@axplusb/kepler 0.0.1 → 1.0.0

package/pulse/lib/claude-reader.ts

@@ -0,0 +1,594 @@
+import fs from 'fs/promises'
+import path from 'path'
+import os from 'os'
+import type {
+  StatsCache,
+  SessionMeta,
+  Facet,
+  HistoryEntry,
+} from '@/types/claude'
+import { slugToPath } from '@/lib/decode'
+
+function stripXmlTags(text: string): string {
+  return text.replace(/<[^>]+>[\s\S]*?<\/[^>]+>/g, '').replace(/<[^>]+\/>/g, '').replace(/<[^>]+>/g, '').trim()
+}
+
+/** Resolve the real filesystem path for a project slug by reading `cwd` from its JSONL files */
+export async function resolveProjectPath(slug: string): Promise<string> {
+  const files = await listProjectJSONLFiles(slug)
+  for (const f of files) {
+    try {
+      const raw = await fs.readFile(f, 'utf-8')
+      const lines = raw.split(/\r?\n/)
+      for (const line of lines.slice(0, 50)) {
+        if (!line.trim()) continue
+        try {
+          const obj = JSON.parse(line)
+          if (obj.cwd && typeof obj.cwd === 'string') return obj.cwd
+        } catch { /* skip malformed line */ }
+      }
+    } catch { /* try next file */ }
+  }
+  return slugToPath(slug)
+}
+
+const CLAUDE_DIR = process.env.ORCA_CONFIG_DIR ?? process.env.CLAUDE_CONFIG_DIR ?? path.join(os.homedir(), '.orca')
+
+export function claudePath(...segments: string[]): string {
+  return path.join(CLAUDE_DIR, ...segments)
+}
+
+// ─── Stats Cache ─────────────────────────────────────────────────────────────
+
+export async function readStatsCache(): Promise<StatsCache | null> {
+  try {
+    const raw = await fs.readFile(claudePath('stats-cache.json'), 'utf-8')
+    return JSON.parse(raw) as StatsCache
+  } catch {
+    return null
+  }
+}
+
+// ─── Sessions from Project JSONL (primary source) ──────────────────────────────
+
+/** Derive session metadata directly from ~/.claude/projects/<project>/<session>.jsonl */
+export async function readSessionsFromProjectJSONL(): Promise<SessionMeta[]> {
+  const results: SessionMeta[] = []
+  try {
+    const slugs = await listProjectSlugs()
+    for (const slug of slugs) {
+      const projectPath = await resolveProjectPath(slug)
+      const files = await listProjectJSONLFiles(slug)
+      for (const filePath of files) {
+        const sessionId = path.basename(filePath, '.jsonl')
+        const meta = await deriveSessionMetaFromJSONL(filePath, sessionId, projectPath)
+        if (meta) results.push(meta)
+      }
+    }
+    return results.sort(
+      (a, b) => new Date(b.start_time).getTime() - new Date(a.start_time).getTime()
+    )
+  } catch {
+    return []
+  }
+}
+
+async function deriveSessionMetaFromJSONL(
+  filePath: string,
+  sessionId: string,
+  projectPath: string
+): Promise<SessionMeta | null> {
+  let startTime = ''
+  let lastTime = ''
+  let userCount = 0
+  let assistantCount = 0
+  const toolCounts: Record<string, number> = {}
+  let inputTokens = 0
+  let outputTokens = 0
+  let cacheRead = 0
+  let cacheWrite = 0
+  let firstPrompt = ''
+  let hasTaskAgent = false
+  let hasMcp = false
+  let hasWebSearch = false
+  let hasWebFetch = false
+  const messageHours: number[] = []
+  const userMessageTimestamps: string[] = []
+
+  try {
+    const raw = await fs.readFile(filePath, 'utf-8')
+    const lines = raw.split(/\r?\n/).filter(Boolean)
+    for (const line of lines) {
+      try {
+        const obj = JSON.parse(line) as Record<string, unknown>
+        const ts = obj.timestamp as string
+        if (ts) {
+          if (!startTime) startTime = ts
+          lastTime = ts
+        }
+        if (obj.type === 'user') {
+          userCount++
+          if (ts) {
+            const d = new Date(ts)
+            if (!isNaN(d.getTime())) {
+              messageHours.push(d.getHours())
+              userMessageTimestamps.push(ts)
+            }
+          }
