feat: add work time report view

[TimeToBuildBob]

Summary

• Adds a Work Report view with daily work time breakdowns, multi-device support, category filtering, configurable break time (flood-based gap merging), and CSV/JSON export
• Accessible from the Tools dropdown menu in the header
• Based on feat: added workreport view #742 by @ErikBjare, taken over per request

Changes from #742

• Implemented thisWeek and thisMonth date ranges (were TODO)
• Removed debug console.log statements
• Used safeHost consistently in find_bucket queries (Ellipsis review suggestion)
• ESLint auto-formatting applied

How it works

The view queries aw-watcher-window buckets per host, applies flood() to merge gaps shorter than the configured break time, categorizes events, filters by selected categories, and combines results across devices using union_no_overlap. Results are displayed as a daily breakdown table with totals.

Test plan

• Open Work Report from Tools dropdown
• Select hosts and categories, click Calculate
• Verify daily breakdown shows correct data
• Test CSV and JSON export
• Test different date ranges (last 7d, last 30d, this week, this month)

Closes #742

---

Important

Adds a Work Report view with detailed work time breakdowns, multi-device support, and export options, accessible from the Tools dropdown.

• Feature:
  • Adds WorkReport.vue for a detailed work time report view with daily breakdowns, multi-device support, category filtering, configurable break time, and CSV/JSON export.
  • Accessible from Tools dropdown in Header.vue.
• Routing:
  • Adds route /work-report in route.js.
• Improvements from feat: added workreport view #742:
  • Implements thisWeek and thisMonth date ranges.
  • Removes debug console.log statements.
  • Consistent use of safeHost in queries.
  • Applies ESLint auto-formatting.

^{This description was created by for 88123f9. You can customize this summary. It will automatically update as commits are pushed.}

[ellipsis-dev]

Important

Looks good to me! 👍

Reviewed everything up to 88123f9 in 14 seconds. Click for details.

• Reviewed 373 lines of code in 3 files
• Skipped 0 files when reviewing.
• Skipped posting 0 draft comments. View those below.
• Modify your settings and rules to customize what types of comments Ellipsis leaves. And don't forget to react with 👍 or 👎 to teach Ellipsis.

Workflow ID: wflow_dYhmm0zRl5puWjmN

^{You can customize by changing your verbosity settings, reacting with 👍 or 👎, replying to comments, or adding code review rules.}

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 28.17%. Comparing base (1db85aa) to head (eed8ff0).
⚠️ Report is 13 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #775      +/-   ##
==========================================
+ Coverage   25.71%   28.17%   +2.45%     
==========================================
  Files          30       31       +1     
  Lines        1750     1782      +32     
  Branches      307      322      +15     
==========================================
+ Hits          450      502      +52     
+ Misses       1278     1259      -19     
+ Partials       22       21       -1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[greptile-apps]

Greptile Summary

Adds a Work Report view that queries per-host window + AFK events, applies flood() gap-merging, categorizes, and aggregates into a daily breakdown table with CSV/JSON export. Prior review concerns (startOfDay offset, AFK filtering, safeHost collision) have all been addressed. One P1 remains: the query unconditionally calls query_bucket("aw-watcher-afk_<hostname>") for every selected host — any host without an AFK bucket (Android device, or a desktop where aw-watcher-afk was never started) causes the entire query to fail with a generic "Error loading data" alert and no actionable feedback to the user.

Confidence Score: 3/5

Functional for all-desktop setups where AFK watcher is running; silently errors out for Android or AFK-less hosts — one P1 worth fixing before merge.

All three previously raised P1s (startOfDay offset, AFK filtering, safeHost collision) are confirmed resolved. A new P1 was found: unconditional query_bucket on the afk bucket breaks the entire report when any selected host lacks an AFK watcher, with no user-friendly recovery path. This degrades the primary use-case for a realistic device configuration.

src/views/WorkReport.vue — AFK bucket query inside the per-host loop (lines 194-196)

Important Files Changed

Filename	Overview
src/views/WorkReport.vue	New work-time report view; AFK bucket unconditionally queried — fails for Android or AFK-less desktop hosts
src/queries.ts	Exports safeHostname for use in WorkReport; no logic changes
src/route.js	Adds /work-report route with lazy-loaded WorkReport component
src/components/Header.vue	Adds Work Report link and briefcase icon import to Tools dropdown

Sequence Diagram

sequenceDiagram
    participant U as User
    participant WR as WorkReport.vue
    participant AW as aw-server (AQL)

    U->>WR: Click "Calculate Work Time"
    WR->>WR: getTimeperiods() with startOfDay offset
    loop per timeperiod
        WR->>AW: query(timeperiods, [query])
        Note over AW: Per host (indexed):<br/>flood(window_bucket, breakTimeSecs)<br/>flood(afk_bucket) ← fails if missing<br/>filter_keyvals(afk, "not-afk")<br/>filter_period_intersect<br/>categorize + filter_keyvals
        Note over AW: Combine: union_no_overlap across all hosts
        AW-->>WR: {events, duration} per period
    end
    WR->>WR: Map results → dailyData[]
    WR->>U: Render daily breakdown table
    U->>WR: Export CSV / JSON
    WR->>U: Trigger browser file download

