mirror of
https://github.com/AgentSeal/codeburn.git
synced 2026-07-22 15:34:29 +00:00
VS Code GitHub Copilot Chat users with no OTel agent-traces.db and no ~/.copilot/session-state/ saw $0.00 cost: codeburn never read VS Code's core chat persistence, the only on-disk source carrying their token counts. Add a fourth Copilot source that reads the VS Code chat delta-journals at workspaceStorage/<hash>/chatSessions/*.jsonl and globalStorage/emptyWindowChatSessions/*.jsonl. The files are a kind:0 snapshot / kind:1 path-set / kind:2 array-append journal; we replay it (prototype-pollution-safe: __proto__/prototype/constructor segments are rejected and containers use Object.create(null)) and read each request's result.metadata.promptTokens / outputTokens (falling back to completionTokens) and resolvedModel for pricing. Dedup so users with multiple sources are not double-counted: prefer OTel (skip chatSessions discovery when an OTel source is present), and skip a workspace's legacy GitHub.copilot-chat/transcripts when that workspace has chatSessions. New env overrides CODEBURN_COPILOT_GLOBAL_STORAGE_DIR and the existing CODEBURN_COPILOT_WS_STORAGE_DIR keep discovery testable. Tests cover the reporter's real request shape (promptTokens 32543 / outputTokens 60 -> non-zero cost), empty sessions, emptyWindow discovery, append-then-edit replay, requestId dedup, prototype-pollution paths, the transcript skip, and the OTel-prefer skip.
This commit is contained in:
ozymandiashh
GitHub
parent
d5002f1deb
commit
b24dd31ede
3 changed files with 694 additions and 36 deletions
|
|
@ -1,23 +1,25 @@
|
|||
# Copilot
|
||||
|
||||
GitHub Copilot Chat (CLI and VS Code extension transcripts).
|
||||
GitHub Copilot Chat (CLI, VS Code core chat sessions, and VS Code extension transcripts).
|
||||
|
||||
- **Source:** `src/providers/copilot.ts`
|
||||
- **Loading:** eager (`src/providers/index.ts:3`)
|
||||
- **Test:** `tests/providers/copilot.test.ts` (401 lines)
|
||||
- **Test:** `tests/providers/copilot.test.ts`
|
||||
|
||||
## Where it reads from
|
||||
|
||||
Two JSONL locations plus an optional OpenTelemetry SQLite source (see below). All
|
||||
discovered sources are walked on every run; results merge and dedupe.
|
||||
Three JSONL locations plus an optional OpenTelemetry SQLite source (see below). OTel is
|
||||
preferred when present; chatSessions are only discovered when no OTel source is found.
|
||||
Other discovered sources are walked on every run; results merge and dedupe.
|
||||
|
||||
1. **Legacy CLI sessions:** `~/.copilot/session-state/`
|
||||
2. **VS Code transcripts:** `~/Library/Application Support/Code/User/workspaceStorage/<hash>/GitHub.copilot-chat/transcripts/` and equivalents on Windows / Linux
|
||||
3. **OTel SQLite store:** VS Code Copilot Chat's `agent-traces.db` (see the OTel section). Preferred when present because it carries full input / output / cache token counts; the JSONL sources only record output tokens.
|
||||
2. **VS Code core chat sessions:** `~/Library/Application Support/Code/User/workspaceStorage/<hash>/chatSessions/*.jsonl` plus `~/Library/Application Support/Code/User/globalStorage/emptyWindowChatSessions/*.jsonl` and equivalents on Windows / Linux
|
||||
3. **VS Code transcripts:** `~/Library/Application Support/Code/User/workspaceStorage/<hash>/GitHub.copilot-chat/transcripts/` and equivalents on Windows / Linux
|
||||
4. **OTel SQLite store:** VS Code Copilot Chat's `agent-traces.db` (see the OTel section). Preferred when present because it carries full input / output / cache token counts; legacy JSONL sources only record output tokens.
|
||||
|
||||
## Storage format
|
||||
|
||||
JSONL in the first two locations (schemas differ; the parser switches by detecting which schema the first event uses), and a SQLite DB for the OTel source.
|
||||
JSONL in the first three locations (schemas differ; the parser switches by source type / event shape), and a SQLite DB for the OTel source. VS Code core chat sessions use a delta journal: `kind:0` sets the root object, `kind:1` writes a value at path `k`, and `kind:2` appends items to an array path.
|
||||
|
||||
## OpenTelemetry (OTel) source
|
||||
|
||||
|
|
@ -26,6 +28,11 @@ breakdowns (input, output, cache-read, cache-creation) from it, which the JSONL
|
|||
not record. Discovery is skipped with `CODEBURN_COPILOT_DISABLE_OTEL=1`, and the DB path
|
||||
can be overridden with `CODEBURN_COPILOT_OTEL_DB`.
|
||||
|
||||
If OTel discovery finds at least one source, workspace `chatSessions/*.jsonl` and
|
||||
`emptyWindowChatSessions/*.jsonl` are skipped. Those journals can mirror the same Copilot
|
||||
turns under IDs that do not match OTel turn IDs, so CodeBurn prefers the richer OTel data
|
||||
instead of trying to dedupe across stores.
|
||||
|
||||
- **Requires Node 22+.** The OTel source uses the built-in `node:sqlite` module (the same
|
||||
backend as Cursor / OpenCode). On Node 20, or if the DB is missing / locked / corrupt /
|
||||
wrong-schema, OTel is skipped and the JSONL/transcript sources are used as a fallback.
|
||||
|
|
@ -43,7 +50,9 @@ None for the JSONL sources. The OTel source uses a durable cache (see above).
|
|||
|
||||
## Deduplication
|
||||
|
||||
Per `messageId` in both formats (`copilot.ts:118` for legacy, `copilot.ts:245` for transcripts).