+          const content = (obj as { message?: { content?: string | unknown[] } }).message?.content
+          if (typeof content === 'string' && !firstPrompt) firstPrompt = stripXmlTags(content).slice(0, 500)
+          else if (Array.isArray(content)) {
+            const text = content.find((c: unknown) => typeof c === 'object' && c !== null && (c as { type?: string }).type === 'text')
+            if (text && typeof (text as { text?: string }).text === 'string' && !firstPrompt) {
+              firstPrompt = stripXmlTags((text as { text: string }).text).slice(0, 500)
+            }
+          }
+        }
+        if (obj.type === 'assistant') {
+          assistantCount++
+          const msg = (obj as { message?: { usage?: Record<string, number>; content?: unknown[] } }).message
+          if (msg?.usage) {
+            inputTokens += msg.usage.input_tokens ?? 0
+            outputTokens += msg.usage.output_tokens ?? 0
+            cacheRead += msg.usage.cache_read_input_tokens ?? 0
+            cacheWrite += msg.usage.cache_creation_input_tokens ?? 0
+          }
+          const content = msg?.content
+          if (Array.isArray(content)) {
+            for (const c of content) {
+              const item = c as { type?: string; name?: string }
+              if (item.type === 'tool_use' && item.name) {
+                toolCounts[item.name] = (toolCounts[item.name] ?? 0) + 1
+                if (item.name.startsWith('Task') || item.name === 'TodoWrite' || item.name === 'Agent') hasTaskAgent = true
+                if (item.name.startsWith('mcp__')) hasMcp = true
+                if (item.name === 'WebSearch') hasWebSearch = true
+                if (item.name === 'WebFetch') hasWebFetch = true
+              }
+            }
+          }
+        }
+      } catch { /* skip malformed line */ }
+    }
+  } catch {
+    return null
+  }
+
+  if (!startTime) return null
+
+  const start = new Date(startTime).getTime()
+  const end = lastTime ? new Date(lastTime).getTime() : start
+  const durationMinutes = (end - start) / 60_000
+
+  return {
+    session_id: sessionId,
+    project_path: projectPath,
+    start_time: startTime,
+    duration_minutes: durationMinutes,
+    user_message_count: userCount,
+    assistant_message_count: assistantCount,
+    tool_counts: toolCounts,
+    languages: {},
+    git_commits: 0,
+    git_pushes: 0,
+    input_tokens: inputTokens,
+    output_tokens: outputTokens,
+    cache_creation_input_tokens: cacheWrite,
+    cache_read_input_tokens: cacheRead,
+    first_prompt: firstPrompt,
+    user_interruptions: 0,
+    user_response_times: [],
+    tool_errors: 0,
+    tool_error_categories: {},
+    uses_task_agent: hasTaskAgent,
+    uses_mcp: hasMcp,
+    uses_web_search: hasWebSearch,
+    uses_web_fetch: hasWebFetch,
+    lines_added: 0,
+    lines_removed: 0,
+    files_modified: 0,
+    message_hours: messageHours,
+    user_message_timestamps: userMessageTimestamps,
+  }
+}
+
+/** Get sessions: prefers JSONL (projects/*.jsonl), falls back to usage-data/session-meta */
+export async function getSessions(): Promise<SessionMeta[]> {
+  const [jsonl, meta] = await Promise.all([
+    readSessionsFromProjectJSONL(),
+    readAllSessionMeta(),
+  ])
+  if (jsonl.length > 0) return jsonl
+  return meta
+}
+
+// ─── Session Meta (usage-data/session-meta — fallback) ────────────────────────
+
+export async function readAllSessionMeta(): Promise<SessionMeta[]> {
+  const dir = claudePath('usage-data', 'session-meta')
+  try {
+    const files = await fs.readdir(dir)
+    const results: SessionMeta[] = []
+    await Promise.all(
+      files
+        .filter(f => f.endsWith('.json'))
+        .map(async f => {
+          try {
+            const raw = await fs.readFile(path.join(dir, f), 'utf-8')
+            const parsed = JSON.parse(raw) as SessionMeta
+            results.push(parsed)
+          } catch { /* skip malformed */ }