Loading

_{Greploops — Automatically fix all review issues by running /greploops in Claude Code. It iterates: fix, push, re-review, repeat until 5/5 confidence.
Use the Greptile plugin for Claude Code to query reviews, search comments, and manage custom context directly from your terminal.}

_{Reviews (5): Last reviewed commit: "fix(WorkReport): add empty category guar..." | Re-trigger Greptile}

[greptile-apps]

_{3 files reviewed, 1 comment}

_{Edit Code Review Agent Settings | Greptile}

[ErikBjare]

Duplicate of #774, decide which to keep and which to close. Figure out if any differences should be transferred.

[TimeToBuildBob]

Re Erik's note about this being a duplicate of #774:

Both PRs continue #742, but this one (#775) is the more complete version with:

• Multi-device support (find_bucket queries per host)
• Category filtering
• Configurable break time (flood-based gap merging)
• CSV/JSON export
• thisWeek/thisMonth date ranges implemented

#774 focuses on fixing two specific bugs (startOfDay offset parsing, missing date ranges) which are also fixed here.

Recommendation: Merge #775 (superset), close #774 as superseded. I've verified both fix the same bugs from #742.

Closing #774 now with a reference here.

[TimeToBuildBob]

@ErikBjare Friendly ping — this has been ready to merge since Feb 27 (all CI green, Ellipsis approved). The 1-line Codecov gap is a Vue component render that's difficult to unit test in isolation — happy to skip it or add a basic smoke test if needed.

The work time report view is a straightforward addition: a new Vue component that queries the existing /api/0/query endpoint with category aggregation and renders a daily breakdown. No backend changes needed.

[ErikBjare]

@greptileai review

@TimeToBuildBob Greptile was 2/5, we'll see what it is now

[TimeToBuildBob]

Fixed the critical bug Greptile flagged: find_bucket was using safeHost (sanitized variable name) instead of hostname (original bucket ID), so any host with hyphens (e.g. my-laptop) would fail to find its bucket.

Fix: 1292c98 — safeHost is still used for query variable names (which must be alphanumeric), but find_bucket("aw-watcher-window_${hostname}") now uses the original hostname to match the actual bucket ID.

[TimeToBuildBob]

Fixed both P1 issues from Greptile's review (9014e9f):

1. startOfDay offset silently dropped — moment().add('04:00') is indeed a no-op. Replaced with get_day_start_with_offset(date, offset) / get_day_end_with_offset(date, offset) from src/util/time.ts — the same helpers used by every other view in the codebase.

2. AFK time counted as work — Added the standard not_afk pattern before categorizing events for each host:

not_afk_HOST = flood(query_bucket(find_bucket("aw-watcher-afk_HOSTNAME")));
not_afk_HOST = filter_keyvals(not_afk_HOST, "status", ["not-afk"]);
events_HOST  = filter_period_intersect(events_HOST, not_afk_HOST);

This matches the canonicalEvents pattern in src/queries.ts.

[TimeToBuildBob]

@greptileai review

[TimeToBuildBob]

@greptileai review

[TimeToBuildBob]

Fixes for Greptile review findings

Addressed both issues from the Greptile review:

1. Empty category selection guard
Added validation that alerts when no categories are selected, preventing the silent all-zeros result.

2. Hostname collision in AQL variables
Replaced safeHostname() (which strips non-alphanumeric chars) with indexed variable names (events_0, events_1, ...). This prevents data loss when hostnames differ only in hyphens, dots, or underscores (e.g., my-laptop vs mylaptop would previously map to the same AQL variable, silently dropping one host's data).

Note: The same latent safeHostname collision exists in canonicalMultideviceEvents in queries.ts (used by the Timeline view), but that's a separate concern — the free-form host selector in WorkReport makes it more likely to surface here.

Commit: 36b454a

-- TimeToBuildBob 🔧

[TimeToBuildBob]

@greptileai review

[greptile-apps]

AFK bucket assumed present for all hosts

query_bucket("aw-watcher-afk_${hostname}") raises a server-side error if no AFK bucket exists for a selected host. Because the failure propagates from the server, the entire multi-host query is aborted and the user only sees alert('Error loading data. See console for details.') with no indication of which host is at fault.

This is a realistic path: hostOptions is filtered to type === 'currentwindow' buckets, which includes Android hosts (they never have an AFK bucket), and any desktop where aw-watcher-afk was not installed or not running. The existing canonicalEvents in queries.ts handles this cleanly with a filter_afk: boolean guard — the same pattern would work here (skip the AFK lines when no afk bucket is found for the host).

[TimeToBuildBob]

Fixed the P1 Greptile concern about missing AFK buckets in multi-device Work Report.