|
||||
Legacy JSONL and transcript sessions dedupe per `messageId`. Core chat sessions dedupe per `copilot-chatsession:<sessionId>:<requestId>`, and are not discovered when an OTel source is present.
|
||||
|
||||
If a workspace hash contains at least one `chatSessions/*.jsonl` file, the provider skips that hash's legacy `GitHub.copilot-chat/transcripts/` directory. The core chat session journal is the modern token-bearing source for the same conversations, so reading both would inflate call counts.
|
||||
|
||||
## Model inference
|
||||
|
||||
|
|
|
|||
|
|
@ -8,11 +8,11 @@
|
|||
// cache-read tokens, and cache-creation tokens are never written there, so
|
||||
// CodeBurn underreports Copilot costs by 60-80%.
|
||||
//
|
||||
// This modified version adds a SECOND data source: VS Code Copilot Chat's
|
||||
// OTel SQLite store (agent-traces.db). When present, it contains full
|
||||
// per-LLM-call token breakdowns (input, output, cache_read, cache_creation)
|
||||
// from the OpenTelemetry GenAI semantic conventions. We prefer OTel data
|
||||
// when available and fall back to the original JSONL parsing.
|
||||
// This modified version adds VS Code sources that can carry fuller token
|
||||
// data: the OTel SQLite store (agent-traces.db), VS Code core chatSessions
|
||||
// journals, and legacy extension transcripts. OTel and chatSessions contain
|
||||
// input/output token breakdowns for Copilot Chat users; legacy JSONL remains
|
||||
// a fallback when richer sources are absent.
|
||||
//
|
||||
// HOW TO ENABLE THE OTEL SQLITE STORE:
|
||||
// TWO settings must both be enabled in VS Code settings.json:
|
||||
|
|
@ -37,11 +37,13 @@
|
|||
// ENVIRONMENT VARIABLES:
|
||||
// CODEBURN_COPILOT_OTEL_DB — Override the agent-traces.db path
|
||||
// CODEBURN_COPILOT_DISABLE_OTEL=1 — Skip OTel entirely, use only JSONL
|
||||
// CODEBURN_COPILOT_WS_STORAGE_DIR — Override VS Code workspaceStorage
|
||||
// CODEBURN_COPILOT_GLOBAL_STORAGE_DIR — Override VS Code globalStorage
|
||||
//
|
||||
// ARCHITECTURE:
|
||||
// discoverSessions() returns BOTH OTel sessions (one per conversation_id)
|
||||
// and JSONL sessions. The OTel sessions are deduped against JSONL by
|
||||
// conversation ID so we don't double-count. OTel sessions carry the full
|
||||
// discoverSessions() returns OTel sessions and legacy JSONL sessions. When
|
||||
// OTel is present, VS Code core chatSessions are skipped because they mirror
|
||||
// the same Copilot turns under different IDs. OTel sessions carry the full
|
||||
// token breakdown; JSONL sessions only carry output tokens (the original
|
||||
// behaviour, as a fallback).
|
||||
//
|
||||
|
|
@ -188,6 +190,9 @@ type CopilotEvent =
|
|||
| { type: 'assistant.message'; data: AssistantMessageData; timestamp?: string }
|
||||
| { type: 'subagent.selected'; data: SubagentSelectedData; timestamp?: string }
|
||||
|
||||
type ChatJournalPathSegment = string | number
|
||||
type ChatSessionRequest = Record<string, unknown>
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Types for OTel span rows from agent-traces.db
|
||||
// ---------------------------------------------------------------------------
|
||||
|
|
@ -334,6 +339,205 @@ function epochToISO(epoch: number): string {
|
|||
return new Date(ms).toISOString()
|
||||
}
|
||||
|
||||
function timestampToISO(raw: unknown): string {
|
||||
if (typeof raw === 'number' && Number.isFinite(raw) && raw > 0) {
|
||||
return epochToISO(raw)
|
||||
}
|
||||
if (typeof raw !== 'string') return ''
|
||||
const trimmed = raw.trim()
|
||||
if (!trimmed) return ''
|
||||
if (/^\d+(\.\d+)?$/.test(trimmed)) {
|
||||
return epochToISO(Number(trimmed))
|
||||
}
|
||||
const parsed = Date.parse(trimmed)
|
||||
return Number.isNaN(parsed) ? '' : new Date(parsed).toISOString()
|
||||
}
|
||||
|
||||
function isRecord(value: unknown): value is Record<string, unknown> {
|
||||
return typeof value === 'object' && value !== null && !Array.isArray(value)
|
||||
}
|
||||
|
||||
function isReplayContainer(value: unknown): value is object {
|
||||
return typeof value === 'object' && value !== null
|
||||
}
|
||||
|
||||
function createReplayObject(): Record<string, unknown> {
|
||||
return Object.create(null) as Record<string, unknown>
|
||||
}
|
||||
|
||||
const FORBIDDEN_CHAT_JOURNAL_KEYS = new Set(['__proto__', 'prototype', 'constructor'])
|
||||
|
||||
function parseChatJournalPath(rawPath: unknown, fallback?: ChatJournalPathSegment[]): ChatJournalPathSegment[] | null {
|
||||
const value = rawPath === undefined ? fallback : rawPath
|
||||
if (!Array.isArray(value)) return null
|
||||
|
||||
const path: ChatJournalPathSegment[] = []
|
||||
for (const segment of value) {
|
||||
if (typeof segment === 'number') {
|
||||
if (!Number.isInteger(segment) || segment < 0) return null
|
||||
path.push(segment)
|
||||
continue
|
||||
}
|
||||
if (typeof segment === 'string') {
|
||||
if (FORBIDDEN_CHAT_JOURNAL_KEYS.has(segment)) return null
|
||||
path.push(segment)
|
||||
continue
|
||||
}
|
||||
return null
|
||||
}
|
||||
return path
|
||||
}
|
||||
|
||||
function getReplayValue(container: object, segment: ChatJournalPathSegment): unknown {
|
||||
return (container as Record<string, unknown>)[String(segment)]
|
||||
}
|
||||
|
||||
function setReplayValue(container: object, segment: ChatJournalPathSegment, value: unknown): void {
|
||||
;(container as Record<string, unknown>)[String(segment)] = value
|
||||
}
|
||||
|
||||
function createContainerForNext(segment: ChatJournalPathSegment): unknown[] | Record<string, unknown> {
|
||||
return typeof segment === 'number' ? [] : createReplayObject()
|
||||
}
|
||||
|
||||
function ensureReplayParent(root: object, path: ChatJournalPathSegment[]): object | null {
|
||||
let current: object = root
|
||||
for (let i = 0; i < path.length - 1; i++) {
|
||||
const segment = path[i]!