+        })
+    )
+    return results.sort(
+      (a, b) => new Date(b.start_time).getTime() - new Date(a.start_time).getTime()
+    )
+  } catch {
+    return []
+  }
+}
+
+export async function readSessionMeta(sessionId: string): Promise<SessionMeta | null> {
+  try {
+    const raw = await fs.readFile(
+      claudePath('usage-data', 'session-meta', `${sessionId}.json`),
+      'utf-8'
+    )
+    return JSON.parse(raw) as SessionMeta
+  } catch {
+    return null
+  }
+}
+
+// ─── Facets ──────────────────────────────────────────────────────────────────
+
+export async function readAllFacets(): Promise<Facet[]> {
+  const dir = claudePath('usage-data', 'facets')
+  try {
+    const files = await fs.readdir(dir)
+    const results: Facet[] = []
+    await Promise.all(
+      files
+        .filter(f => f.endsWith('.json'))
+        .map(async f => {
+          try {
+            const raw = await fs.readFile(path.join(dir, f), 'utf-8')
+            results.push(JSON.parse(raw) as Facet)
+          } catch { /* skip */ }
+        })
+    )
+    return results
+  } catch {
+    return []
+  }
+}
+
+export async function readFacet(sessionId: string): Promise<Facet | null> {
+  try {
+    const raw = await fs.readFile(
+      claudePath('usage-data', 'facets', `${sessionId}.json`),
+      'utf-8'
+    )
+    return JSON.parse(raw) as Facet
+  } catch {
+    return null
+  }
+}
+
+// ─── Projects ─────────────────────────────────────────────────────────────────
+
+export async function listProjectSlugs(): Promise<string[]> {
+  try {
+    const entries = await fs.readdir(claudePath('projects'), { withFileTypes: true })
+    return entries
+      .filter(e => e.isDirectory())
+      .map(e => e.name)
+  } catch {
+    return []
+  }
+}
+
+export async function listProjectJSONLFiles(slug: string): Promise<string[]> {
+  try {
+    const dir = claudePath('projects', slug)
+    const files = await fs.readdir(dir)
+    return files
+      .filter(f => f.endsWith('.jsonl'))
+      .map(f => path.join(dir, f))
+  } catch {
+    return []
+  }
+}
+
+/** Stream a JSONL file line by line, calling cb for each parsed line */
+export async function readJSONLLines(
+  filePath: string,
+  cb: (line: Record<string, unknown>) => void
+): Promise<void> {
+  try {
+    const raw = await fs.readFile(filePath, 'utf-8')
+    for (const line of raw.split(/\r?\n/)) {
+      if (!line.trim()) continue
+      try {
+        cb(JSON.parse(line))
+      } catch { /* skip malformed */ }
+    }
+  } catch { /* file missing */ }
+}
+
+/** Find which project slug contains a given session ID */
+export async function findSessionSlug(sessionId: string): Promise<string | null> {
+  const slugs = await listProjectSlugs()
+  for (const slug of slugs) {
+    const files = await listProjectJSONLFiles(slug)
+    for (const f of files) {
+      if (path.basename(f).startsWith(sessionId)) return slug
+    }
+  }
+  return null
+}
+
+/** Find the JSONL file path for a given session ID */
+export async function findSessionJSONL(sessionId: string): Promise<string | null> {
+  const slugs = await listProjectSlugs()
+  for (const slug of slugs) {
+    const files = await listProjectJSONLFiles(slug)
+    for (const f of files) {
+      if (path.basename(f, '.jsonl') === sessionId) return f
+    }
+  }
+  return null
+}
+
+// ─── Plans ───────────────────────────────────────────────────────────────────
+
+export interface PlanFile {
+  path: string
+  name: string
+  content: string
+  mtime: string
+}
+
+export async function readPlans(): Promise<PlanFile[]> {
+  const results: PlanFile[] = []
+  try {
+    const dir = claudePath('plans')
+    const files = await fs.readdir(dir)
+    for (const f of files.filter((x) => x.endsWith('.md'))) {
+      try {
+        const fullPath = path.join(dir, f)
+        const [raw, stat] = await Promise.all([
