implemented a lot of features

commands/ReleaseStats.ts

@@ -0,0 +1,57 @@
+import { BaseCommand } from '@adonisjs/core/ace'
+import { CommandOptions } from '@adonisjs/core/types/ace'
+import db from '@adonisjs/lucid/services/db'
+
+/** Prints how well files are being grouped into releases. */
+export default class ReleaseStats extends BaseCommand {
+  public static commandName = 'release:stats'
+  public static description = 'Shows release recognition statistics.'
+
+  public static options: CommandOptions = {
+    startApp: true,
+  }
+
+  public async run() {
+    const { rows } = await db.rawQuery(`
+      WITH r AS (
+        SELECT r.id, r.file_total, r.poster IS NULL AS weak, r.source,
+               count(b.id) AS binaries,
+               count(b.id) FILTER (WHERE b.completed_at IS NOT NULL) AS completed
+        FROM releases r LEFT JOIN binaries b ON b.release_id = r.id
+        GROUP BY r.id
+      )
+      SELECT
+        count(*) AS releases,
+        count(*) FILTER (WHERE completed > 0) AS with_files,
+        count(*) FILTER (WHERE completed >= 2) AS multi_file,
+        count(*) FILTER (WHERE file_total IS NOT NULL AND completed >= file_total) AS complete_by_counter,
+        count(*) FILTER (WHERE weak) AS from_obfuscated,
+        count(*) FILTER (WHERE weak AND completed > 0) AS from_obfuscated_with_files
+      FROM r`)
+    const { rows: coverage } = await db.rawQuery(`
+      SELECT count(*) AS completed_files,
+             count(*) FILTER (WHERE release_id IS NOT NULL) AS in_release
+      FROM binaries WHERE completed_at IS NOT NULL`)
+    const { rows: bySource } = await db.rawQuery(`
+      SELECT coalesce(r.source, 'unknown') AS source, count(DISTINCT r.id) AS releases,
+             count(b.id) FILTER (WHERE b.completed_at IS NOT NULL) AS completed_files
+      FROM releases r JOIN binaries b ON b.release_id = r.id GROUP BY 1 ORDER BY 2 DESC`)
+    // Readable names: descriptive as posted, or recovered from contents.
+    const { rows: naming } = await db.rawQuery(`
+      WITH f AS (
+        SELECT real_name IS NOT NULL AS recovered,
+               is_readable_name(coalesce(real_name, filename)) AS readable_name,
+               regexp_replace(coalesce(real_name, filename),
+                 '(\\.(part\\d+\\.rar|r\\d{2,3}|rar|vol\\d+[+-]\\d+\\.par2|par2|7z(\\.\\d{3})?|zip|\\d{3}|nfo|sfv|nzb|tar|zst|mkv|mp4|avi|flac|mp3|jpg))+$', '', 'i') AS stem,
+               coalesce(real_name, filename) AS name
+        FROM files
+      )
+      SELECT count(*) AS files,
+             count(*) FILTER (WHERE recovered) AS names_recovered,
+             count(*) FILTER (WHERE name ~ '\\.[A-Za-z0-9]{1,5}$' AND readable_name) AS readable
+      FROM f`)
+    this.logger.info(JSON.stringify({ ...rows[0], ...coverage[0] }))
+    this.logger.info(`naming: ${JSON.stringify(naming[0])}`)
+    this.logger.info(`by source: ${JSON.stringify(bySource)}`)
+  }
+}