From d6f558ca9130a9a4f3748ee4837229687cbab271 Mon Sep 17 00:00:00 2001
From: BigSimmo <87357024+BigSimmo@users.noreply.github.com>
Date: Thu, 2 Jul 2026 17:17:51 +0800
Subject: [PATCH] chore: remove accidentally committed scratch/ debugging
 files; gitignore scratch/

The 12 files under scratch/ (one-off SQL and TS debugging scripts) landed
on main via the 'Save Codex local changes' commit (0aa8f5d7e) and were
never meant to be tracked. Remove them and gitignore scratch/ so local
debugging files cannot be committed again. The tsconfig 'scratch/**'
exclude from PR #139 stays: tsc does not respect .gitignore, so the
exclude keeps untracked local scratch files out of typecheck.

Files remain recoverable from git history if any turn out to matter.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
---
 .gitignore                      |   2 +
 scratch/apply-fast-schema.sql   | 575 ------------------------------
 scratch/audit-guidelines.ts     | 107 ------
 scratch/check-indexes.ts        |  22 --
 scratch/check-queued-docs.ts    |  22 --
 scratch/count-files.ts          |  59 ----
 scratch/count-status.ts         |  54 ---
 scratch/create-hnsw-indexes.sql |  21 --
 scratch/fix-indexes.ts          |  15 -
 scratch/queue-queued-docs.ts    |  57 ---
 scratch/run-migrations-safe.sql | 603 --------------------------------
 scratch/show-samples.ts         |  22 --
 scratch/update-health-check.sql | 132 -------
 13 files changed, 2 insertions(+), 1689 deletions(-)
 delete mode 100644 scratch/apply-fast-schema.sql
 delete mode 100644 scratch/audit-guidelines.ts
 delete mode 100644 scratch/check-indexes.ts
 delete mode 100644 scratch/check-queued-docs.ts
 delete mode 100644 scratch/count-files.ts
 delete mode 100644 scratch/count-status.ts
 delete mode 100644 scratch/create-hnsw-indexes.sql
 delete mode 100644 scratch/fix-indexes.ts
 delete mode 100644 scratch/queue-queued-docs.ts
 delete mode 100644 scratch/run-migrations-safe.sql
 delete mode 100644 scratch/show-samples.ts
 delete mode 100644 scratch/update-health-check.sql

diff --git a/.gitignore b/.gitignore
index d6b6f2e68..25e923da9 100644
--- a/.gitignore
+++ b/.gitignore
@@ -65,6 +65,8 @@ next-env.d.ts
 
 # agent/QA artifacts
 .codex-screenshots/
+# Local debugging scratch space — never commit (accidentally landed once via 'Save Codex local changes')
+scratch/
 .qa-smoke/
 *.pid
 tmp_output.txt
diff --git a/scratch/apply-fast-schema.sql b/scratch/apply-fast-schema.sql
deleted file mode 100644
index e76dbc9a4..000000000
--- a/scratch/apply-fast-schema.sql
+++ /dev/null
@@ -1,575 +0,0 @@
-SET search_path = public, extensions;
-
--- ============================================================
--- 1. RAG RETRIEVAL LOGS TABLE
--- Per-query diagnostics: candidate scores, latency breakdown,
--- miss detection.
--- ============================================================
-
-create table if not exists public.rag_retrieval_logs (
-  id uuid primary key default gen_random_uuid(),
-  owner_id uuid references auth.users(id) on delete set null,
-  query text not null,
-  normalized_query text,
-  query_class text,
-  retrieval_strategy text,
-
-  -- Candidate scores
-  candidate_count integer not null default 0,
-  top_similarity double precision,
-  top_text_rank double precision,
-  top_hybrid_score double precision,
-  top_rrf_score double precision,
-  mean_hybrid_score double precision,
-
-  -- Selected citations
-  selected_chunk_ids uuid[] not null default '{}',
-  selected_document_ids uuid[] not null default '{}',
-  selected_count integer not null default 0,
-
-  -- Latency breakdown
-  embedding_latency_ms integer,
-  rpc_latency_ms integer,
-  rerank_latency_ms integer,
-  total_latency_ms integer,
-
-  -- Source counts
-  vector_candidate_count integer,
-  text_candidate_count integer,
-  memory_card_count integer,
-  index_unit_count integer,
-  embedding_field_count integer,
-
-  -- Miss detection
-  is_miss boolean not null default false,
-  miss_reason text,
-
-  -- Metadata
-  embedding_cache_hit boolean,
-  metadata jsonb not null default '{}'::jsonb,
-  created_at timestamptz not null default now()
-);
-
-create index if not exists rag_retrieval_logs_owner_created_idx
-  on public.rag_retrieval_logs(owner_id, created_at desc);
-create index if not exists rag_retrieval_logs_miss_idx
-  on public.rag_retrieval_logs(is_miss, created_at desc)
-  where is_miss = true;
-create index if not exists rag_retrieval_logs_strategy_idx
-  on public.rag_retrieval_logs(retrieval_strategy, created_at desc);
-
-alter table public.rag_retrieval_logs enable row level security;
-grant select, insert, update, delete on table public.rag_retrieval_logs to service_role;
-grant select on table public.rag_retrieval_logs to authenticated;
-
-create policy "rag retrieval logs owner read" on public.rag_retrieval_logs
-  for select to authenticated using (owner_id = (select auth.uid()));
-
--- ============================================================
--- 2. CONTENT_HASH DEDUP ON DOCUMENT_EMBEDDING_FIELDS
--- Replace the problematic unique(document_id, source_chunk_id,
--- field_type, content) constraint that indexes full text.
--- ============================================================
-
-alter table public.document_embedding_fields
-  add column if not exists content_hash text;
-
--- Populate for existing rows
-update public.document_embedding_fields
-  set content_hash = md5(content)
-  where content_hash is null;
-
--- Drop the old constraint.
-alter table public.document_embedding_fields
-  drop constraint if exists document_embedding_fields_document_id_source_chunk_id_field_key;
-alter table public.document_embedding_fields
-  drop constraint if exists document_embedding_fields_document_id_source_chunk_id_fiel_key;
-
--- Create hash-based unique index
-create unique index if not exists document_embedding_fields_dedup_idx
-  on public.document_embedding_fields(document_id, source_chunk_id, field_type, content_hash);
-
--- ============================================================
--- 3. DROP DUPLICATE/REDUNDANT INDEXES
--- These single-column indexes are covered by composite indexes
--- with the same leading column.