+          fs.readFile(fullPath, 'utf-8'),
+          fs.stat(fullPath),
+        ])
+        results.push({
+          path: fullPath,
+          name: f.replace(/\.md$/, ''),
+          content: raw,
+          mtime: stat.mtime.toISOString(),
+        })
+      } catch { /* skip */ }
+    }
+    return results.sort((a, b) => b.mtime.localeCompare(a.mtime))
+  } catch {
+    return []
+  }
+}
+
+// ─── Todos ───────────────────────────────────────────────────────────────────
+
+export interface TodoFile {
+  path: string
+  name: string
+  data: unknown
+  mtime: string
+}
+
+export async function readTodos(): Promise<TodoFile[]> {
+  const results: TodoFile[] = []
+  try {
+    const dir = claudePath('todos')
+    const files = await fs.readdir(dir)
+    for (const f of files.filter((x) => x.endsWith('.json'))) {
+      try {
+        const fullPath = path.join(dir, f)
+        const [raw, stat] = await Promise.all([
+          fs.readFile(fullPath, 'utf-8'),
+          fs.stat(fullPath),
+        ])
+        results.push({
+          path: fullPath,
+          name: f.replace(/\.json$/, ''),
+          data: JSON.parse(raw),
+          mtime: stat.mtime.toISOString(),
+        })
+      } catch { /* skip */ }
+    }
+    return results.sort((a, b) => b.mtime.localeCompare(a.mtime))
+  } catch {
+    return []
+  }
+}
+
+// ─── History ─────────────────────────────────────────────────────────────────
+
+export async function readHistory(limit = 200): Promise<HistoryEntry[]> {
+  const entries: HistoryEntry[] = []
+  try {
+    const raw = await fs.readFile(claudePath('history.jsonl'), 'utf-8')
+    const lines = raw.split(/\r?\n/).filter(Boolean)
+    for (const line of lines.slice(-limit)) {
+      try {
+        entries.push(JSON.parse(line) as HistoryEntry)
+      } catch { /* skip */ }
+    }
+  } catch { /* file missing */ }
+  return entries
+}
+
+// ─── Skills ───────────────────────────────────────────────────────────────────
+
+export interface SkillInfo {
+  name: string
+  description: string
+  triggers: string
+  hasSkillMd: boolean
+}
+
+export async function readSkills(): Promise<SkillInfo[]> {
+  const skillsDir = claudePath('skills')
+  try {
+    const entries = await fs.readdir(skillsDir, { withFileTypes: true })
+    const dirs = entries.filter(e => e.isDirectory() && !e.name.startsWith('.') && e.name !== 'nebius-skills-workspace')
+    const results: SkillInfo[] = []
+    for (const dir of dirs) {
+      const skillMdPath = path.join(skillsDir, dir.name, 'SKILL.md')
+      let description = ''
+      let triggers = ''
+      let hasSkillMd = false
+      try {
+        const raw = await fs.readFile(skillMdPath, 'utf-8')
+        hasSkillMd = true
+        const descMatch = raw.match(/^#\s+(.+)$/m)
+        if (descMatch) description = descMatch[1].trim()
+        const triggerMatch = raw.match(/(?:TRIGGER|trigger)[^\n]*\n([\s\S]*?)(?:\n#{1,3}\s|\n---|\n\*\*DO NOT|$)/m)
+        if (triggerMatch) triggers = triggerMatch[1].replace(/\s+/g, ' ').trim().slice(0, 200)
+      } catch { /* no SKILL.md */ }
+      results.push({ name: dir.name, description, triggers, hasSkillMd })
+    }
+    return results.sort((a, b) => a.name.localeCompare(b.name))
+  } catch {
+    return []
+  }
+}
+
+// ─── Plugins ──────────────────────────────────────────────────────────────────
+
+export interface PluginInfo {
+  id: string
+  scope: string
+  version: string
+  installedAt: string
+}
+
+export async function readInstalledPlugins(): Promise<PluginInfo[]> {
+  try {
+    const raw = await fs.readFile(claudePath('plugins', 'installed_plugins.json'), 'utf-8')
+    const json = JSON.parse(raw) as { plugins: Record<string, Array<{ scope: string; version: string; installedAt: string }>> }
+    return Object.entries(json.plugins).flatMap(([id, installs]) =>
+      installs.map(inst => ({ id, scope: inst.scope, version: inst.version, installedAt: inst.installedAt }))