|
||||
const nextSegment = path[i + 1]!
|
||||
let child = getReplayValue(current, segment)
|
||||
if (!isReplayContainer(child)) {
|
||||
const created = createContainerForNext(nextSegment)
|
||||
setReplayValue(current, segment, created)
|
||||
current = created
|
||||
continue
|
||||
}
|
||||
current = child
|
||||
}
|
||||
return current
|
||||
}
|
||||
|
||||
function applyChatJournalSet(root: unknown, path: ChatJournalPathSegment[], value: unknown): unknown {
|
||||
if (path.length === 0) return value
|
||||
|
||||
const workingRoot = isReplayContainer(root) ? root : createReplayObject()
|
||||
const parent = ensureReplayParent(workingRoot, path)
|
||||
if (!parent) return workingRoot
|
||||
setReplayValue(parent, path[path.length - 1]!, value)
|
||||
return workingRoot
|
||||
}
|
||||
|
||||
function applyChatJournalAppend(root: unknown, path: ChatJournalPathSegment[], items: unknown[]): unknown {
|
||||
const workingRoot = isReplayContainer(root) ? root : createReplayObject()
|
||||
|
||||
if (path.length === 0) {
|
||||
if (Array.isArray(workingRoot)) {
|
||||
for (const item of items) workingRoot.push(item)
|
||||
}
|
||||
return workingRoot
|
||||
}
|
||||
|
||||
const parent = ensureReplayParent(workingRoot, path)
|
||||
if (!parent) return workingRoot
|
||||
|
||||
const last = path[path.length - 1]!
|
||||
let target = getReplayValue(parent, last)
|
||||
const targetArray: unknown[] = Array.isArray(target) ? target : []
|
||||
if (target !== targetArray) {
|
||||
setReplayValue(parent, last, targetArray)
|
||||
}
|
||||
for (const item of items) targetArray.push(item)
|
||||
return workingRoot
|
||||
}
|
||||
|
||||
function replayChatSessionJournal(content: string): unknown {
|
||||
let root: unknown = createReplayObject()
|
||||
const lines = content.split('\n').filter((l) => l.trim())
|
||||
|
||||
for (const line of lines) {
|
||||
let entry: unknown
|
||||
try {
|
||||
entry = JSON.parse(line) as unknown
|
||||
} catch {
|
||||
continue
|
||||
}
|
||||
if (!isRecord(entry)) continue
|
||||
|
||||
const kind = entry['kind']
|
||||
if (kind === 0) {
|
||||
root = entry['v']
|
||||
continue
|
||||
}
|
||||
|
||||
if (kind === 1) {
|
||||
const path = parseChatJournalPath(entry['k'])
|
||||
if (!path) continue
|
||||
root = applyChatJournalSet(root, path, entry['v'])
|
||||
continue
|
||||
}
|
||||
|
||||
if (kind === 2) {
|
||||
const hasPath = Object.prototype.hasOwnProperty.call(entry, 'k')
|
||||
const path = parseChatJournalPath(hasPath ? entry['k'] : undefined, ['requests'])
|
||||
const items = Array.isArray(entry['v']) ? entry['v'] : []
|
||||
if (!path) continue
|
||||
root = applyChatJournalAppend(root, path, items)
|
||||
}
|
||||
}
|
||||
|
||||
return root
|
||||
}
|
||||
|
||||
function numberOrZero(raw: unknown): number {
|
||||
return typeof raw === 'number' && Number.isFinite(raw) && raw > 0 ? raw : 0
|
||||
}
|
||||
|
||||
function readString(raw: unknown): string {
|
||||
return typeof raw === 'string' ? raw : ''
|
||||
}
|
||||
|
||||
function modelFromChatSessionRequest(req: ChatSessionRequest, metadata: Record<string, unknown>): string {
|
||||
const resolved = readString(metadata['resolvedModel'])
|
||||
if (resolved) return resolved
|
||||
|
||||
const modelId = readString(req['modelId']).replace(/^copilot\//, '')
|
||||
return modelId || 'unknown'
|
||||
}
|
||||
|
||||
function extractChatSessionTools(metadata: Record<string, unknown>): string[] {
|
||||
const rounds = metadata['toolCallRounds']
|
||||
if (!Array.isArray(rounds)) return []
|
||||
|
||||
const names = new Set<string>()
|
||||
const addName = (raw: unknown): void => {
|
||||
if (typeof raw === 'string' && raw.trim()) names.add(normalizeTool(raw))
|
||||
}
|
||||
const addFromRecord = (record: Record<string, unknown>): void => {
|
||||
addName(record['toolName'])
|
||||
addName(record['name'])
|
||||
addName(record['tool'])
|
||||
}
|
||||
|
||||
for (const round of rounds) {
|
||||
if (!isRecord(round)) continue
|
||||
addFromRecord(round)
|
||||
|
||||
for (const key of ['tools', 'toolCalls', 'toolRequests']) {
|
||||
const entries = round[key]
|
||||
if (!Array.isArray(entries)) continue
|
||||
for (const entry of entries) {
|
||||
if (typeof entry === 'string') {
|
||||
addName(entry)
|
||||
} else if (isRecord(entry)) {
|
||||
addFromRecord(entry)
|
||||
}
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
return [...names]
|
||||
}
|
||||
|
||||
/**
|
||||
* Extract a shell command string from an OTel execute_tool span's
|
||||
* `gen_ai.tool.call.arguments` attribute. The attribute is a JSON-encoded
|
||||
|
|
@ -571,6 +775,71 @@ function createJsonlParser(
|
|||
}
|
||||
}
|
||||
|
||||
function createChatSessionParser(
|
||||
source: SessionSource,
|
||||
seenKeys: Set<string>