--- ============================================================
-
-drop index if exists document_chunks_document_id_idx;
-drop index if exists document_sections_document_id_idx;
-drop index if exists document_memory_cards_document_id_idx;
-drop index if exists document_images_document_id_idx;
-drop index if exists document_labels_document_id_idx;
-drop index if exists document_embedding_fields_document_id_idx;
-drop index if exists document_table_facts_document_id_idx;
-drop index if exists document_index_quality_document_id_idx;
-
--- ============================================================
--- 4. OTHER EXPECTED SCHEMA INDEXES FROM OTHER MIGRATIONS
--- ============================================================
-create index if not exists ingestion_jobs_status_next_run_idx
-  on public.ingestion_jobs(status, next_run_at asc)
-  where status in ('pending', 'processing');
-
-create index if not exists ingestion_jobs_document_status_idx
-  on public.ingestion_jobs(document_id, status);
-
-create index if not exists import_batches_status_created_idx
-  on public.import_batches(status, created_at desc);
-
-create index if not exists storage_cleanup_jobs_status_created_idx
-  on public.storage_cleanup_jobs(status, created_at desc);
-
--- ============================================================
--- 5. ENHANCED HYBRID SCORING
--- ============================================================
-
-create or replace function public.match_document_chunks_hybrid(
-  query_embedding extensions.vector(1536),
-  query_text text,
-  match_count integer default 12,
-  min_similarity double precision default 0.12,
-  document_filters uuid[] default null,
-  owner_filter uuid default null
-)
-returns table (
-  id uuid,
-  document_id uuid,
-  title text,
-  file_name text,
-  page_number integer,
-  chunk_index integer,
-  section_heading text,
-  content text,
-  retrieval_synopsis text,
-  image_ids uuid[],
-  source_metadata jsonb,
-  similarity double precision,
-  text_rank double precision,
-  hybrid_score double precision,
-  rrf_score double precision,
-  images jsonb
-)
-language sql
-stable
-set search_path = public, extensions, pg_temp
-as $$
-  with query as (
-    select websearch_to_tsquery('english', coalesce(query_text, '')) as tsq
-  ),
-  vector_ranked as (
-    select
-      c.id,
-      c.document_id,
-      c.page_number,
-      c.chunk_index,
-      c.section_heading,
-      c.content,
-      c.retrieval_synopsis,
-      c.image_ids,
-      1 - (c.embedding <=> query_embedding) as similarity,
-      (
-        ts_rank_cd(c.search_tsv, query.tsq) +
-        (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-      )::double precision as text_rank,
-      row_number() over (order by c.embedding <=> query_embedding) as vector_rank,
-      null::bigint as text_match_rank,
-      coalesce((d.metadata->'rag_indexing_version') is not null, false) as has_deep_index,
-      d.updated_at as doc_updated_at,
-      coalesce(
-        (select q.quality_score from public.document_index_quality q where q.document_id = c.document_id),
-        0.7
-      ) as quality_score
-    from public.document_chunks c
-    join public.documents d on d.id = c.document_id
-    cross join query
-    where (document_filters is null or c.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and d.status = 'indexed'
-      and 1 - (c.embedding <=> query_embedding) >= min_similarity
-    order by c.embedding <=> query_embedding
-    limit greatest(match_count * 6, 48)
-  ),
-  text_ranked as (
-    select
-      c.id,
-      c.document_id,
-      c.page_number,
-      c.chunk_index,
-      c.section_heading,
-      c.content,
-      c.retrieval_synopsis,
-      c.image_ids,
-      1 - (c.embedding <=> query_embedding) as similarity,
-      (
-        ts_rank_cd(c.search_tsv, query.tsq) +
-        (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-      )::double precision as text_rank,
-      null::bigint as vector_rank,
-      row_number() over (
-        order by
-          (
-            ts_rank_cd(c.search_tsv, query.tsq) +
-            (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-          ) desc,
-          c.embedding <=> query_embedding
-      ) as text_match_rank,
-      coalesce((d.metadata->'rag_indexing_version') is not null, false) as has_deep_index,
-      d.updated_at as doc_updated_at,
-      coalesce(
-        (select q.quality_score from public.document_index_quality q where q.document_id = c.document_id),
-        0.7
-      ) as quality_score
-    from public.document_chunks c
-    join public.documents d on d.id = c.document_id
-    cross join query
-    where (document_filters is null or c.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and d.status = 'indexed'
-      and (c.search_tsv @@ query.tsq or d.title_search_tsv @@ query.tsq)
-    order by (
-      ts_rank_cd(c.search_tsv, query.tsq) +
-      (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-    ) desc
-    limit greatest(match_count * 6, 48)
-  ),
-  combined as (
-    select * from vector_ranked
-    union all
-    select * from text_ranked
-  ),
-  scored as (
-    select
-      id,
-      document_id,
-      page_number,
-      chunk_index,
-      section_heading,
-      content,
-      retrieval_synopsis,
-      image_ids,
-      max(similarity)::double precision as similarity,
-      max(text_rank)::double precision as text_rank,
-      min(vector_rank) as vector_rank,
-      min(text_match_rank) as text_match_rank,
-      max(quality_score)::double precision as quality_score,
-      bool_or(has_deep_index) as has_deep_index,
-      max(doc_updated_at) as doc_updated_at
-    from combined
-    group by id, document_id, page_number, chunk_index, section_heading, content, retrieval_synopsis, image_ids
-  ),
-  scored_metrics as (
-    select
-      scored.*,
-      (
-        (scored.similarity * 0.62)
-        + (least(scored.text_rank, 1) * 0.22)
-        + (scored.quality_score * 0.10)
-        + (case when scored.doc_updated_at > now() - interval '90 days' then 0.06 else 0 end)
-      )::double precision as hybrid_score,
-      (
-        coalesce(1.0 / (60 + scored.vector_rank), 0) +
-        coalesce(1.0 / (60 + scored.text_match_rank), 0)
-      )::double precision as rrf_score
-    from scored
-  ),
-  hybrid_candidates as (
-    select id
-    from scored_metrics
-    order by hybrid_score desc, similarity desc, text_rank desc
-    limit match_count
-  ),
-  vector_candidates as (
-    select id
-    from scored_metrics
-    order by similarity desc, hybrid_score desc
-    limit match_count
-  ),
-  text_candidates as (
-    select id
-    from scored_metrics
-    order by text_rank desc, hybrid_score desc
-    limit match_count
-  ),
-  rrf_candidates as (
-    select id
-    from scored_metrics
-    order by rrf_score desc, hybrid_score desc
-    limit match_count
-  ),
-  candidate_ids as (
-    select id from hybrid_candidates
-    union
-    select id from vector_candidates
-    union
-    select id from text_candidates
-    union
-    select id from rrf_candidates
-  )
-  select
-    c.id,
-    c.document_id,
-    d.title,
-    d.file_name,
-    c.page_number,
-    c.chunk_index,
-    c.section_heading,
-    c.content,
-    c.retrieval_synopsis,
-    c.image_ids,
-    d.metadata as source_metadata,
-    c.similarity,
-    c.text_rank,
-    c.hybrid_score,
-    c.rrf_score,
-    public.chunk_image_metadata(c.image_ids) as images
-  from scored_metrics c
-  join candidate_ids candidates on candidates.id = c.id
-  join public.documents d on d.id = c.document_id
-  order by c.hybrid_score desc, c.rrf_score desc, c.similarity desc, c.text_rank desc
-  limit match_count;
-$$;
-
-create or replace function public.match_document_index_units_hybrid(
-  query_embedding extensions.vector(1536),
-  query_text text,
-  match_count integer default 24,
-  min_similarity double precision default 0.1,
-  document_filters uuid[] default null,
-  owner_filter uuid default null
-)
-returns table (
-  id uuid,
-  document_id uuid,
-  source_chunk_id uuid,
-  source_image_id uuid,
-  unit_type text,
-  title text,
-  content text,
-  page_start integer,
-  page_end integer,
-  heading_path text[],
-  normalized_terms text[],
-  source_span jsonb,
-  quality_score real,
-  extraction_mode text,
-  similarity double precision,
-  text_rank double precision,
