Daan MeijerandClaude Opus 5.5
f0ff8e200f
Copying all existing keys in the migration read the files rows from the HDD for over half an hour and held up the deploy. gc:scheduler now copies them in slices (FILE_KEYS_FILL_SECONDS per sweep) and FileStore checks every partition until the fill is complete. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
102 lines
4.3 KiB
TypeScript
102 lines
4.3 KiB
TypeScript
import db from '@adonisjs/lucid/services/db'
|
|
import FileKeys from '#services/FileKeys'
|
|
import type { QueryClientContract } from '@adonisjs/lucid/types/database'
|
|
import { packSegments, unpackSegments, type Segment } from '#lib/nzb/segments'
|
|
|
|
/**
|
|
* The only code that knows how `files` keeps its segment lists (CONCEPT.md
|
|
* D31): compressed in `files.segments`, with the first segment's message-id
|
|
* in the indexed `first_message_id` for lookups. Read them with
|
|
* `segmentsOf(file)`; write new files with `FileStore.insert`.
|
|
*/
|
|
|
|
/** A new `files` row, with its segments as a list. */
|
|
export type NewFile = Record<string, unknown> & { segments: Segment[] }
|
|
|
|
const unpacked = new WeakMap<object, Segment[]>()
|
|
|
|
/** The segments of a `files` row (a row from `db.from('files')`), in order. */
|
|
export function segmentsOf(file: { segments?: Buffer | Segment[] | null }): Segment[] {
|
|
const { segments } = file
|
|
if (!segments) return []
|
|
if (Array.isArray(segments)) return segments
|
|
let list = unpacked.get(file)
|
|
if (!list) {
|
|
list = unpackSegments(segments)
|
|
unpacked.set(file, list)
|
|
}
|
|
return list
|
|
}
|
|
|
|
/** Stored form of a segment list, for the File model's column. */
|
|
export function storedSegments(segments: Segment[]) {
|
|
return packSegments(segments)
|
|
}
|
|
|
|
export function loadedSegments(stored: Buffer | null): Segment[] {
|
|
return stored ? unpackSegments(stored) : []
|
|
}
|
|
|
|
/** Rows per INSERT, well below Postgres' 65,535 bind parameters (tests lower it). */
|
|
export const insertLimits = { rows: 2000 }
|
|
|
|
class FileStore {
|
|
/**
|
|
* Inserts files, skipping those whose `file_key` exists. Returns the rows
|
|
* that were inserted (id, filename, file_key). `files` is partitioned by id
|
|
* (CONCEPT.md D39) and each partition has its own unique file_key index:
|
|
* once file_keys (a 64-bit hash of every key) is complete, it says which keys
|
|
* may exist, and only those are looked up in the partitions; two
|
|
* writers adding the same new key at once meet in the newest partition's index.
|
|
*/
|
|
async insert(
|
|
files: NewFile[],
|
|
client: Pick<QueryClientContract, 'knexQuery' | 'rawQuery'> = db
|
|
): Promise<{ id: number; filename: string; file_key: string | null }[]> {
|
|
if (files.length === 0) return []
|
|
const keys = [...new Set(files.map((file) => file.file_key).filter((key): key is string => typeof key === 'string'))]
|
|
const known = new Set<string>()
|
|
const filtered = await FileKeys.complete()
|
|
for (let i = 0; i < keys.length; i += 5000) {
|
|
const chunk = keys.slice(i, i + 5000)
|
|
if (!filtered) {
|
|
const found = await client.knexQuery().from('files').whereIn('file_key', chunk).select('file_key')
|
|
for (const row of found) known.add(row.file_key)
|
|
continue
|
|
}
|
|
const { rows } = await client.rawQuery(
|
|
`SELECT k.key FROM unnest(?::text[]) AS k(key)
|
|
WHERE EXISTS (SELECT 1 FROM file_keys f WHERE f.key_hash = hashtextextended(k.key, 0))`,
|
|
[chunk]
|
|
)
|
|
const maybe = rows.map((row: { key: string }) => row.key)
|
|
if (maybe.length === 0) continue
|
|
const found = await client.knexQuery().from('files').whereIn('file_key', maybe).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 []
|
|
// A statement takes at most 65,535 parameters: ~5,400 rows of 12 columns.
|
|
const inserted: { id: number; filename: string; file_key: string | null }[] = []
|
|
for (let i = 0; i < rows.length; i += insertLimits.rows) {
|
|
const chunk = await client.knexQuery().table('files').insert(rows.slice(i, i + insertLimits.rows)).onConflict().ignore().returning(['id', 'filename', 'file_key'])
|
|
for (const row of chunk) inserted.push({ id: Number(row.id), filename: row.filename, file_key: row.file_key ?? null })
|
|
}
|
|
return inserted
|
|
}
|
|
}
|
|
|
|
export default new FileStore()
|