Cursor: per-project breakdown by workspace (closes per-project half of #196) (#296)

Cursor's chat history showed as a single row labeled 'cursor' in
the dashboard because the global state.vscdb has no workspace
field on individual bubbles. The fix joins through Cursor's
per-workspace storage:

1. Walk ~/Library/Application Support/Cursor/User/workspaceStorage/*
2. For each hash dir, read workspace.json -> folder URI
3. Open that dir's state.vscdb, read
   ItemTable['composer.composerData'] -> allComposers list
4. Build Map<composerId, folder URI>
5. emit one SessionSource per workspace plus a catch-all 'cursor'
   source for composers that did not register against any
   workspace (multi-root workspaces, no-folder-open windows,
   deleted workspaces with surviving global rows)

The parser decodes source.path's #cursor-ws= tag, filters the
parsed bubbles to the composerIds that belong to this workspace,
and yields only those. The orphan-tag source negates the filter so
it captures every composer not in any workspace.

In passing, fix a real bug in the old code: parseBubbles set
`sessionId: row.conversation_id ?? 'unknown'`, but the JSON
`conversationId` field is empty in current Cursor builds, so every
call shipped with `sessionId: 'unknown'`. We now derive the
composer id from the row key (`bubbleId:<composerId>:<bubbleUuid>`)
which is what the workspace map joins on. The old behavior masked
the bug because every call went into a single 'cursor' project
anyway; with per-workspace bucketing the bug becomes load-bearing.
Cache version bumped 2 -> 3 to invalidate caches that still record
'unknown' as the session id.

Live-tested against my real 1.9 GB Cursor DB: the single 'cursor'
row with 1904 calls / $4.08 now breaks into 5 workspaces plus an
orphan bucket, totals reconcile exactly. 8 fixture-based tests
cover multi-workspace routing, orphan filtering, legacy bare DB
path backwards compat, multi-root workspace skip, vscode-remote
URI slugification, and total reconciliation across all sources.

Full suite: 46 files, 653 tests passing.

Author: iamtoruk

CHANGELOG.md

@@ -41,6 +41,23 @@
   reconcile. Closes #279.
 
 ### Fixed (CLI)
+- **Cursor sessions break down by project, not one row called "cursor".**
+  Cursor's chat history sat under a single dashboard row labeled `cursor`
+  because the provider had no way to attribute bubbles to a workspace.
+  The fix walks `~/Library/Application Support/Cursor/User/workspaceStorage/*`
+  for each workspace's `workspace.json` (folder URI) and
+  `composer.composerData` (the composer ids opened in that workspace),
+  then joins those composer ids against the global bubbles. Each
+  workspace becomes its own project row, sanitized into the same slug
+  shape Claude uses (e.g. `-Users-you-myproject`); composers that have
+  no workspace mapping (multi-root workspaces, "no folder open"
+  sessions, deleted workspaces) remain under a catch-all `cursor` row.
+  As part of this the cursor parser now derives `sessionId` from the
+  bubble row key (`bubbleId:<composerId>:<bubbleUuid>`) instead of the
+  empty `conversationId` JSON field, which was always falling back to
+  `'unknown'`. Cursor result cache version bumped to 3 to invalidate
+  prior caches that recorded the old session id. Closes the per-project
+  half of #196.
 - **Cursor cost shown for every model, not just Auto.** Cursor emits model
   names in a `claude-<dot-version>-<tier>` shape (`claude-4.6-sonnet`,
   `claude-4.5-opus`, `claude-4.5-opus-high-thinking`, etc.) plus its own

src/cursor-cache.ts

@@ -5,7 +5,13 @@ import { randomBytes } from 'crypto'
 
 import type { ParsedProviderCall } from './providers/types.js'
 
-const CURSOR_CACHE_VERSION = 2
+// Bumped to 3 for the workspace-aware breakdown change: the cursor parser
+// now derives `sessionId` from the bubble row key (the real composer id)
+// rather than the empty `conversationId` JSON field, and the workspace
+// router relies on those composer ids to bucket calls per project.
+// Version 2 caches contain `sessionId: 'unknown'` for every call and would
+// route everything to the orphan project, so we invalidate them.
+const CURSOR_CACHE_VERSION = 3
 
 type ResultCache = {
   version?: number

src/daily-cache.ts

@@ -5,7 +5,14 @@ import { homedir } from 'os'
 import { join } from 'path'
 import type { DateRange, ProjectSummary } from './types.js'
 
-export const DAILY_CACHE_VERSION = 4
+// Bumped to 5 alongside the Cursor per-project breakdown: prior daily
+// entries recorded every Cursor session under a single 'cursor' project
+// label. After the upgrade, the breakdown produces per-workspace project
+// labels for new days; without invalidation the dashboard would show
+// 'cursor' for historical days and `-Users-you-myproject` for new ones
+// in the same window, producing a confusing mixed projection. v5 forces a
+// full recompute.
+export const DAILY_CACHE_VERSION = 5
 const MIN_SUPPORTED_VERSION = 2
 const DAILY_CACHE_FILENAME = 'daily-cache.json'