-  hybrid_score double precision,
-  metadata jsonb
-)
-language sql
-stable
-set search_path = public, extensions, pg_temp
-as $$
-  with query as (
-    select websearch_to_tsquery('english', coalesce(query_text, '')) as tsq,
-      regexp_split_to_array(lower(coalesce(query_text, '')), '\s+') as terms
-  ),
-  ranked as (
-    select u.id, u.document_id, u.source_chunk_id, u.source_image_id, u.unit_type, u.title, u.content, u.page_start,
-      u.page_end, u.heading_path, u.normalized_terms, u.source_span, u.quality_score, u.extraction_mode,
-      (1 - (u.embedding <=> query_embedding))::double precision as similarity,
-      (ts_rank_cd(u.search_tsv, query.tsq)
-        + case when u.normalized_terms && query.terms then 0.25 else 0 end
-        + case when u.unit_type in ('askable_question', 'table_fact', 'clinical_fact') then 0.06
-               when u.unit_type = 'section_summary' then 0.03
-               else 0 end
-      )::double precision as text_rank,
-      u.metadata
-    from public.document_index_units u
-    join public.documents d on d.id = u.document_id
-    cross join query
-    where d.status = 'indexed'
-      and (document_filters is null or u.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and u.source_chunk_id is not null
-      and (u.search_tsv @@ query.tsq or u.normalized_terms && query.terms)
-    order by text_rank desc, similarity desc
-    limit greatest(match_count * 3, 48)
-  )
-  select id, document_id, source_chunk_id, source_image_id, unit_type, title, content, page_start, page_end, heading_path,
-    normalized_terms, source_span, quality_score, extraction_mode, similarity, text_rank,
-    (
-      (similarity * 0.52)
-      + (least(text_rank, 1) * 0.28)
-      + (quality_score * 0.12)
-      + (case when extraction_mode in ('model_heavy', 'hybrid') then 0.04 else 0 end)
-      + (case when unit_type = 'askable_question' then 0.04 else 0 end)
-    )::double precision as hybrid_score,
-    metadata
-  from ranked
-  order by hybrid_score desc, similarity desc, text_rank desc
-  limit match_count;
-$$;
-
-create or replace function public.analyze_rag_tables()
-returns void
-language plpgsql
-set search_path = public, extensions, pg_temp
-as $$
-begin
-  analyze public.document_chunks;
-  analyze public.document_memory_cards;
-  analyze public.document_index_units;
-  analyze public.document_embedding_fields;
-  analyze public.document_table_facts;
-  analyze public.documents;
-end;
-$$;
-
-revoke execute on function public.analyze_rag_tables() from public, anon, authenticated;
-grant execute on function public.analyze_rag_tables() to service_role;
-
-create or replace function public.reset_document_index(p_document_id uuid)
-returns void
-language plpgsql
-set search_path = public, extensions, pg_temp
-as $$
-begin
-  perform set_config('statement_timeout', '180000', true);
-  delete from public.document_index_units where document_id = p_document_id;
-  delete from public.document_memory_cards where document_id = p_document_id;
-  delete from public.document_sections where document_id = p_document_id;
-  delete from public.document_table_facts where document_id = p_document_id;
-  delete from public.document_embedding_fields where document_id = p_document_id;
-  delete from public.document_index_quality where document_id = p_document_id;
-  delete from public.document_chunks where document_id = p_document_id;
-  delete from public.document_images where document_id = p_document_id;
-  delete from public.document_pages where document_id = p_document_id;
-end;
-$$;
-
--- Update health check
-create or replace function public.search_schema_health()
-returns jsonb
-language plpgsql
-stable
-set search_path = public, extensions, pg_catalog, pg_temp
-as $$
-declare
-  missing text[] := array[]::text[];
-  vector_type_oid oid;
-  vector_schema text;
-  index_name text;
-begin
-  select t.oid, n.nspname
-  into vector_type_oid, vector_schema
-  from pg_type t
-  join pg_namespace n on n.oid = t.typnamespace
-  where t.typname = 'vector'
-    and n.nspname = 'extensions'
-  limit 1;
-
-  if vector_type_oid is null then
-    missing := array_append(missing, 'extensions.vector_type');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_memory_cards_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_memory_cards_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_index_units_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_index_units_hybrid.extensions_vector_signature');
-  end if;
-
-  if to_regclass('public.document_index_units') is null then
-    missing := array_append(missing, 'document_index_units.table');
-  end if;
-  if to_regclass('public.rag_retrieval_logs') is null then
-    missing := array_append(missing, 'rag_retrieval_logs.table');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'documents_title_trgm_idx') then
-    missing := array_append(missing, 'documents_title_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_chunks_content_trgm_idx') then
-    missing := array_append(missing, 'document_chunks_content_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_labels_label_trgm_idx') then
-    missing := array_append(missing, 'document_labels_label_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_summaries_summary_trgm_idx') then
-    missing := array_append(missing, 'document_summaries_summary_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_index_units_embedding_hnsw_idx') then
-    missing := array_append(missing, 'document_index_units_embedding_hnsw_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_embedding_fields_owner_idx', 'document_embedding_fields_owner_id_idx')) then
-    missing := array_append(missing, 'document_embedding_fields_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_table_facts_owner_idx', 'document_table_facts_owner_id_idx')) then
-    missing := array_append(missing, 'document_table_facts_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_table_facts_source_image_idx') then
-    missing := array_append(missing, 'document_table_facts_source_image_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_embedding_fields_dedup_idx') then
-    missing := array_append(missing, 'document_embedding_fields_dedup_idx');
-  end if;
-  foreach index_name in array array[
-    'document_pages_document_idx',
-    'document_images_document_idx',
-    'document_sections_document_idx',
-    'document_memory_cards_document_idx',
-    'document_chunks_document_idx',
-    'document_table_facts_document_idx',
-    'document_embedding_fields_document_idx',
-    'document_index_units_document_idx',
-    'ingestion_jobs_status_next_run_idx',
-    'ingestion_jobs_document_status_idx',
-    'documents_owner_status_idx',
-    'import_batches_status_created_idx',
-    'storage_cleanup_jobs_status_created_idx'
-  ] loop
-    if not exists (select 1 from pg_class where relname = index_name) then
-      missing := array_append(missing, index_name);
-    end if;
-  end loop;
-
-  return jsonb_build_object(
-    'ok', cardinality(missing) = 0,
-    'missing', missing,
-    'vector_extension_schema', vector_schema,
-    'checked_at', now()
-  );
-end;
-$$;
-
-revoke execute on function public.search_schema_health() from public, anon, authenticated;
-grant execute on function public.search_schema_health() to service_role;
diff --git a/scratch/audit-guidelines.ts b/scratch/audit-guidelines.ts
deleted file mode 100644
index 985389e68..000000000
--- a/scratch/audit-guidelines.ts
+++ /dev/null
@@ -1,107 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-import * as fs from "fs";
-import * as path from "path";
-
-loadEnvConfig(process.cwd());
-
-function getFilesRecursively(dir: string): string[] {
-  let results: string[] = [];
-  if (!fs.existsSync(dir)) return [];
-  const list = fs.readdirSync(dir);
-  for (const file of list) {
-    const filePath = path.join(dir, file);
-    const stat = fs.statSync(filePath);
-    if (stat && stat.isDirectory()) {
-      results = results.concat(getFilesRecursively(filePath));
-    } else {
-      if (/\.(pdf|docx|doc)$/i.test(file)) {
-        results.push(filePath);
-      }
-    }
-  }
-  return results;
-}
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  // 1. Scan the local directory
-  const rootDir = "C:\\Users\\joshs\\OneDrive\\Medicine\\Guidelines";
-  console.log(`Scanning local files in ${rootDir}...`);
