feat(claude): show the Fable weekly limit and fix /usage parsing for tall screens#217
Merged
Merged
Conversation
…tall screens The Claude CLI and OAuth usage API now report a Fable 5 weekly limit that ClaudeBar ignored. Parse it from both probes so it appears as a quota card in the window and as a selectable menu-bar metric (everything downstream of parsing is already generic over QuotaType.modelSpecific): - CLI probe: parse the "Current week (Fable)" section (paren-open anchor so a future "(Fable 5)" label still matches), with its own reset time falling back to the all-models weekly reset. - API probe: the endpoint reports model limits via a generic `limits` array (kind "weekly_scoped" + scope.model.display_name) instead of the legacy seven_day_<model> fields, which are now null. Parse it generically — any future model-scoped limit appears automatically — keyed on the first word of the display name to stay in sync with the CLI probe's key, deduped against legacy fields, and without clamping over-quota negative remaining. Also fix the CLI probe failing on every tick with recent CLI versions: the /usage screen grew past the probe's 50-row terminal (usage-contribution report), scrolling the quota sections into scrollback, which the terminal renderer never read. Render the whole buffer (visible rows + scrollback) instead of just the visible screen. Verified against a real /usage PTY capture and a live probe: session, weekly, and Fable quotas all parse (quotas=3, "Fable: 67% remaining"). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Contributor
📝 WalkthroughWalkthroughThis PR adds parsing of a Claude Fable 5 weekly quota from both the CLI ChangesFable weekly quota parsing and scrollback support
Estimated code review effort: 3 (Moderate) | ~25 minutes Sequence Diagram(s)sequenceDiagram
participant API as OAuth Usage API
participant CLI as Claude CLI /usage
participant Renderer as TerminalRenderer
participant Probe as Claude Usage Probes
API->>Probe: usage JSON with limits array
Probe->>Probe: filter weekly_scoped, derive model quota
CLI->>Renderer: render output taller than visible rows
Renderer->>Renderer: capture rows + scrollback lines
Renderer-->>Probe: full text incl. off-screen quota sections
Probe->>Probe: extract fablePct, resolve reset time
Probe-->>Probe: append modelSpecific("fable") quota
🚥 Pre-merge checks | ✅ 5✅ Passed checks (5 passed)
✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
🧹 Nitpick comments (1)
Sources/Infrastructure/Claude/ClaudeAPIUsageProbe.swift (1)
474-505: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick winCross-file "fable" key derivation is duplicated and only comment-linked.
The model key here is derived by splitting/lowercasing
displayName, whileClaudeUsageProbe.swifthardcodes the literal"fable"string for the CLI path. The only thing keeping them in sync is a comment (// Key on the first word... must stay in sync with the key the CLI probe hardcodes). If Anthropic changes the API'sdisplay_nameformat (e.g., to "Claude Fable 5"), this loop would silently derive"claude"instead of"fable", breaking the menu-bar persistence contract without any compiler or test signal outside the two now-passing but independently-maintained fixtures.Consider extracting the key-derivation into a single shared helper (e.g. a
QuotaTypestatic function) that both probes call, so any future model-name format change is caught in one place.🤖 Prompt for AI Agents
Verify each finding against current code. Fix only still-valid issues, skip the rest with a brief reason, keep changes minimal, and validate. In `@Sources/Infrastructure/Claude/ClaudeAPIUsageProbe.swift` around lines 474 - 505, The model key derivation for the Claude Fable quota is duplicated between this loop and ClaudeUsageProbe.swift, relying only on a comment to stay in sync. Extract the display-name-to-model-key logic into a shared helper or QuotaType utility and have both the weekly_scoped parsing path and the CLI probe use it, so the persisted model:<name> selection remains consistent if the API displayName format changes.
🤖 Prompt for all review comments with AI agents
Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.
Nitpick comments:
In `@Sources/Infrastructure/Claude/ClaudeAPIUsageProbe.swift`:
- Around line 474-505: The model key derivation for the Claude Fable quota is
duplicated between this loop and ClaudeUsageProbe.swift, relying only on a
comment to stay in sync. Extract the display-name-to-model-key logic into a
shared helper or QuotaType utility and have both the weekly_scoped parsing path
and the CLI probe use it, so the persisted model:<name> selection remains
consistent if the API displayName format changes.
ℹ️ Review info
⚙️ Run configuration
Configuration used: defaults
Review profile: CHILL
Plan: Pro
Run ID: 57a2ef99-7ca3-4947-9aff-446a944bb0a5
📒 Files selected for processing (7)
CHANGELOG.mdSources/Infrastructure/Claude/ClaudeAPIUsageProbe.swiftSources/Infrastructure/Claude/ClaudeUsageProbe.swiftSources/Infrastructure/Shared/TerminalRenderer.swiftTests/DomainTests/Provider/QuotaTypeTests.swiftTests/InfrastructureTests/Claude/ClaudeAPIUsageProbeTests.swiftTests/InfrastructureTests/Claude/ClaudeUsageProbeParsingTests.swift
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## main #217 +/- ##
==========================================
+ Coverage 81.65% 81.83% +0.17%
==========================================
Files 116 116
Lines 8713 8758 +45
==========================================
+ Hits 7115 7167 +52
+ Misses 1598 1591 -7
🚀 New features to boost your workflow:
|
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The Claude CLI and OAuth usage API now report a Fable 5 weekly limit that ClaudeBar ignored. This PR parses it from both probes so it appears as a quota card in the window and as a selectable menu-bar metric. It also fixes a bug where the CLI probe failed on every tick with recent Claude CLI versions.
Fable weekly limit
Current week (Fable)section (paren-open anchor so a future(Fable 5)label still matches), with its own reset time and fallback to the all-models weekly reset.limitsarray (kind: "weekly_scoped"+scope.model.display_name) instead of the legacyseven_day_<model>fields, which are nownull(verified against the live endpoint). Parsing is generic — any future model-scoped limit appears automatically — keyed on the first word of the display name to stay in sync with the CLI probe's key, deduped against legacy fields, and preserving negativepercentRemainingas the over-quota signal per theUsageQuotacontract.QuotaType.modelSpecific, quota cards, the menu-bar quota picker, quota keys likemodel:fable) is already generic.Parse-failure fix (blocks Fable in CLI mode)
Recent Claude CLI versions render a
/usagescreen ~81 lines tall (new usage-contribution report). On the probe's 50-row PTY the quota sections scroll into scrollback, whichTerminalRenderer.extractScreenTextnever read — so parsing failed with "could not find 'Current session'" on every tick. The renderer now reads the whole buffer (visible rows + scrollback, capacity raised to 2000 lines).Verification
limitsarray incl. dedupe/malformed/over-quota cases, QuotaType round-trip); full Domain + Infrastructure suites pass./usagePTY capture (160×50 raw bytes) and a live app run: probe succeeds withquotas=3— Session 35% / Weekly 82% / Fable 67% remaining — matching the CLI exactly.Follow-up (not in this PR)
Summary by CodeRabbit
New Features
Bug Fixes