.message);
+
+  await Promise.all(
+    (documents ?? []).map(async (document) => {
+      const { error } = await supabase
+        .from("documents")
+        .update({
+          owner_id: null,
+          metadata: withPublicCorpusMetadata(document.metadata),
+          updated_at: updatedAt,
+        })
+        .eq("id", document.id);
+      if (error) throw new Error(error.message);
+    }),
+  );
+
+  const artifactUpdates = await Promise.all([
+    supabase.from("document_labels").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+    supabase.from("document_summaries").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+    supabase.from("document_sections").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+    supabase.from("document_memory_cards").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+    supabase.from("document_table_facts").update({ owner_id: null }).in("document_id", ids),
+    supabase.from("document_embedding_fields").update({ owner_id: null }).in("document_id", ids),
+    supabase.from("document_index_quality").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+    supabase.from("document_index_units").update({ owner_id: null, updated_at: updatedAt }).in("document_id", ids),
+  ]);
+  for (const { error } of artifactUpdates) {
+    if (error) throw new Error(error.message);
+  }
 }
 
 async function main() {

scripts/promote-public-documents.ts

@@ -34,8 +34,7 @@ async function countCandidates(supabase: Awaited<ReturnType<typeof loadAdminClie
     .from("documents")
     .select("id", { count: "exact", head: true })
     .eq("status", "indexed")
-    .not("owner_id", "is", null)
-    .in("metadata->>clinical_validation_status", ["locally_reviewed", "approved"]);
+    .not("owner_id", "is", null);
 
   if (ownerId) query = query.eq("owner_id", ownerId);
 
@@ -70,7 +69,7 @@ async function main() {
     candidateCount,
   );
   console.log(
-    "Apply migration 20260705220000_promote_locally_reviewed_documents_public.sql with `npx supabase db push --linked`.",
+    "Promote remaining owned indexed documents with `npm run promote:public-documents:batch` or apply migration 20260706120000_promote_remaining_indexed_documents_public.sql.",
   );
 }
 

supabase/migrations/20260706120000_promote_remaining_indexed_documents_public.sql

@@ -0,0 +1,64 @@
+-- Promote all remaining indexed owned documents to the public corpus (owner_id IS NULL)
+-- regardless of clinical_validation_status, so anonymous callers can access the full indexed corpus.
+
+begin;
+
+create temporary table promoted_public_documents on commit drop as
+with promoted as (
+  update public.documents d
+  set
+    owner_id = null,
+    metadata = jsonb_set(
+      coalesce(d.metadata, '{}'::jsonb),
+      '{public_corpus}',
+      'true'::jsonb,
+      true
+    ),
+    updated_at = now()
+  where d.status = 'indexed'
+    and d.owner_id is not null
+  returning d.id, d.owner_id as previous_owner_id
+)
+select id, previous_owner_id from promoted;
+
+update public.document_labels dl
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where dl.document_id = pd.id;
+
+update public.document_summaries ds
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where ds.document_id = pd.id;
+
+update public.document_sections ds
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where ds.document_id = pd.id;
+
+update public.document_memory_cards dmc
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where dmc.document_id = pd.id;
+
+update public.document_table_facts dtf
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where dtf.document_id = pd.id;
+
+update public.document_embedding_fields def
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where def.document_id = pd.id;
+
+update public.document_index_quality diq
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where diq.document_id = pd.id;
+
+update public.document_index_units diu
+set owner_id = null, updated_at = now()
+from promoted_public_documents pd
+where diu.document_id = pd.id;
+
+commit;