-  const localFiles = getFilesRecursively(rootDir);
-  console.log(`Found ${localFiles.length} local guideline files (PDF/DOC/DOCX).`);
-  
-  // 2. Fetch all documents from the database
-  console.log("Fetching documents from Supabase...");
-  const { data: dbDocs, error } = await supabase
-    .from("documents")
-    .select("id, title, status, metadata");
-    
-  if (error) {
-    console.error("Error fetching database documents:", error);
-    return;
-  }
-  
-  console.log(`Found ${dbDocs?.length || 0} documents in Supabase.`);
-  
-  // Create a map of lowercased source path -> db doc
-  const dbDocsByPath = new Map<string, typeof dbDocs[number]>();
-  for (const doc of dbDocs || []) {
-    const meta = doc.metadata && typeof doc.metadata === "object" ? (doc.metadata as any) : {};
-    if (meta.source_path) {
-      dbDocsByPath.set(meta.source_path.toLowerCase(), doc);
-    }
-  }
-  
-  // 3. Match them up
-  const categories = ["BMJ", "EMHS", "KEMH", "NMHS", "PHC", "RKPG", "SMHS"];
-  const stats: Record<string, { total: number; indexed: number; processing: number; queued: number; missing: number; missingFiles: string[] }> = {};
-  
-  for (const cat of categories) {
-    stats[cat] = { total: 0, indexed: 0, processing: 0, queued: 0, missing: 0, missingFiles: [] };
-  }
-  
-  for (const file of localFiles) {
-    // Determine category from path
-    const relative = path.relative(rootDir, file);
-    const topFolder = relative.split(path.sep)[0];
-    
-    if (stats[topFolder]) {
-      stats[topFolder].total++;
-      const matched = dbDocsByPath.get(file.toLowerCase());
-      if (matched) {
-        if (matched.status === "indexed") {
-          stats[topFolder].indexed++;
-        } else if (matched.status === "processing") {
-          stats[topFolder].processing++;
-        } else if (matched.status === "queued") {
-          stats[topFolder].queued++;
-        } else {
-          stats[topFolder].missing++;
-          stats[topFolder].missingFiles.push(relative);
-        }
-      } else {
-        stats[topFolder].missing++;
-        stats[topFolder].missingFiles.push(relative);
-      }
-    }
-  }
-  
-  console.log("=== Auditing Summary by Directory ===");
-  for (const [cat, info] of Object.entries(stats)) {
-    console.log(`\nDirectory: ${cat}`);
-    console.log(`  Total Local Files: ${info.total}`);
-    console.log(`  Indexed (Done):    ${info.indexed}`);
-    console.log(`  Processing:        ${info.processing}`);
-    console.log(`  Queued:            ${info.queued}`);
-    console.log(`  Not in Database:   ${info.missing}`);
-    if (info.missing > 0 && info.missing <= 5) {
-      console.log(`  Missing files:`, info.missingFiles);
-    } else if (info.missing > 5) {
-      console.log(`  Missing files (first 5):`, info.missingFiles.slice(0, 5));
-    }
-  }
-}
-
-main().catch(console.error);
diff --git a/scratch/check-indexes.ts b/scratch/check-indexes.ts
deleted file mode 100644
index b31ad49b0..000000000
--- a/scratch/check-indexes.ts
+++ /dev/null
@@ -1,22 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  // We can query pg_indexes view using Supabase client if it is exposed, or we can use RPC
-  // Wait, let's see if we can query from a system catalog or check what indexes exist.
-  // Actually, pg_indexes might not be directly exposed. Let's see if we can do a select from pg_indexes.
-  const { data, error } = await supabase
-    .from("pg_indexes") // this might fail if not in public schema, but system catalogs aren't exposed in PostgREST by default.
-    .select("*");
-    
-  if (error) {
-    console.error("System query error (expected if pg_indexes is not exposed):", error.message);
-  } else {
-    console.log(data);
-  }
-}
-
-main().catch(console.error);
diff --git a/scratch/check-queued-docs.ts b/scratch/check-queued-docs.ts
deleted file mode 100644
index a0b962b9a..000000000
--- a/scratch/check-queued-docs.ts
+++ /dev/null
@@ -1,22 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  const { data: jobs, error } = await supabase
-    .from("ingestion_jobs")
-    .select("id, status, error_message, attempt_count, locked_at, document_id, documents(title)")
-    .in("status", ["pending", "processing"]);
-    
-  if (error) {
-    console.error(error);
-    return;
-  }
-  
-  console.log("=== Active Ingestion Jobs ===");
-  console.log(JSON.stringify(jobs, null, 2));
-}
-
-main().catch(console.error);
diff --git a/scratch/count-files.ts b/scratch/count-files.ts
deleted file mode 100644
index a8a73dc94..000000000
--- a/scratch/count-files.ts
+++ /dev/null
@@ -1,59 +0,0 @@
-import { readdir, stat } from "node:fs/promises";
-import path from "node:path";
-
-const paths = [
-  'C:\\Users\\joshs\\OneDrive\\Medicine\\Guidelines\\NMHS',
-  'C:\\Users\\joshs\\OneDrive\\Medicine\\Guidelines\\PHC',
-  'C:\\Users\\joshs\\OneDrive\\Medicine\\Guidelines\\RKPG',
-  'C:\\Users\\joshs\\OneDrive\\Medicine\\Guidelines\\SMHS'
-];
-
-const extensions = [".pdf", ".docx", ".xlsx", ".txt"];
-
-async function countFiles(dir: string): Promise<{ count: number; bytes: number }> {
-  let count = 0;
-  let bytes = 0;
-
-  async function visit(current: string) {
-    let entries;
-    try {
-      entries = await readdir(current, { withFileTypes: true });
-    } catch {
-      return;
-    }
-    for (const entry of entries) {
-      const fullPath = path.join(current, entry.name);
-      if (entry.isDirectory()) {
-        await visit(fullPath);
-      } else if (entry.isFile()) {
-        const ext = path.extname(entry.name).toLowerCase();
-        if (extensions.includes(ext)) {
-          count++;
-          try {
-            const s = await stat(fullPath);
-            bytes += s.size;
-          } catch {}
-        }
-      }
-    }
-  }
-
-  await visit(dir);
-  return { count, bytes };
-}
-
-async function main() {
-  let grandTotalCount = 0;
-  let grandTotalBytes = 0;
-  for (const dir of paths) {
-    const { count, bytes } = await countFiles(dir);
-    const mb = (bytes / (1024 * 1024)).toFixed(2);
-    console.log(`${dir}: ${count} files (${mb} MB)`);
-    grandTotalCount += count;
-    grandTotalBytes += bytes;
-  }
-  const grandTotalMB = (grandTotalBytes / (1024 * 1024)).toFixed(2);
-  console.log(`Grand Total: ${grandTotalCount} files (${grandTotalMB} MB)`);
-}
-
-main().catch(console.error);
diff --git a/scratch/count-status.ts b/scratch/count-status.ts
deleted file mode 100644
index e4265b86c..000000000
--- a/scratch/count-status.ts
+++ /dev/null
@@ -1,54 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  // Total in documents table
-  const { data: docCounts, error: docError } = await supabase
-    .from("documents")
-    .select("status, metadata");
-  
-  if (docError) {
-    console.error("Error fetching documents:", docError);
-    return;
-  }
-  
-  const statusCounts: Record<string, number> = {};
-  let enrichmentPending = 0;
-  
-  for (const doc of docCounts || []) {
-    statusCounts[doc.status] = (statusCounts[doc.status] || 0) + 1;
-    const meta = doc.metadata && typeof doc.metadata === "object" ? (doc.metadata as any) : {};
-    if (meta.enrichment_status === "pending") {
-      enrichmentPending++;
-    }
-  }
-  
-  console.log("=== Document Table Status ===");
-  console.log("Total rows in documents table:", docCounts?.length);
-  console.log("Breakdown by status:", statusCounts);
-  console.log("Documents with metadata.enrichment_status === 'pending':", enrichmentPending);
-  
-  // Total in ingestion_jobs
-  const { data: jobCounts, error: jobError } = await supabase
-    .from("ingestion_jobs")
-    .select("status");
-    
-  if (jobError) {
-    console.error("Error fetching jobs:", jobError);
-    return;
-  }
-  
-  const jobStatusCounts: Record<string, number> = {};
-  for (const job of jobCounts || []) {
-    jobStatusCounts[job.status] = (jobStatusCounts[job.status] || 0) + 1;
-  }
-  
-  console.log("=== Ingestion Jobs Status ===");
-  console.log("Total jobs:", jobCounts?.length);
-  console.log("Breakdown by status:", jobStatusCounts);
-}
-
-main().catch(console.error);
diff --git a/scratch/create-hnsw-indexes.sql b/scratch/create-hnsw-indexes.sql
deleted file mode 100644
index 0c478aa82..000000000
--- a/scratch/create-hnsw-indexes.sql
+++ /dev/null
@@ -1,21 +0,0 @@
--- SQL script to recreate the missing HNSW vector indexes.