|
||||
): SessionParser {
|
||||
return {
|
||||
async *parse(): AsyncGenerator<ParsedProviderCall> {
|
||||
const content = await readSessionFile(source.path)
|
||||
if (!content) return
|
||||
|
||||
const root = replayChatSessionJournal(content)
|
||||
if (!isRecord(root)) return
|
||||
|
||||
const sessionId = readString(root['sessionId']) || basename(source.path, '.jsonl')
|
||||
const sessionCreatedAt = timestampToISO(root['creationDate'])
|
||||
const requests = Array.isArray(root['requests']) ? root['requests'] : []
|
||||
|
||||
for (let index = 0; index < requests.length; index++) {
|
||||
const rawReq = requests[index]
|
||||
if (!isRecord(rawReq)) continue
|
||||
|
||||
const result = rawReq['result']
|
||||
const resultRecord = isRecord(result) ? result : null
|
||||
const rawMetadata = resultRecord?.['metadata']
|
||||
const metadata = isRecord(rawMetadata) ? rawMetadata : createReplayObject()
|
||||
|
||||
const inputTokens = numberOrZero(metadata['promptTokens'])
|
||||
const metadataOutputTokens = numberOrZero(metadata['outputTokens'])
|
||||
const outputTokens = metadataOutputTokens || numberOrZero(rawReq['completionTokens'])
|
||||
|
||||
if (inputTokens === 0 && outputTokens === 0) continue
|
||||
|
||||
const requestId = readString(rawReq['requestId']) || `request-${index}`
|
||||
const dedupKey = `copilot-chatsession:${sessionId}:${requestId}`
|
||||
if (seenKeys.has(dedupKey)) continue
|
||||
seenKeys.add(dedupKey)
|
||||
|
||||
const model = modelFromChatSessionRequest(rawReq, metadata)
|
||||
const costUSD = calculateCost(model, inputTokens, outputTokens, 0, 0, 0)
|
||||
const timestamp = timestampToISO(rawReq['timestamp']) || sessionCreatedAt
|
||||
|
||||
yield {
|
||||
provider: 'copilot',
|
||||
sessionId,
|
||||
project: source.project,
|
||||
model,
|
||||
inputTokens,
|
||||
outputTokens,
|
||||
cacheCreationInputTokens: 0,
|
||||
cacheReadInputTokens: 0,
|
||||
cachedInputTokens: 0,
|
||||
reasoningTokens: 0,
|
||||
webSearchRequests: 0,
|
||||
costUSD,
|
||||
tools: extractChatSessionTools(metadata),
|
||||
bashCommands: [],
|
||||
timestamp,
|
||||
speed: 'standard' as const,
|
||||
deduplicationKey: dedupKey,
|
||||
userMessage: '',
|
||||
}
|
||||
}
|
||||
},
|
||||
}
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// OTel SQLite parser — reads agent-traces.db for FULL token data
|
||||
// ---------------------------------------------------------------------------
|
||||
|
|
@ -820,10 +1089,18 @@ interface JsonlSessionSource extends SessionSource {
|
|||
sourceType: 'jsonl'
|
||||
}
|
||||
|
||||
interface ChatSessionSource extends SessionSource {
|
||||
sourceType: 'chatsession'
|
||||
}
|
||||
|
||||
function isOtelSource(source: SessionSource): source is OTelSessionSource {
|
||||
return (source as OTelSessionSource).sourceType === 'otel'
|
||||
}
|
||||
|
||||
function isChatSessionSource(source: SessionSource): source is ChatSessionSource {
|
||||
return (source as ChatSessionSource).sourceType === 'chatsession'
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Session discovery: JSONL (original)
|
||||
// ---------------------------------------------------------------------------
|
||||
|
|
@ -921,6 +1198,140 @@ export function getVSCodeWorkspaceStorageDirs(home: string, os: string): string[
|
|||
]
|
||||
}
|
||||
|
||||
export function getVSCodeGlobalStorageDirs(home: string, os: string): string[] {
|
||||
const j = os === 'win32' ? win32.join : posix.join
|
||||
if (os === 'darwin') {
|
||||
return [
|
||||
j(home, 'Library', 'Application Support', 'Code', 'User', 'globalStorage'),
|
||||
j(home, 'Library', 'Application Support', 'Code - Insiders', 'User', 'globalStorage'),
|
||||
j(home, 'Library', 'Application Support', 'VSCodium', 'User', 'globalStorage'),
|
||||
]
|
||||
}
|
||||
if (os === 'linux') {
|
||||
return [
|
||||
j(home, '.config', 'Code', 'User', 'globalStorage'),
|
||||
j(home, '.config', 'Code - Insiders', 'User', 'globalStorage'),
|
||||
j(home, '.config', 'VSCodium', 'User', 'globalStorage'),
|
||||
]
|
||||
}
|
||||
return [
|
||||
j(home, 'AppData', 'Roaming', 'Code', 'User', 'globalStorage'),
|
||||
j(home, 'AppData', 'Roaming', 'Code - Insiders', 'User', 'globalStorage'),
|
||||
j(home, 'AppData', 'Roaming', 'VSCodium', 'User', 'globalStorage'),
|
||||
]
|
||||
}
|
||||
|
||||
async function resolveWorkspaceProject(wsDir: string, hashDir: string): Promise<string> {
|
||||
let project = hashDir
|
||||
try {
|
||||
const wsJson = await readSessionFile(join(wsDir, hashDir, 'workspace.json'))
|
||||
if (wsJson) {
|
||||
const data = JSON.parse(wsJson) as { folder?: string }
|
||||
if (typeof data.folder === 'string') {
|
||||
// folder is a URI like 'file:///home/user/myapp' or 'file:///C:/Users/...'
