Deferred review findings from the CLI and download PR series

オープン
#173 コメント 0 件 リアクション 0 件 担当者 0 名 GitHub で見る

メンテナーはふだん 1 日以内に返信

まだ誰も着手していません。

評価

難易度
4/5
見積もり時間
3〜5日
初心者へのやさしさ
25/100
issue の種類
リファクタリング
明瞭さ
説明が足りない
活発さ
活発
技術スタック
python
領域
cli

調査の方向性

Start by reading the cited CLI and download PRs (#167–#172) and structural issues #159, #161, #162, #163; the findings span viral stores, exit statuses, shared options, and data commands. Check which findings remain relevant after those issues, then address a narrowly scoped item and verify its behavior with the relevant CLI tests. Several findings are explicitly superseded or folded into other work, so the issue does not define one standalone completion target.

索引モデルが issue の本文から書いたものです。

説明

enhancement

Code review of the CLI and download PR series (#167–#172) raised findings that were deliberately not fixed there. They are pre-existing gaps, pathological states, or cleanups that the structural issues (#159, #161, #162) will make easier or unnecessary. The blocking findings (regressions, data loss, false claims) were fixed in the PRs.

Viral stores (#170): mostly superseded once #159 moves collections onto VersionedDatasetRegistry

  • datacache.bundle_store is imported although DataCache doesn't export it (the import is isolated in _bundle_store() with a clear error). The durable fix is to export BundleStore and a diagnose/repair hook from DataCache.
  • _store_problem checks stores with rules that can differ from DataCache's own BundleStore checks, for example on stray .datacache-* files or a hard-linked or FIFO marker.
  • A future-dated stray entry under bundles/ sets the names of later bundles, and prune skips it.
  • select_viral_references copies from the collection bundle outside the collection lock, so a concurrent explicit prune_viral_references(keep=1) can delete its source mid-copy.
  • The store marker is written without fsync, and prune_viral_references reimplements prune_bundles.
  • prune_viral_references only reaches the current bundle version and index schema.
  • A damaged selection is hashed twice before it is rebuilt.
  • Replaced bundles are kept until an explicit prune (about 10 MB per refresh, by design).

Exit statuses (#172)

  • --reference and --asset on search-reference commands are checked separately, so a valid but mismatched pair exits 1 instead of 2.
  • Restricted-terms checks are duplicated per command (require_terms). A TermsNotAccepted(ValueError) raised by download_chain and fetch_data and mapped by ArgumentParser.fail would cover every caller.
  • Registry-file wrapping is duplicated in catalog_cli and dataset_cli, and ASSET=PATH parsing in catalog_cli, genomes_cli and download_cli (fold into #162 and #163).
  • Some option values are still validated only by library code and exit 1, as docs/cli.md says.

Shared options (#171)

  • share_cache_root must be called by hand at the end of each group's parser setup; a subcommand added after the call loses the option. --registry still works only before the subcommand.

Data commands (#168, #169)

  • A repair leaves the damaged DataCache generation on disk until the cache is pruned (documented).
  • _show writes its own truncation footer next to print_table's.
  • The focused commands (data catalog, data dataset, …) still need --force to repair a damaged install (#162 turns them into data fetch aliases).
主要言語
Python
スター
1
フォーク
1
平均マージ
10時間 11分
マージ済み PR(30日)
51

環境構築

このプロジェクトには開発コンテナ、Dockerfile、コントリビューションガイドがありません。まず README を読み、一般的な手順ははじめてのコントリビューションガイドを参照してください。

はじめの一歩

  1. issue を最後まで読み、次にプロジェクトのコントリビューションガイドを読みます。
  2. 着手することを issue にコメントします — 二人が同じ作業をするのを防げます。
  3. リポジトリをフォークし、ブランチを切って変更します。
  4. issue 番号を参照したプルリクエストを送ります。

openvax/orfdata のほかの issue

openvax/orfdata の issue をすべて見る

似ている issue

Python の issue をもっと見る

新しい issue をメールで受け取る

初心者向けの GitHub issue を短くまとめたダイジェスト。