+    )
+  } catch {
+    return []
+  }
+}
+
+// ─── Settings ─────────────────────────────────────────────────────────────────
+
+export async function readSettings(): Promise<Record<string, unknown>> {
+  try {
+    const raw = await fs.readFile(claudePath('settings.json'), 'utf-8')
+    return JSON.parse(raw)
+  } catch {
+    return {}
+  }
+}
+
+// ─── Memory ───────────────────────────────────────────────────────────────────
+
+export type MemoryType = 'user' | 'feedback' | 'project' | 'reference' | 'index' | 'unknown'
+
+export interface MemoryEntry {
+  file: string
+  projectSlug: string
+  projectPath: string
+  name: string
+  type: MemoryType
+  description: string
+  body: string
+  mtime: string
+  isIndex: boolean
+}
+
+function parseFrontmatter(raw: string): { meta: Record<string, string>; body: string } {
+  const match = raw.match(/^---\r?\n([\s\S]*?)\r?\n---\r?\n?([\s\S]*)$/)
+  if (!match) return { meta: {}, body: raw }
+  const meta: Record<string, string> = {}
+  for (const line of match[1].split('\n')) {
+    const colon = line.indexOf(':')
+    if (colon === -1) continue
+    const key = line.slice(0, colon).trim()
+    const val = line.slice(colon + 1).trim()
+    if (key) meta[key] = val
+  }
+  return { meta, body: match[2].trim() }
+}
+
+export async function readMemories(): Promise<MemoryEntry[]> {
+  const results: MemoryEntry[] = []
+  try {
+    const slugs = await listProjectSlugs()
+    await Promise.all(
+      slugs.map(async slug => {
+        const memDir = claudePath('projects', slug, 'memory')
+        try {
+          const files = await fs.readdir(memDir)
+          const mdFiles = files.filter(f => f.endsWith('.md'))
+          await Promise.all(
+            mdFiles.map(async file => {
+              try {
+                const fullPath = path.join(memDir, file)
+                const [raw, stat] = await Promise.all([
+                  fs.readFile(fullPath, 'utf-8'),
+                  fs.stat(fullPath),
+                ])
+                const isIndex = file === 'MEMORY.md'
+                const { meta, body } = parseFrontmatter(raw)
+                const projectPath = slugToPath(slug)
+                const h1Match = body.match(/^#\s+(.+)$/m)
+                const titleFromBody = h1Match ? h1Match[1].trim() : null
+                results.push({
+                  file,
+                  projectSlug: slug,
+                  projectPath,
+                  name: meta.name ?? titleFromBody ?? (isIndex ? 'Memory Index' : file.replace(/\.md$/, '')),
+                  type: (meta.type as MemoryType) ?? (isIndex ? 'index' : 'unknown'),
+                  description: meta.description ?? '',
+                  body,
+                  mtime: stat.mtime.toISOString(),
+                  isIndex,
+                })
+              } catch { /* skip */ }
+            })
+          )
+        } catch { /* no memory dir */ }
+      })
+    )
+  } catch { /* skip */ }
+  return results.sort((a, b) => b.mtime.localeCompare(a.mtime))
+}
+
+// ─── Storage size ─────────────────────────────────────────────────────────────
+
+export async function getClaudeStorageBytes(): Promise<number> {
+  async function dirSize(dirPath: string): Promise<number> {
+    let total = 0
+    try {
+      const entries = await fs.readdir(dirPath, { withFileTypes: true })
+      await Promise.all(
+        entries.map(async e => {
+          const full = path.join(dirPath, e.name)
+          if (e.isDirectory()) {
+            total += await dirSize(full)
+          } else {
+            try {
+              const stat = await fs.stat(full)
+              total += stat.size
+            } catch { /* skip */ }
+          }
+        })
+      )
+    } catch { /* skip inaccessible dirs */ }
+    return total
+  }
+  return dirSize(CLAUDE_DIR)
+}