|
||||
const folder = data.folder.replace(/^file:\/\//, '').replace(/\/+$/, '')
|
||||
const name = basename(folder)
|
||||
if (name) project = name
|
||||
}
|
||||
}
|
||||
} catch {
|
||||
// workspace.json may be absent or malformed
|
||||
}
|
||||
return project
|
||||
}
|
||||
|
||||
async function hasChatSessionFiles(chatSessionsDir: string): Promise<boolean> {
|
||||
let files: string[]
|
||||
try {
|
||||
files = await readdir(chatSessionsDir)
|
||||
} catch {
|
||||
return false
|
||||
}
|
||||
|
||||
for (const file of files) {
|
||||
if (!file.endsWith('.jsonl')) continue
|
||||
const s = await stat(join(chatSessionsDir, file)).catch(() => null)
|
||||
if (s?.isFile()) return true
|
||||
}
|
||||
return false
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Session discovery: VS Code core chatSessions
|
||||
// ---------------------------------------------------------------------------
|
||||
|
||||
async function discoverWorkspaceChatSessions(
|
||||
workspaceStorageDirs: string[]
|
||||
): Promise<ChatSessionSource[]> {
|
||||
const sources: ChatSessionSource[] = []
|
||||
|
||||
for (const wsDir of workspaceStorageDirs) {
|
||||
let hashDirs: string[]
|
||||
try {
|
||||
hashDirs = await readdir(wsDir)
|
||||
} catch {
|
||||
continue
|
||||
}
|
||||
|
||||
for (const hashDir of hashDirs) {
|
||||
const chatSessionsDir = join(wsDir, hashDir, 'chatSessions')
|
||||
let files: string[]
|
||||
try {
|
||||
files = await readdir(chatSessionsDir)
|
||||
} catch {
|
||||
continue
|
||||
}
|
||||
|
||||
const project = await resolveWorkspaceProject(wsDir, hashDir)
|
||||
for (const file of files) {
|
||||
if (!file.endsWith('.jsonl')) continue
|
||||
const path = join(chatSessionsDir, file)
|
||||
const s = await stat(path).catch(() => null)
|
||||
if (!s?.isFile()) continue
|
||||
sources.push({
|
||||
path,
|
||||
project,
|
||||
provider: 'copilot',
|
||||
sourceType: 'chatsession',
|
||||
})
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
return sources
|
||||
}
|
||||
|
||||
async function discoverEmptyWindowChatSessions(
|
||||
globalStorageDirs: string[]
|
||||
): Promise<ChatSessionSource[]> {
|
||||
const sources: ChatSessionSource[] = []
|
||||
|
||||
for (const globalDir of globalStorageDirs) {
|
||||
const chatSessionsDir = join(globalDir, 'emptyWindowChatSessions')
|
||||
let files: string[]
|
||||
try {
|
||||
files = await readdir(chatSessionsDir)
|
||||
} catch {
|
||||
continue
|
||||
}
|
||||
|
||||
for (const file of files) {
|
||||
if (!file.endsWith('.jsonl')) continue
|
||||
const path = join(chatSessionsDir, file)
|
||||
const s = await stat(path).catch(() => null)
|
||||
if (!s?.isFile()) continue
|
||||
sources.push({
|
||||
path,
|
||||
project: 'copilot-chat',
|
||||
provider: 'copilot',
|
||||
sourceType: 'chatsession',
|
||||
})
|
||||
}
|
||||
}
|
||||
|
||||
return sources
|
||||
}
|
||||
|
||||
// ---------------------------------------------------------------------------
|
||||
// Session discovery: VS Code workspace transcripts
|
||||
// ---------------------------------------------------------------------------
|
||||
|
|
@ -944,24 +1355,11 @@ async function discoverTranscriptSessions(
|
|||
}
|
||||
|
||||
for (const hashDir of hashDirs) {
|
||||
const transcriptsDir = join(wsDir, hashDir, 'GitHub.copilot-chat', 'transcripts')
|
||||
const chatSessionsDir = join(wsDir, hashDir, 'chatSessions')
|
||||
if (await hasChatSessionFiles(chatSessionsDir)) continue
|
||||
|
||||
// Resolve project name from workspace.json
|
||||
let project = hashDir
|
||||
try {
|
||||
const wsJson = await readSessionFile(join(wsDir, hashDir, 'workspace.json'))
|
||||
if (wsJson) {
|
||||
const data = JSON.parse(wsJson) as { folder?: string }
|
||||
if (typeof data.folder === 'string') {
|
||||
// folder is a URI like 'file:///home/user/myapp' or 'file:///C:/Users/...'
