Partition files by age (id)

New rows went into free space scattered through the 9.6 GB heap on the HDD
(left behind by inspection's updates): one random HDD read per insert, and
seed batches took a median 13 s to store. The existing table becomes the
first partition as it is (a CHECK scan, no rewrite; indexes, its foreign
key and the triggers are taken over by the partitioned parent); new
partitions of 4M ids go in the same tablespace and stay in memory while
they fill.

- Every partition has a unique file_key index; FileStore.insert looks keys
  up in all partitions first (a key can't be unique across partitions).
- FilePartitions.ensure() (gc:scheduler) keeps two partitions ahead.
- Stats count files from the partitions; seed:node waits ~20 minutes for
  the API before giving a range up.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

app/services/FileStore.ts

@@ -39,15 +39,34 @@ export function loadedSegments(stored: Buffer | null): Segment[] {
 class FileStore {
   /**
    * Inserts files, skipping those whose `file_key` exists. Returns the rows
-   * that were inserted (id and filename).
+   * that were inserted (id and filename). `files` is partitioned by id
+   * (CONCEPT.md D39) and each partition has its own unique file_key index:
+   * keys already in any partition are looked up first; two writers adding
+   * the same new key at once meet in the newest partition's index.
    */
   async insert(files: NewFile[], client: Pick<QueryClientContract, 'knexQuery'> = db): Promise<{ id: number; filename: string }[]> {
     if (files.length === 0) return []
-    const rows = files.map(({ segments, ...file }) => {
-      const sorted = [...segments].sort((a, b) => a.number - b.number)
-      return { ...file, segments: packSegments(sorted), first_message_id: sorted[0]?.id ?? null }
-    })
-    const inserted = await client.knexQuery().table('files').insert(rows).onConflict('file_key').ignore().returning(['id', 'filename'])
+    const keys = [...new Set(files.map((file) => file.file_key).filter((key): key is string => typeof key === 'string'))]
+    const known = new Set<string>()
+    for (let i = 0; i < keys.length; i += 5000) {
+      const found = await client.knexQuery().from('files').whereIn('file_key', keys.slice(i, i + 5000)).select('file_key')
+      for (const row of found) known.add(row.file_key)
+    }
+    const seen = new Set<string>()
+    const rows = files
+      .filter((file) => {
+        const key = file.file_key as string | null | undefined
+        if (!key) return true
+        if (known.has(key) || seen.has(key)) return false
+        seen.add(key)
+        return true
+      })
+      .map(({ segments, ...file }) => {
+        const sorted = [...segments].sort((a, b) => a.number - b.number)
+        return { ...file, segments: packSegments(sorted), first_message_id: sorted[0]?.id ?? null }
+      })
+    if (rows.length === 0) return []
+    const inserted = await client.knexQuery().table('files').insert(rows).onConflict().ignore().returning(['id', 'filename'])
     return inserted.map((row: any) => ({ id: Number(row.id), filename: row.filename }))
   }
 }