{
 "flight": {
  "id": 588726,
  "agent": "openai-codex",
  "state": "open",
  "title": "Fix inference OOM by streaming dataset",
  "number": 1489,
  "source": "aidev",
  "commits": 4,
  "debrief": null,
  "outcome": "stale",
  "signals": [
   "aidev:OpenAI_Codex"
  ],
  "index_ok": false,
  "additions": 155,
  "closed_at": null,
  "deletions": 15,
  "merged_at": null,
  "opened_at": "2025-07-27T18:17:29+00:00",
  "task_type": "fix",
  "body_chars": 287,
  "confidence": "dataset",
  "updated_at": "2026-11-03T17:53:33.703657+00:00",
  "enriched_at": null,
  "body_excerpt": "## Summary\n- avoid loading entire dataset in memory when running classification inference\n- stream input shards in batches and write results incrementally\n\n## Testing\n- `pytest tests/processing/tokenize -q`\n\n\n------\nhttps://chatgpt.com/codex/tasks/task_e_68866679275483318b55e211a2a91eee",
  "review_count": 2,
  "files_changed": 4,
  "hours_to_merge": null,
  "metrics_source": "aidev_commit_sum",
  "change_requests": 0,
  "og_generated_at": null,
  "reviewer_excerpt": "looks good, just need to add back the support for parquet. thanks!",
  "review_comment_count": 1,
  "reviewer_excerpt_url": "https://github.com/marin-community/marin/pull/1489#pullrequestreview-3059735779"
 },
 "repo": {
  "name": "marin",
  "owner": "marin-community",
  "stars": 4080,
  "license": "Apache-2.0",
  "full_name": "marin-community/marin",
  "primary_language": "Python"
 },
 "author": {
  "login": "dlwh",
  "name": "David Hall",
  "github_profile": "https://github.com/dlwh",
  "page": "https://agentriders.com/logs/author/dlwh/"
 },
 "claim": null,
 "events": [
  {
   "at": null,
   "url": null,
   "kind": "committed",
   "actor_login": "David Hall"
  },
  {
   "at": null,
   "url": null,
   "kind": "reviewed",
   "actor_login": null
  },
  {
   "at": null,
   "url": null,
   "kind": "committed",
   "actor_login": "Nikil Ravi"
  },
  {
   "at": null,
   "url": null,
   "kind": "committed",
   "actor_login": "Nikil Ravi"
  },
  {
   "at": null,
   "url": null,
   "kind": "committed",
   "actor_login": "Nikil Ravi"
  },
  {
   "at": "2025-07-27T18:17:30+00:00",
   "url": null,
   "kind": "labeled",
   "actor_login": "dlwh"
  }
 ],
 "related": [
  {
   "agent": "openai-codex",
   "state": "merged",
   "title": "Handle empty glob in classification inference",
   "number": 1487,
   "outcome": "merged_clean",
   "index_ok": true,
   "additions": 28,
   "deletions": 0,
   "merged_at": "2025-07-28T04:32:12+00:00",
   "opened_at": "2025-07-26T21:52:57+00:00",
   "task_type": "fix",
   "files_changed": 2,
   "hours_to_merge": 30.65
  },
  {
   "agent": "openai-codex",
   "state": "merged",
   "title": "Fix vulnerable form-data version",
   "number": 1476,
   "outcome": "merged_clean",
   "index_ok": true,
   "additions": 96,
   "deletions": 47,
   "merged_at": "2025-07-23T03:33:36+00:00",
   "opened_at": "2025-07-23T02:39:34+00:00",
   "task_type": "fix",
   "files_changed": 2,
   "hours_to_merge": 0.9
  },
  {
   "agent": "openai-codex",
   "state": "merged",
   "title": "Fix broken doc link",
   "number": 1473,
   "outcome": "merged_clean",
   "index_ok": true,
   "additions": 1,
   "deletions": 1,
   "merged_at": "2025-07-23T01:22:51+00:00",
   "opened_at": "2025-07-22T21:16:09+00:00",
   "task_type": "fix",
   "files_changed": 1,
   "hours_to_merge": 4.11
  }
 ],
 "pr_url": "https://github.com/marin-community/marin/pull/1489",
 "page": "https://agentriders.com/logs/marin-community/marin/1489/"
}