|
||||
const folder = data.folder.replace(/^file:\/\//, '').replace(/\/+$/, '')
|
||||
const name = basename(folder)
|
||||
if (name) project = name
|
||||
}
|
||||
}
|
||||
} catch {
|
||||
// workspace.json may be absent or malformed
|
||||
}
|
||||
const transcriptsDir = join(wsDir, hashDir, 'GitHub.copilot-chat', 'transcripts')
|
||||
const project = await resolveWorkspaceProject(wsDir, hashDir)
|
||||
|
||||
let transcriptFiles: string[]
|
||||
try {
|
||||
|
|
@ -987,7 +1385,11 @@ async function discoverTranscriptSessions(
|
|||
return sources
|
||||
}
|
||||
|
||||
export function createCopilotProvider(sessionStateDir?: string, workspaceStorageDir?: string): Provider {
|
||||
export function createCopilotProvider(
|
||||
sessionStateDir?: string,
|
||||
workspaceStorageDir?: string,
|
||||
globalStorageDir?: string
|
||||
): Provider {
|
||||
// jsonlDir is resolved lazily inside discoverSessions so that env-var
|
||||
// overrides set after module load (e.g. in tests) are respected.
|
||||
|
||||
|
|
@ -1005,6 +1407,13 @@ export function createCopilotProvider(sessionStateDir?: string, workspaceStorage
|
|||
return getVSCodeWorkspaceStorageDirs(homedir(), platform())
|
||||
}
|
||||
|
||||
function getGlobalDirs(): string[] {
|
||||
if (globalStorageDir !== undefined) return [globalStorageDir]
|
||||
const envDir = process.env['CODEBURN_COPILOT_GLOBAL_STORAGE_DIR']
|
||||
if (envDir) return [envDir]
|
||||
return getVSCodeGlobalStorageDirs(homedir(), platform())
|
||||
}
|
||||
|
||||
return {
|
||||
name: 'copilot',
|
||||
displayName: 'Copilot',
|
||||
|
|
@ -1023,6 +1432,7 @@ export function createCopilotProvider(sessionStateDir?: string, workspaceStorage
|
|||
|
||||
async discoverSessions(): Promise<SessionSource[]> {
|
||||
const sources: SessionSource[] = []
|
||||
let discoveredOtel = false
|
||||
|
||||
// 1. Discover OTel sessions (preferred — full token data)
|
||||
const disableOtel = process.env['CODEBURN_COPILOT_DISABLE_OTEL'] === '1'
|
||||
|
|
@ -1031,6 +1441,7 @@ export function createCopilotProvider(sessionStateDir?: string, workspaceStorage
|
|||
if (dbPath) {
|
||||
try {
|
||||
const otelSources = await discoverOtelSessions(dbPath)
|
||||
discoveredOtel = otelSources.length > 0
|
||||
sources.push(...otelSources)
|
||||
} catch {
|
||||
// OTel discovery failed — fall through to JSONL
|
||||
|
|
@ -1047,7 +1458,27 @@ export function createCopilotProvider(sessionStateDir?: string, workspaceStorage
|
|||
// JSONL discovery failed
|
||||
}
|
||||
|
||||
// 3. Discover VS Code workspace transcript sessions
|
||||
// Prefer OTel over chatSessions: they can mirror the same turns under
|
||||
// incompatible IDs, and OTel carries richer token/cache data.
|
||||
if (!discoveredOtel) {
|
||||
// 3. Discover VS Code core chatSessions journals
|
||||
try {
|
||||
const chatSessionSources = await discoverWorkspaceChatSessions(getWsDirs())
|
||||
sources.push(...chatSessionSources)
|
||||
} catch {
|
||||
// Workspace chatSessions discovery failed
|
||||
}
|
||||
|
||||
// 4. Discover VS Code empty-window chatSessions journals
|
||||
try {
|
||||
const emptyWindowSources = await discoverEmptyWindowChatSessions(getGlobalDirs())
|
||||
sources.push(...emptyWindowSources)
|
||||
} catch {
|
||||
// Empty-window chatSessions discovery failed
|
||||
}
|
||||
}
|
||||
|
||||
// 5. Discover VS Code workspace transcript sessions
|
||||
try {
|
||||
const transcriptSources = await discoverTranscriptSessions(getWsDirs())
|
||||
sources.push(...transcriptSources)
|
||||
|
|
@ -1069,6 +1500,9 @@ export function createCopilotProvider(sessionStateDir?: string, workspaceStorage
|
|||
if (isOtelSource(source)) {
|
||||
return createOtelParser(source, seenKeys)
|
||||
}
|
||||
if (isChatSessionSource(source)) {
|
||||
return createChatSessionParser(source, seenKeys)
|
||||
}
|
||||
return createJsonlParser(source, seenKeys)
|
||||
},
|
||||
}
|
||||
|
|
|
|||
|
|
@ -4,7 +4,7 @@ import { join, posix, win32 } from 'path'
|
|||
import { tmpdir } from 'os'
|
||||
import { createRequire } from 'node:module'
|
||||
|
||||
import { copilot, createCopilotProvider, getVSCodeWorkspaceStorageDirs } from '../../src/providers/copilot.js'
|
||||
import { copilot, createCopilotProvider, getVSCodeGlobalStorageDirs, getVSCodeWorkspaceStorageDirs } from '../../src/providers/copilot.js'
|
||||
import { isSqliteAvailable } from '../../src/sqlite.js'
|
||||
import type { ParsedProviderCall } from '../../src/providers/types.js'
|
||||
|
||||
|
|
@ -63,6 +63,34 @@ function transcriptAssistantMessage(opts: { messageId: string; content?: string;
|
|||
})
|
||||
}
|
||||
|
||||
function chatSessionSampleRequest(overrides: Record<string, unknown> = {}) {
|
||||
return {
|
||||
requestId: 'request_8c8ce017-6e3f-460a-9931-5a16825d231a',
|
||||
modelId: 'copilot/claude-sonnet-4.6',
|
||||
completionTokens: 490,
|
||||
result: {
|
||||
metadata: {
|
||||
promptTokens: 32543,
|
||||
outputTokens: 60,
|
||||
resolvedModel: 'claude-sonnet-4-6',
|
||||