--- Run this script in the Supabase Dashboard SQL Editor for your project:
--- https://supabase.com/dashboard/project/sjrfecxgysukkwxsowpy/sql/new
-
-SET statement_timeout = 0; -- disable statement timeout for index creation
-
--- 1. Create HNSW index on document_chunks (69,000+ rows)
--- This is a large build and must be run without session timeout.
-CREATE INDEX IF NOT EXISTS document_chunks_embedding_hnsw_idx
-  ON public.document_chunks USING hnsw (embedding vector_cosine_ops)
-  WITH (m = 24, ef_construction = 128);
-
--- 2. Create HNSW index on document_memory_cards (10,000+ rows)
-CREATE INDEX IF NOT EXISTS document_memory_cards_embedding_hnsw_idx
-  ON public.document_memory_cards USING hnsw (embedding vector_cosine_ops)
-  WITH (m = 24, ef_construction = 128);
-
--- 3. Create HNSW index on document_embedding_fields (3,000+ rows)
-CREATE INDEX IF NOT EXISTS document_embedding_fields_embedding_hnsw_idx
-  ON public.document_embedding_fields USING hnsw (embedding vector_cosine_ops)
-  WITH (m = 24, ef_construction = 128);
diff --git a/scratch/fix-indexes.ts b/scratch/fix-indexes.ts
deleted file mode 100644
index 348f296cb..000000000
--- a/scratch/fix-indexes.ts
+++ /dev/null
@@ -1,15 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  console.log("Applying missing schema indexes...");
-  
-  // 1. Create document_embedding_fields_owner_idx
-  const { error: error1 } = await supabase.rpc("search_schema_health"); // Just a test, but we can execute raw SQL if we have a way.
-  // Wait, does Supabase JS client support raw SQL execution? No, unless we use RPC or run it via schema migration.
-  // Ah, wait! Is there a function we can use, or does the migration schema script apply them?
-  // Wait, let's see how the test suite applies migrations or runs SQL.
-}
diff --git a/scratch/queue-queued-docs.ts b/scratch/queue-queued-docs.ts
deleted file mode 100644
index 3b389ca8c..000000000
--- a/scratch/queue-queued-docs.ts
+++ /dev/null
@@ -1,57 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("@/lib/supabase/admin");
-  const { env } = await import("@/lib/env");
-  const supabase = createAdminClient();
-
-  const { data: docs, error } = await supabase
-    .from("documents")
-    .select("id, metadata")
-    .eq("status", "queued");
-
-  if (error) {
-    console.error("Error fetching queued documents:", error);
-    return;
-  }
-
-  if (!docs || docs.length === 0) {
-    console.log("No queued documents found to create jobs for.");
-    return;
-  }
-
-  for (const doc of docs) {
-    const metadata = doc.metadata as Record<string, any>;
-    const batchId = metadata?.import_batch_id || null;
-
-    console.log(`Creating ingestion job for document ${doc.id} (batch: ${batchId})`);
-
-    const { data: existingJobs } = await supabase
-      .from("ingestion_jobs")
-      .select("id")
-      .eq("document_id", doc.id);
-
-    if (existingJobs && existingJobs.length > 0) {
-      console.log(`Job already exists for document ${doc.id}, skipping.`);
-      continue;
-    }
-
-    const { error: insertError } = await supabase.from("ingestion_jobs").insert({
-      document_id: doc.id,
-      batch_id: batchId,
-      status: "pending",
-      stage: "queued",
-      progress: 0,
-      max_attempts: env.WORKER_MAX_ATTEMPTS || 3,
-    });
-
-    if (insertError) {
-      console.error(`Error inserting job for document ${doc.id}:`, insertError);
-    } else {
-      console.log(`Successfully created ingestion job for document ${doc.id}`);
-    }
-  }
-}
-
-main().catch(console.error);
diff --git a/scratch/run-migrations-safe.sql b/scratch/run-migrations-safe.sql
deleted file mode 100644
index 1096201c4..000000000
--- a/scratch/run-migrations-safe.sql
+++ /dev/null
@@ -1,603 +0,0 @@
--- Set no timeout for the session
-SET statement_timeout = 0;
-SET search_path = public, extensions;
-
--- ============================================================
--- 1. HNSW INDEX TUNING
--- Rebuild with m=24 (graph connectivity) and ef_construction=128
--- (build-time recall). Default was m=16, ef_construction=64.
--- ============================================================
-
-drop index if exists document_chunks_embedding_hnsw_idx;
-create index document_chunks_embedding_hnsw_idx
-  on public.document_chunks using hnsw (embedding vector_cosine_ops)
-  with (m = 24, ef_construction = 128);
-
-drop index if exists document_memory_cards_embedding_hnsw_idx;
-create index document_memory_cards_embedding_hnsw_idx
-  on public.document_memory_cards using hnsw (embedding vector_cosine_ops)
-  with (m = 24, ef_construction = 128);
-
-drop index if exists document_index_units_embedding_hnsw_idx;
-create index document_index_units_embedding_hnsw_idx
-  on public.document_index_units using hnsw (embedding vector_cosine_ops)
-  with (m = 24, ef_construction = 128);
-
-drop index if exists document_embedding_fields_embedding_hnsw_idx;
-create index document_embedding_fields_embedding_hnsw_idx
-  on public.document_embedding_fields using hnsw (embedding vector_cosine_ops)
-  with (m = 24, ef_construction = 128);
-
--- ============================================================
--- 2. RAG RETRIEVAL LOGS TABLE
--- Per-query diagnostics: candidate scores, latency breakdown,
--- miss detection.
