Drop garbage uploads early (D42)

Random-named uploads that never get a real name or password were ~90% of
what seeding stores. Two filters:

- before storing (seeding, ingest): random-named uploads of a dropped
  signature (posting tool from the Message-ID, first group) or of a
  release dropped after its probe aren't stored; 2% of a dropped
  signature's releases are still kept and probed. Stored files of dropped
  signatures aren't probed.
- after a probe: a random-named release that is still unnamed, without
  password, names from files or spot, with all its probes done, and
  encrypted, a container or missing its first article, is deleted and
  tombstoned in dropped_releases.

Signatures are learned by gc:scheduler from a sample of probed releases
(signature_probes); garbage:learn --seed fills it from existing files.

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

config/indexer.ts

@@ -95,6 +95,25 @@ export default {
     maxHddGb: env.get('BACKFILL_MAX_HDD_GB', 150),
   },
 
+  /**
+   * Dropping garbage uploads early (CONCEPT.md D42). A signature (posting tool, first
+   * group) is dropped once at least minProbes of its probed random-named releases are
+   * known and at most maxNamedRate of them got a real name or password; its
+   * random-named files are then not stored, except explorePercent of its releases.
+   */
+  garbage: {
+    /** Off: nothing is filtered or deleted (the learning still runs). */
+    enabled: env.get('GARBAGE_FILTER', true),
+    explorePercent: env.get('GARBAGE_EXPLORE_PERCENT', 2),
+    minProbes: 100,
+    maxNamedRate: 0.005,
+    /** Probed releases learned from: 1 in this many (by release id), for the last windowDays. */
+    probeSample: 4,
+    windowDays: 30,
+    /** How often gc:scheduler learns the signatures again. */
+    learnIntervalMs: env.get('GARBAGE_LEARN_INTERVAL_MS', 3 * 60 * 60 * 1000),
+  },
+
   /** worker:deferred leaves a file alone until it has had no new articles for this long. */
   deferredSettleMs: env.get('DEFERRED_SETTLE_MS', 15 * 60 * 1000),
   /** Files of at most this many parts are downloaded completely (small archives, NZBs, par2 index files). */