toolCallRounds: [{ thinking: { tokens: 0 }, modelId: 'claude-sonnet-4.6' }],
|
||||
agentId: 'github.copilot.editsAgent',
|
||||
},
|
||||
},
|
||||
...overrides,
|
||||
}
|
||||
}
|
||||
|
||||
async function createChatSessionFile(filePath: string, entries: unknown[]) {
|
||||
await writeFile(filePath, entries.map(entry => JSON.stringify(entry)).join('\n') + '\n')
|
||||
}
|
||||
|
||||
async function collectCalls(source: { path: string; project: string; provider: string; sourceType?: string }, seenKeys = new Set<string>()) {
|
||||
const calls: ParsedProviderCall[] = []
|
||||
for await (const call of copilot.createSessionParser(source, seenKeys).parse()) calls.push(call)
|
||||
return calls
|
||||
}
|
||||
|
||||
describe('copilot provider - JSONL parsing', () => {
|
||||
beforeEach(async () => {
|
||||
tmpDir = await mkdtemp(join(tmpdir(), 'copilot-test-'))
|
||||
|
|
@ -362,6 +390,181 @@ describe('copilot provider - JSONL parsing', () => {
|
|||
})
|
||||
})
|
||||
|
||||
describe('copilot provider - chatSessions parsing', () => {
|
||||
beforeEach(async () => {
|
||||
tmpDir = await mkdtemp(join(tmpdir(), 'copilot-chatsessions-test-'))
|
||||
vi.stubEnv('CODEBURN_COPILOT_DISABLE_OTEL', '1')
|
||||
})
|
||||
|
||||
afterEach(async () => {
|
||||
await rm(tmpDir, { recursive: true, force: true })
|
||||
vi.unstubAllEnvs()
|
||||
})
|
||||
|
||||
it('parses sample journal token counts and cost', async () => {
|
||||
const filePath = join(tmpDir, 'sample.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-session-1', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
|
||||
const calls = await collectCalls({ path: filePath, project: 'myproject', provider: 'copilot', sourceType: 'chatsession' })
|
||||
|
||||
expect(calls).toHaveLength(1)
|
||||
expect(calls[0]!.inputTokens).toBe(32543)
|
||||
expect(calls[0]!.outputTokens).toBe(60)
|
||||
expect(calls[0]!.model).toBe('claude-sonnet-4-6')
|
||||
expect(calls[0]!.costUSD).toBeGreaterThan(0)
|
||||
})
|
||||
|
||||
it('returns no calls for an empty reconstructed requests array', async () => {
|
||||
const filePath = join(tmpDir, 'empty.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-empty', requests: [] } },
|
||||
])
|
||||
|
||||
const calls = await collectCalls({ path: filePath, project: 'myproject', provider: 'copilot', sourceType: 'chatsession' })
|
||||
|
||||
expect(calls).toHaveLength(0)
|
||||
})
|
||||
|
||||
it('discovers and parses emptyWindowChatSessions from globalStorage', async () => {
|
||||
const globalDir = join(tmpDir, 'globalStorage')
|
||||
const emptyWindowDir = join(globalDir, 'emptyWindowChatSessions')
|
||||
await mkdir(emptyWindowDir, { recursive: true })
|
||||
const filePath = join(emptyWindowDir, 'empty-window.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'empty-window-session', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
|
||||
const provider = createCopilotProvider('/nonexistent/legacy', '/nonexistent/ws', globalDir)
|
||||
const sessions = await provider.discoverSessions()
|
||||
|
||||
expect(sessions).toHaveLength(1)
|
||||
expect(sessions[0]!.project).toBe('copilot-chat')
|
||||
expect((sessions[0] as { sourceType?: string }).sourceType).toBe('chatsession')
|
||||
|
||||
const calls: ParsedProviderCall[] = []
|
||||
for await (const call of provider.createSessionParser(sessions[0]!, new Set()).parse()) calls.push(call)
|
||||
expect(calls).toHaveLength(1)
|
||||
expect(calls[0]!.inputTokens).toBe(32543)
|
||||
})
|
||||
|
||||
it('skips chatSessions discovery when an OTel source is present', async () => {
|
||||
if (!isSqliteAvailable()) return
|
||||
|
||||
vi.unstubAllEnvs()
|
||||
const dbPath = join(tmpDir, 'agent-traces.db')
|
||||
vi.stubEnv('CODEBURN_COPILOT_OTEL_DB', dbPath)
|
||||
vi.stubEnv('CODEBURN_COPILOT_DISABLE_OTEL', '')
|
||||
createOtelDb(dbPath)
|
||||
insertSpan(dbPath, {
|
||||
spanId: 'span-chatsession-skip',
|
||||
traceId: 'trace-chatsession-skip',
|
||||
operationName: 'chat',
|
||||
startTimeMs: 1000,
|
||||
attrs: {
|
||||
'gen_ai.conversation.id': 'conv-chatsession-skip',
|
||||
'gen_ai.response.model': 'gpt-4.1',
|
||||
'gen_ai.usage.input_tokens': 100,
|
||||
'gen_ai.usage.output_tokens': 10,
|
||||
},
|
||||
})
|
||||
|
||||
const wsDir = join(tmpDir, 'vscode-ws')
|
||||
const hashDir = join(wsDir, 'abc123')
|
||||
const workspaceChatSessionsDir = join(hashDir, 'chatSessions')
|
||||
const globalDir = join(tmpDir, 'globalStorage')
|
||||
const emptyWindowDir = join(globalDir, 'emptyWindowChatSessions')
|
||||
await mkdir(workspaceChatSessionsDir, { recursive: true })
|
||||
await mkdir(emptyWindowDir, { recursive: true })
|
||||
await writeFile(join(hashDir, 'workspace.json'), JSON.stringify({ folder: 'file:///home/user/myapp' }))
|
||||
await createChatSessionFile(join(workspaceChatSessionsDir, 'workspace.jsonl'), [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-workspace', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