--- ============================================================
-
-create table if not exists public.rag_retrieval_logs (
-  id uuid primary key default gen_random_uuid(),
-  owner_id uuid references auth.users(id) on delete set null,
-  query text not null,
-  normalized_query text,
-  query_class text,
-  retrieval_strategy text,
-
-  -- Candidate scores
-  candidate_count integer not null default 0,
-  top_similarity double precision,
-  top_text_rank double precision,
-  top_hybrid_score double precision,
-  top_rrf_score double precision,
-  mean_hybrid_score double precision,
-
-  -- Selected citations
-  selected_chunk_ids uuid[] not null default '{}',
-  selected_document_ids uuid[] not null default '{}',
-  selected_count integer not null default 0,
-
-  -- Latency breakdown
-  embedding_latency_ms integer,
-  rpc_latency_ms integer,
-  rerank_latency_ms integer,
-  total_latency_ms integer,
-
-  -- Source counts
-  vector_candidate_count integer,
-  text_candidate_count integer,
-  memory_card_count integer,
-  index_unit_count integer,
-  embedding_field_count integer,
-
-  -- Miss detection
-  is_miss boolean not null default false,
-  miss_reason text,
-
-  -- Metadata
-  embedding_cache_hit boolean,
-  metadata jsonb not null default '{}'::jsonb,
-  created_at timestamptz not null default now()
-);
-
-create index if not exists rag_retrieval_logs_owner_created_idx
-  on public.rag_retrieval_logs(owner_id, created_at desc);
-create index if not exists rag_retrieval_logs_miss_idx
-  on public.rag_retrieval_logs(is_miss, created_at desc)
-  where is_miss = true;
-create index if not exists rag_retrieval_logs_strategy_idx
-  on public.rag_retrieval_logs(retrieval_strategy, created_at desc);
-
-alter table public.rag_retrieval_logs enable row level security;
-grant select, insert, update, delete on table public.rag_retrieval_logs to service_role;
-grant select on table public.rag_retrieval_logs to authenticated;
-
-create policy "rag retrieval logs owner read" on public.rag_retrieval_logs
-  for select to authenticated using (owner_id = (select auth.uid()));
-
--- ============================================================
--- 3. CONTENT_HASH DEDUP ON DOCUMENT_EMBEDDING_FIELDS
--- Replace the problematic unique(document_id, source_chunk_id,
--- field_type, content) constraint that indexes full text.
--- ============================================================
-
-alter table public.document_embedding_fields
-  add column if not exists content_hash text;
-
--- Populate for existing rows
-update public.document_embedding_fields
-  set content_hash = md5(content)
-  where content_hash is null;
-
--- Drop the old constraint.
-alter table public.document_embedding_fields
-  drop constraint if exists document_embedding_fields_document_id_source_chunk_id_field_key;
-alter table public.document_embedding_fields
-  drop constraint if exists document_embedding_fields_document_id_source_chunk_id_fiel_key;
-
--- Create hash-based unique index
-create unique index if not exists document_embedding_fields_dedup_idx
-  on public.document_embedding_fields(document_id, source_chunk_id, field_type, content_hash);
-
--- ============================================================
--- 4. DROP DUPLICATE/REDUNDANT INDEXES
--- These single-column indexes are covered by composite indexes
--- with the same leading column.
--- ============================================================
-
-drop index if exists document_chunks_document_id_idx;
-drop index if exists document_sections_document_id_idx;
-drop index if exists document_memory_cards_document_id_idx;
-drop index if exists document_images_document_id_idx;
-drop index if exists document_labels_document_id_idx;
-drop index if exists document_embedding_fields_document_id_idx;
-drop index if exists document_table_facts_document_id_idx;
-drop index if exists document_index_quality_document_id_idx;
-
--- ============================================================
--- 5. OTHER EXPECTED SCHEMA INDEXES FROM OTHER MIGRATIONS
--- ============================================================
-create index if not exists ingestion_jobs_status_next_run_idx
-  on public.ingestion_jobs(status, next_run_at asc)
-  where status in ('pending', 'processing');
-
-create index if not exists ingestion_jobs_document_status_idx
-  on public.ingestion_jobs(document_id, status);
-
-create index if not exists import_batches_status_created_idx
-  on public.import_batches(status, created_at desc);
-
-create index if not exists storage_cleanup_jobs_status_created_idx
-  on public.storage_cleanup_jobs(status, created_at desc);
-
--- ============================================================
--- 6. ENHANCED HYBRID SCORING
--- ============================================================
-
-create or replace function public.match_document_chunks_hybrid(
-  query_embedding extensions.vector(1536),
-  query_text text,
-  match_count integer default 12,
-  min_similarity double precision default 0.12,
-  document_filters uuid[] default null,
-  owner_filter uuid default null
-)
-returns table (
-  id uuid,
-  document_id uuid,
-  title text,
-  file_name text,
-  page_number integer,
-  chunk_index integer,
-  section_heading text,
-  content text,
-  retrieval_synopsis text,
-  image_ids uuid[],
-  source_metadata jsonb,
-  similarity double precision,
-  text_rank double precision,
-  hybrid_score double precision,
-  rrf_score double precision,
-  images jsonb
-)
-language sql
-stable
-set search_path = public, extensions, pg_temp
-as $$
-  with query as (
-    select websearch_to_tsquery('english', coalesce(query_text, '')) as tsq
-  ),
-  vector_ranked as (
-    select
-      c.id,
-      c.document_id,
-      c.page_number,
-      c.chunk_index,
-      c.section_heading,
-      c.content,
-      c.retrieval_synopsis,
-      c.image_ids,
-      1 - (c.embedding <=> query_embedding) as similarity,
-      (
-        ts_rank_cd(c.search_tsv, query.tsq) +
-        (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-      )::double precision as text_rank,
-      row_number() over (order by c.embedding <=> query_embedding) as vector_rank,
-      null::bigint as text_match_rank,
-      coalesce((d.metadata->'rag_indexing_version') is not null, false) as has_deep_index,
-      d.updated_at as doc_updated_at,
-      coalesce(
-        (select q.quality_score from public.document_index_quality q where q.document_id = c.document_id),
-        0.7
-      ) as quality_score
-    from public.document_chunks c
-    join public.documents d on d.id = c.document_id
-    cross join query
-    where (document_filters is null or c.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and d.status = 'indexed'
-      and 1 - (c.embedding <=> query_embedding) >= min_similarity
-    order by c.embedding <=> query_embedding
-    limit greatest(match_count * 6, 48)
-  ),
-  text_ranked as (
-    select
-      c.id,
-      c.document_id,
-      c.page_number,
-      c.chunk_index,
-      c.section_heading,
-      c.content,
-      c.retrieval_synopsis,
-      c.image_ids,
-      1 - (c.embedding <=> query_embedding) as similarity,
-      (
-        ts_rank_cd(c.search_tsv, query.tsq) +
-        (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-      )::double precision as text_rank,
-      null::bigint as vector_rank,
-      row_number() over (
-        order by
-          (
-            ts_rank_cd(c.search_tsv, query.tsq) +
-            (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-          ) desc,
-          c.embedding <=> query_embedding
-      ) as text_match_rank,
-      coalesce((d.metadata->'rag_indexing_version') is not null, false) as has_deep_index,
-      d.updated_at as doc_updated_at,
-      coalesce(
-        (select q.quality_score from public.document_index_quality q where q.document_id = c.document_id),
-        0.7
-      ) as quality_score
-    from public.document_chunks c
-    join public.documents d on d.id = c.document_id
-    cross join query
-    where (document_filters is null or c.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and d.status = 'indexed'
