gsoc26: Manifest Replay for download, delete, and deploy (#65)

[DhanashreePetare]

Pull Request

Description

Implements databusclient manifest replay <path> for all three commands — download, delete, and deploy.

• Download replay: re-executes a saved download using the recorded replayParams (compression, format conversion, graph-name, base-uri, checksum validation). Credentials (--vault-token/--databus-key) and --localdir are never read from the manifest — always supplied fresh at replay time.
• Delete replay: interactive y/yes confirmation by default before deleting, per mentor discussion. --force (CLI-only, never stored in the manifest) skips the prompt for scripted/CI use. dry_run is read automatically from whatever was recorded in the original delete --dry-run --manifest ... run; a replay-time --dry-run override can additionally force a preview.
• Deploy replay: supports classic and metadata-file modes. Per mentor guidance, the manifest stores fully-resolved deployment metadata (checksum, size, format, compression already computed) rather than raw pre-construction inputs, so replay never re-downloads or recomputes file checksums — it reconstructs create_dataset() inputs directly from the stored data. WebDAV/Nextcloud mode is explicitly not supported for replay, since locally uploaded files may no longer exist at their original paths by replay time.

All three replay paths raise a clean, actionable error (never a raw traceback) for: missing/corrupted manifest files, missing required fields, missing credentials, and unsupported commands.

Related Issues
#65

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• This change requires a documentation update
• Housekeeping

Checklist:

• My code follows the ruff code style of this project.
• I have performed a self-review of my own code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (if applicable)
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
  • poetry run pytest - all tests passed
  • poetry run ruff check - no linting errors

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on base/target branches other than the default branch.

Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 9d493f6f-28ab-4475-a468-2c9fe978dd57

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[Integer-Ctrl]

Usability tests did not reveal any bugs

[DhanashreePetare]

Apart from the review comment fixes, the latest commit also contains a small improvement to manifest summary: when an operation failed, the summary now also shows the actual error message (the "cause"), not just "Status: failed" — so users don't have to open the raw manifest file to see why something failed.

Also included two small, unrelated-to-replay fixes, to include here rather than as separate issues since they're small and genuinely affect users:

1. BOM fix for deploy --metadata: some Windows tools save JSON files with a hidden UTF-8 BOM character at the start. The deploy command was failing to read those files. Now it handles both cases correctly.
2. Clean error messages for deploy: if a deploy failed for any reason (bad API key, network issue, etc.), the command was printing a full raw Python error trace instead of a simple message. Now it shows just the actual error, the same clean way download and manifest replay already do.