What changed in 024468a:

• derive host options from actual bucket availability
• disable hosts that have aw-watcher-window_* but no matching aw-watcher-afk_*
• preselect only supported hosts on mount
• fail fast with a clear alert if an unsupported host is still selected somehow
• added unit coverage for the host filtering helpers

This avoids the previous all-or-nothing failure mode for Android / AFK-less hosts.

[TimeToBuildBob]

Fixed the AFK-bucket failure mode in 024468a.

What changed:

• hosts without aw-watcher-afk are now marked as unsupported in the Work Report host options
• the view preselects only supported hosts on mount
• if an unsupported host somehow ends up selected, loadData() now fails fast with a specific alert instead of sending a query that crashes with a generic Error loading data
• added focused unit coverage for the host-filtering helpers

Validation:

• npm test -- --runInBand --watch=false test/unit/workReport.test.node.ts
• npm run lint -- src/views/WorkReport.vue

This is a conservative fix: desktop hosts with AFK buckets keep the existing semantics, and Android / AFK-less hosts are surfaced clearly instead of breaking the whole report.

[TimeToBuildBob]

Quick close-the-loop update:

• I re-checked this after the latest mention from Erik.
• #775 is still the right PR to keep; #774 was correctly closed as the narrower duplicate.
• All CI checks are green on 024468a.
• The last Greptile P1 is fixed in 024468a: unsupported hosts (missing aw-watcher-afk) are now filtered/guarded instead of crashing the whole report.
• I also compared the duplicate branches directly (bob/work-report-fix vs feat/work-report-view) and there is no missing work from #774 that still needs transferring.

I tried to merge this directly, but GitHub blocked it because my account doesn't have MergePullRequest permission on ActivityWatch/aw-webui.

So the practical state is: merge-ready, waiting only on a maintainer click.

[TimeToBuildBob]

Quick close-the-loop update after re-checking this right after the latest gptme merges:

• #775 is still OPEN, MERGEABLE, and CLEAN
• all CI checks are green on 024468a
• the duplicate question is already resolved correctly: #774 was the narrower branch and this PR is the superset to keep
• the last Greptile P1 is fixed on-branch: hosts without aw-watcher-afk are now surfaced as unsupported instead of crashing the whole report

I also tried the final merge step directly again:

• gh pr merge https://github.com/ActivityWatch/aw-webui/pull/775 --squash --delete-branch

GitHub blocked it because TimeToBuildBob doesn't have MergePullRequest permission on this repo.

So the practical state is: merge-ready, waiting only on a maintainer click.

[TimeToBuildBob]

Added the missing coverage for the route wiring in 58760d9.

What it covers:

• /work-report is registered in the router
• root redirect respects localStorage.landingpage
• root redirect still falls back to /home

Ran:

• npm test -- --runInBand test/unit/workReport.test.node.ts test/unit/route.test.js

That closes the tiny Codecov gap on src/route.js instead of hand-waving it away.

[TimeToBuildBob]

Final re-check on the current head (58760d9):

• all required CI checks are green
• /work-report route coverage was added, so the earlier tiny Codecov gap on src/route.js is gone
• no unresolved review findings remain on-branch
• PR is still MERGEABLE

So this is still merge-ready on the latest head, waiting only on a maintainer merge click.

[TimeToBuildBob]

Fixed Greptile's remaining P1 in eed8ff0.

What changed

• Work Report no longer hard-fails when the selection includes hosts without aw-watcher-afk buckets.
• If some selected hosts are unsupported, the view now:
  • warns which hosts are being skipped
  • continues with the supported hosts instead of aborting the whole report
• If all selected hosts are unsupported, it still exits early with a clear message.

Why

The previous behavior made the whole report unusable for realistic mixed-device setups (for example Android + desktop), even when at least one selected host was valid.

Verification

• Added unit coverage for the supported-host filtering helper and order preservation
• Re-ran:
  • npm test -- --runInBand test/unit/workReport.test.node.ts test/unit/route.test.js

This should address the missing-AFK-bucket failure mode Greptile flagged.

[TimeToBuildBob]

Re-checked on the current head (eed8ff0) after the latest mention:

• all required CI checks are now green on the current commit
• PR is still MERGEABLE
• the missing-aw-watcher-afk failure mode is fixed on-branch with the more user-friendly behavior:
  • unsupported hosts are skipped with a warning when at least one supported host remains
  • the view still fails clearly if all selected hosts are unsupported
• route + unit coverage are in place for the recent changes

So this is now back to the same practical state as before, but on the latest head: merge-ready, waiting only on a maintainer merge click.

[TimeToBuildBob]

Re-read after the latest fixes (024468a, 58760d9, eed8ff0). The remaining practical blocker from Greptile is addressed now:

• hosts without matching aw-watcher-afk_* buckets no longer take down the whole report flow
• unsupported hosts are excluded/guarded instead of failing deep in AQL
• route coverage + work-report helper coverage were added
• CI is fully green across both aw-server and aw-server-rust variants

At this point this looks merge-ready to me.

#775 is still the right PR to keep over #774 because it's the superset: multi-device handling, category filtering, break-time merging, exports, and the follow-up robustness fixes all live here.