-      and (c.search_tsv @@ query.tsq or d.title_search_tsv @@ query.tsq)
-    order by (
-      ts_rank_cd(c.search_tsv, query.tsq) +
-      (ts_rank_cd(d.title_search_tsv, query.tsq) * 3.0)
-    ) desc
-    limit greatest(match_count * 6, 48)
-  ),
-  combined as (
-    select * from vector_ranked
-    union all
-    select * from text_ranked
-  ),
-  scored as (
-    select
-      id,
-      document_id,
-      page_number,
-      chunk_index,
-      section_heading,
-      content,
-      retrieval_synopsis,
-      image_ids,
-      max(similarity)::double precision as similarity,
-      max(text_rank)::double precision as text_rank,
-      min(vector_rank) as vector_rank,
-      min(text_match_rank) as text_match_rank,
-      max(quality_score)::double precision as quality_score,
-      bool_or(has_deep_index) as has_deep_index,
-      max(doc_updated_at) as doc_updated_at
-    from combined
-    group by id, document_id, page_number, chunk_index, section_heading, content, retrieval_synopsis, image_ids
-  ),
-  scored_metrics as (
-    select
-      scored.*,
-      (
-        (scored.similarity * 0.62)
-        + (least(scored.text_rank, 1) * 0.22)
-        + (scored.quality_score * 0.10)
-        + (case when scored.doc_updated_at > now() - interval '90 days' then 0.06 else 0 end)
-      )::double precision as hybrid_score,
-      (
-        coalesce(1.0 / (60 + scored.vector_rank), 0) +
-        coalesce(1.0 / (60 + scored.text_match_rank), 0)
-      )::double precision as rrf_score
-    from scored
-  ),
-  hybrid_candidates as (
-    select id
-    from scored_metrics
-    order by hybrid_score desc, similarity desc, text_rank desc
-    limit match_count
-  ),
-  vector_candidates as (
-    select id
-    from scored_metrics
-    order by similarity desc, hybrid_score desc
-    limit match_count
-  ),
-  text_candidates as (
-    select id
-    from scored_metrics
-    order by text_rank desc, hybrid_score desc
-    limit match_count
-  ),
-  rrf_candidates as (
-    select id
-    from scored_metrics
-    order by rrf_score desc, hybrid_score desc
-    limit match_count
-  ),
-  candidate_ids as (
-    select id from hybrid_candidates
-    union
-    select id from vector_candidates
-    union
-    select id from text_candidates
-    union
-    select id from rrf_candidates
-  )
-  select
-    c.id,
-    c.document_id,
-    d.title,
-    d.file_name,
-    c.page_number,
-    c.chunk_index,
-    c.section_heading,
-    c.content,
-    c.retrieval_synopsis,
-    c.image_ids,
-    d.metadata as source_metadata,
-    c.similarity,
-    c.text_rank,
-    c.hybrid_score,
-    c.rrf_score,
-    public.chunk_image_metadata(c.image_ids) as images
-  from scored_metrics c
-  join candidate_ids candidates on candidates.id = c.id
-  join public.documents d on d.id = c.document_id
-  order by c.hybrid_score desc, c.rrf_score desc, c.similarity desc, c.text_rank desc
-  limit match_count;
-$$;
-
-create or replace function public.match_document_index_units_hybrid(
-  query_embedding extensions.vector(1536),
-  query_text text,
-  match_count integer default 24,
-  min_similarity double precision default 0.1,
-  document_filters uuid[] default null,
-  owner_filter uuid default null
-)
-returns table (
-  id uuid,
-  document_id uuid,
-  source_chunk_id uuid,
-  source_image_id uuid,
-  unit_type text,
-  title text,
-  content text,
-  page_start integer,
-  page_end integer,
-  heading_path text[],
-  normalized_terms text[],
-  source_span jsonb,
-  quality_score real,
-  extraction_mode text,
-  similarity double precision,
-  text_rank double precision,
-  hybrid_score double precision,
-  metadata jsonb
-)
-language sql
-stable
-set search_path = public, extensions, pg_temp
-as $$
-  with query as (
-    select websearch_to_tsquery('english', coalesce(query_text, '')) as tsq,
-      regexp_split_to_array(lower(coalesce(query_text, '')), '\s+') as terms
-  ),
-  ranked as (
-    select u.id, u.document_id, u.source_chunk_id, u.source_image_id, u.unit_type, u.title, u.content, u.page_start,
-      u.page_end, u.heading_path, u.normalized_terms, u.source_span, u.quality_score, u.extraction_mode,
-      (1 - (u.embedding <=> query_embedding))::double precision as similarity,
-      (ts_rank_cd(u.search_tsv, query.tsq)
-        + case when u.normalized_terms && query.terms then 0.25 else 0 end
-        + case when u.unit_type in ('askable_question', 'table_fact', 'clinical_fact') then 0.06
-               when u.unit_type = 'section_summary' then 0.03
-               else 0 end
-      )::double precision as text_rank,
-      u.metadata
-    from public.document_index_units u
-    join public.documents d on d.id = u.document_id
-    cross join query
-    where d.status = 'indexed'
-      and (document_filters is null or u.document_id = any(document_filters))
-      and (owner_filter is null or d.owner_id = owner_filter)
-      and u.source_chunk_id is not null
-      and (u.search_tsv @@ query.tsq or u.normalized_terms && query.terms)
-    order by text_rank desc, similarity desc
-    limit greatest(match_count * 3, 48)
-  )
-  select id, document_id, source_chunk_id, source_image_id, unit_type, title, content, page_start, page_end, heading_path,
-    normalized_terms, source_span, quality_score, extraction_mode, similarity, text_rank,
-    (
-      (similarity * 0.52)
-      + (least(text_rank, 1) * 0.28)
-      + (quality_score * 0.12)
-      + (case when extraction_mode in ('model_heavy', 'hybrid') then 0.04 else 0 end)
-      + (case when unit_type = 'askable_question' then 0.04 else 0 end)
-    )::double precision as hybrid_score,
-    metadata
-  from ranked
-  order by hybrid_score desc, similarity desc, text_rank desc
-  limit match_count;
-$$;
-
-create or replace function public.analyze_rag_tables()
-returns void
-language plpgsql
-set search_path = public, extensions, pg_temp
-as $$
-begin
-  analyze public.document_chunks;
-  analyze public.document_memory_cards;
-  analyze public.document_index_units;
-  analyze public.document_embedding_fields;
-  analyze public.document_table_facts;
-  analyze public.documents;
-end;
-$$;
-
-revoke execute on function public.analyze_rag_tables() from public, anon, authenticated;
-grant execute on function public.analyze_rag_tables() to service_role;
-
-create or replace function public.reset_document_index(p_document_id uuid)
-returns void
-language plpgsql
-set search_path = public, extensions, pg_temp
-as $$
-begin
-  perform set_config('statement_timeout', '180000', true);
-  delete from public.document_index_units where document_id = p_document_id;
-  delete from public.document_memory_cards where document_id = p_document_id;
-  delete from public.document_sections where document_id = p_document_id;
-  delete from public.document_table_facts where document_id = p_document_id;
-  delete from public.document_embedding_fields where document_id = p_document_id;
-  delete from public.document_index_quality where document_id = p_document_id;
-  delete from public.document_chunks where document_id = p_document_id;
-  delete from public.document_images where document_id = p_document_id;
-  delete from public.document_pages where document_id = p_document_id;
-end;
-$$;
-
--- Update health check with correct indexes
-create or replace function public.search_schema_health()
-returns jsonb
-language plpgsql
-stable
-set search_path = public, extensions, pg_catalog, pg_temp
-as $$
-declare
-  missing text[] := array[]::text[];
-  vector_type_oid oid;
-  vector_schema text;
-  index_name text;
-begin
-  select t.oid, n.nspname
-  into vector_type_oid, vector_schema
-  from pg_type t
-  join pg_namespace n on n.oid = t.typnamespace
-  where t.typname = 'vector'
-    and n.nspname = 'extensions'
-  limit 1;
-
-  if vector_type_oid is null then
-    missing := array_append(missing, 'extensions.vector_type');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_memory_cards_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_memory_cards_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_index_units_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_index_units_hybrid.extensions_vector_signature');