await createChatSessionFile(join(emptyWindowDir, 'empty-window.jsonl'), [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-empty-window', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest({ requestId: 'request-empty-window' })] },
|
||||
])
|
||||
|
||||
const provider = createCopilotProvider('/nonexistent/legacy', wsDir, globalDir)
|
||||
const sources = await provider.discoverSessions()
|
||||
|
||||
expect(sources.filter(s => (s as { sourceType?: string }).sourceType === 'otel')).toHaveLength(1)
|
||||
expect(sources.filter(s => (s as { sourceType?: string }).sourceType === 'chatsession')).toHaveLength(0)
|
||||
})
|
||||
|
||||
it('applies append-then-edit journal updates', async () => {
|
||||
const filePath = join(tmpDir, 'append-edit.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-edit', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
{ kind: 1, k: ['requests', 0, 'result', 'metadata', 'outputTokens'], v: 88 },
|
||||
])
|
||||
|
||||
const calls = await collectCalls({ path: filePath, project: 'myproject', provider: 'copilot', sourceType: 'chatsession' })
|
||||
|
||||
expect(calls).toHaveLength(1)
|
||||
expect(calls[0]!.outputTokens).toBe(88)
|
||||
})
|
||||
|
||||
it('deduplicates by requestId across parser runs', async () => {
|
||||
const filePath = join(tmpDir, 'dedupe.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-dedupe', requests: [] } },
|
||||
{ kind: 2, v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
const source = { path: filePath, project: 'myproject', provider: 'copilot', sourceType: 'chatsession' }
|
||||
const seenKeys = new Set<string>()
|
||||
|
||||
const calls1 = await collectCalls(source, seenKeys)
|
||||
const calls2 = await collectCalls(source, seenKeys)
|
||||
|
||||
expect(calls1).toHaveLength(1)
|
||||
expect(calls2).toHaveLength(0)
|
||||
})
|
||||
|
||||
it('ignores prototype-pollution journal paths without crashing', async () => {
|
||||
const filePath = join(tmpDir, 'proto.jsonl')
|
||||
await createChatSessionFile(filePath, [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-proto', requests: [] } },
|
||||
{ kind: 1, k: ['__proto__', 'polluted'], v: true },
|
||||
{ kind: 1, k: ['constructor', 'prototype', 'polluted'], v: true },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
|
||||
expect(({} as { polluted?: unknown }).polluted).toBeUndefined()
|
||||
const calls = await collectCalls({ path: filePath, project: 'myproject', provider: 'copilot', sourceType: 'chatsession' })
|
||||
|
||||
expect(calls).toHaveLength(1)
|
||||
expect(({} as { polluted?: unknown }).polluted).toBeUndefined()
|
||||
})
|
||||
|
||||
it('skips legacy transcripts for a workspace hash that has chatSessions', async () => {
|
||||
const wsDir = join(tmpDir, 'vscode-ws')
|
||||
const hashDir = join(wsDir, 'abc123')
|
||||
const chatSessionsDir = join(hashDir, 'chatSessions')
|
||||
const transcriptsDir = join(hashDir, 'GitHub.copilot-chat', 'transcripts')
|
||||
await mkdir(chatSessionsDir, { recursive: true })
|
||||
await mkdir(transcriptsDir, { recursive: true })
|
||||
await writeFile(join(hashDir, 'workspace.json'), JSON.stringify({ folder: 'file:///home/user/myapp' }))
|
||||
await createChatSessionFile(join(chatSessionsDir, 'chat.jsonl'), [
|
||||
{ kind: 0, v: { version: 3, creationDate: 1780157113020, sessionId: 'chat-modern', requests: [] } },
|
||||
{ kind: 2, k: ['requests'], v: [chatSessionSampleRequest()] },
|
||||
])
|
||||
await writeFile(join(transcriptsDir, 'legacy.jsonl'), transcriptSessionStart('legacy') + '\n')
|
||||
|
||||
const provider = createCopilotProvider('/nonexistent/legacy', wsDir, '/nonexistent/global')
|
||||
const sessions = await provider.discoverSessions()
|
||||
|
||||
expect(sessions).toHaveLength(1)
|
||||
expect((sessions[0] as { sourceType?: string }).sourceType).toBe('chatsession')
|
||||
expect(sessions[0]!.path).toContain(`${join('abc123', 'chatSessions')}`)
|
||||
})
|
||||
})
|
||||
|
||||
describe('copilot provider - discoverSessions', () => {
|
||||
beforeEach(async () => {
|
||||
tmpDir = await mkdtemp(join(tmpdir(), 'copilot-test-'))
|
||||
|
|
@ -450,6 +653,18 @@ describe('copilot provider - discoverSessions', () => {
|
|||
posix.join('/home/test', '.config', 'VSCodium', 'User', 'workspaceStorage'),
|
||||
)
|
||||
})
|
||||
|
||||
it('includes VSCodium globalStorage paths on all supported platforms', () => {
|
||||
expect(getVSCodeGlobalStorageDirs('/Users/test', 'darwin')).toContain(
|
||||
posix.join('/Users/test', 'Library', 'Application Support', 'VSCodium', 'User', 'globalStorage'),
|
||||
)
|
||||
expect(getVSCodeGlobalStorageDirs('C:\\Users\\test', 'win32')).toContain(
|
||||
win32.join('C:\\Users\\test', 'AppData', 'Roaming', 'VSCodium', 'User', 'globalStorage'),
|
||||
)
|
||||
expect(getVSCodeGlobalStorageDirs('/home/test', 'linux')).toContain(
|
||||
posix.join('/home/test', '.config', 'VSCodium', 'User', 'globalStorage'),
|
||||
)
|
||||
})
|
||||
})
|
||||
|
||||
describe('copilot provider - metadata', () => {
|
||||
|
|
|
|||
Loading…
Add table
Add a link
Reference in a new issue