-  end if;
-
-  if to_regclass('public.document_index_units') is null then
-    missing := array_append(missing, 'document_index_units.table');
-  end if;
-  if to_regclass('public.rag_retrieval_logs') is null then
-    missing := array_append(missing, 'rag_retrieval_logs.table');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'documents_title_trgm_idx') then
-    missing := array_append(missing, 'documents_title_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_chunks_content_trgm_idx') then
-    missing := array_append(missing, 'document_chunks_content_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_labels_label_trgm_idx') then
-    missing := array_append(missing, 'document_labels_label_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_summaries_summary_trgm_idx') then
-    missing := array_append(missing, 'document_summaries_summary_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_index_units_embedding_hnsw_idx') then
-    missing := array_append(missing, 'document_index_units_embedding_hnsw_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_embedding_fields_owner_idx', 'document_embedding_fields_owner_id_idx')) then
-    missing := array_append(missing, 'document_embedding_fields_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_table_facts_owner_idx', 'document_table_facts_owner_id_idx')) then
-    missing := array_append(missing, 'document_table_facts_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_table_facts_source_image_idx') then
-    missing := array_append(missing, 'document_table_facts_source_image_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_embedding_fields_dedup_idx') then
-    missing := array_append(missing, 'document_embedding_fields_dedup_idx');
-  end if;
-  foreach index_name in array array[
-    'document_pages_document_idx',
-    'document_images_document_idx',
-    'document_sections_document_idx',
-    'document_memory_cards_document_idx',
-    'document_chunks_document_idx',
-    'document_table_facts_document_idx',
-    'document_embedding_fields_document_idx',
-    'document_index_units_document_idx',
-    'ingestion_jobs_status_next_run_idx',
-    'ingestion_jobs_document_status_idx',
-    'documents_owner_status_idx',
-    'import_batches_status_created_idx',
-    'storage_cleanup_jobs_status_created_idx'
-  ] loop
-    if not exists (select 1 from pg_class where relname = index_name) then
-      missing := array_append(missing, index_name);
-    end if;
-  end loop;
-
-  return jsonb_build_object(
-    'ok', cardinality(missing) = 0,
-    'missing', missing,
-    'vector_extension_schema', vector_schema,
-    'checked_at', now()
-  );
-end;
-$$;
-
-revoke execute on function public.search_schema_health() from public, anon, authenticated;
-grant execute on function public.search_schema_health() to service_role;
diff --git a/scratch/show-samples.ts b/scratch/show-samples.ts
deleted file mode 100644
index 830a9bff5..000000000
--- a/scratch/show-samples.ts
+++ /dev/null
@@ -1,22 +0,0 @@
-import { loadEnvConfig } from "@next/env";
-loadEnvConfig(process.cwd());
-
-async function main() {
-  const { createAdminClient } = await import("../src/lib/supabase/admin");
-  const supabase = createAdminClient();
-  
-  const { data: sampleDocs, error } = await supabase
-    .from("documents")
-    .select("id, title, status, metadata")
-    .limit(10);
-    
-  if (error) {
-    console.error(error);
-    return;
-  }
-  
-  console.log("=== Sample Documents ===");
-  console.log(JSON.stringify(sampleDocs, null, 2));
-}
-
-main().catch(console.error);
diff --git a/scratch/update-health-check.sql b/scratch/update-health-check.sql
deleted file mode 100644
index 6da2830c0..000000000
--- a/scratch/update-health-check.sql
+++ /dev/null
@@ -1,132 +0,0 @@
-create or replace function public.search_schema_health()
-returns jsonb
-language plpgsql
-stable
-set search_path = public, extensions, pg_catalog, pg_temp
-as $$
-declare
-  missing text[] := array[]::text[];
-  vector_type_oid oid;
-  vector_schema text;
-  index_name text;
-begin
-  select t.oid, n.nspname
-  into vector_type_oid, vector_schema
-  from pg_type t
-  join pg_namespace n on n.oid = t.typnamespace
-  where t.typname = 'vector'
-    and n.nspname = 'extensions'
-  limit 1;
-
-  if vector_type_oid is null then
-    missing := array_append(missing, 'extensions.vector_type');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_chunks_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_chunks_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_memory_cards_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_memory_cards_hybrid.extensions_vector_signature');
-  end if;
-
-  if vector_type_oid is not null and not exists (
-    select 1
-    from pg_proc p
-    join pg_namespace n on n.oid = p.pronamespace
-    where n.nspname = 'public'
-      and p.proname = 'match_document_index_units_hybrid'
-      and p.proargtypes[0] = vector_type_oid
-  ) then
-    missing := array_append(missing, 'match_document_index_units_hybrid.extensions_vector_signature');
-  end if;
-
-  if to_regclass('public.document_index_units') is null then
-    missing := array_append(missing, 'document_index_units.table');
-  end if;
-  if to_regclass('public.rag_retrieval_logs') is null then
-    missing := array_append(missing, 'rag_retrieval_logs.table');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'documents_title_trgm_idx') then
-    missing := array_append(missing, 'documents_title_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_chunks_content_trgm_idx') then
-    missing := array_append(missing, 'document_chunks_content_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_labels_label_trgm_idx') then
-    missing := array_append(missing, 'document_labels_label_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_summaries_summary_trgm_idx') then
-    missing := array_append(missing, 'document_summaries_summary_trgm_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_index_units_embedding_hnsw_idx') then
-    missing := array_append(missing, 'document_index_units_embedding_hnsw_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_embedding_fields_owner_idx', 'document_embedding_fields_owner_id_idx')) then
-    missing := array_append(missing, 'document_embedding_fields_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname in ('document_table_facts_owner_idx', 'document_table_facts_owner_id_idx')) then
-    missing := array_append(missing, 'document_table_facts_owner_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_table_facts_source_image_idx') then
-    missing := array_append(missing, 'document_table_facts_source_image_idx');
-  end if;
-  if not exists (select 1 from pg_class where relname = 'document_embedding_fields_dedup_idx') then
-    missing := array_append(missing, 'document_embedding_fields_dedup_idx');
-  end if;
-  foreach index_name in array array[
-    'document_pages_document_idx',
-    'document_images_document_idx',
-    'document_sections_document_idx',
-    'document_memory_cards_document_idx',
-    'document_chunks_document_idx',
-    'document_table_facts_document_idx',
-    'document_embedding_fields_document_idx',
-    'document_index_units_document_idx',
-    'ingestion_jobs_status_next_run_idx',
-    'ingestion_jobs_document_status_idx',
-    'documents_owner_status_idx',
-    'import_batches_status_created_idx',
-    'storage_cleanup_jobs_status_created_idx'
-  ] loop
-    if not exists (select 1 from pg_class where relname = index_name) then
-      missing := array_append(missing, index_name);
-    end if;
-  end loop;
-
-  return jsonb_build_object(
-    'ok', cardinality(missing) = 0,
-    'missing', missing,
-    'vector_extension_schema', vector_schema,
-    'checked_at', now()
-  );
-end;
-$$;
-
-revoke execute on function public.search_schema_health() from public, anon, authenticated;
-grant execute on function public.search_schema_health() to service_role;