nlm-memory 0.5.0 → 0.5.2

package/src/core/adapters/pi.ts

@@ -1,264 +0,0 @@
-/**
- * Pi adapter.
- *
- * Reads ~/.pi/agent/sessions/<cwd-slug>/<timestamp>_<uuid>.jsonl. Pi writes
- * session files even when a run aborts — those still ingest, but the adapter
- * flags them via the `gitBranch: "aborted"` sentinel (SessionChunk has no
- * status field; storage layer decodes the sentinel later).
- *
- * File shape (v3, confirmed 2026-05-18): one JSON object per line. Five
- * event types: session, model_change, thinking_level_change, message,
- * custom_message. Only `message` produces turns; the rest are configuration
- * or extension-injected (custom_message must be excluded).
- *
- * Discovery is recursive (`<sessions>/<cwd-slug>/<file>.jsonl`).
- * $PI_SESSIONS_PATH overrides the default path.
- */
-
-import { promises as fs, existsSync, statSync } from "node:fs";
-import { extname, join } from "node:path";
-import { homedir } from "node:os";
-import type {
-  DetectionResult,
-  DiscoverOptions,
-  SessionChunk,
-  TranscriptAdapter,
-} from "@ports/transcript-adapter.js";
-import {
-  durationMinutes,
-  normalizeTimestamp,
-  safeSessionId,
-} from "./common.js";
-
-const TOOL_RESULT_PREVIEW_CHARS = 240;
-
-export interface PiAdapterOptions {
-  readonly sessionsPath?: string;
-  readonly idleMinutes?: number;
-}
-
-interface Turn {
-  readonly role: "user" | "assistant";
-  readonly text: string;
-  readonly timestamp: string;
-}
-
-function defaultSessionsPath(): string {
-  return (
-    process.env["PI_SESSIONS_PATH"] ??
-    join(homedir(), ".pi", "agent", "sessions")
-  );
-}
-
-export class PiAdapter implements TranscriptAdapter {
-  readonly name = "pi";
-  readonly runtimeVersion = "pi/1.0";
-  readonly transcriptKind = "pi-jsonl";
-
-  private readonly sessionsPath: string;
-  readonly idleMinutes: number;
-
-  constructor(opts: PiAdapterOptions = {}) {
-    this.sessionsPath = opts.sessionsPath ?? defaultSessionsPath();
-    this.idleMinutes = opts.idleMinutes ?? 15;
-  }
-
-  detect(): DetectionResult {
-    const p = defaultSessionsPath();
-    if (existsSync(p) && statSync(p).isDirectory()) {
-      return { adapterName: this.name, enabled: true, path: p, hint: null };
-    }
-    return {
-      adapterName: this.name,
-      enabled: false,
-      path: null,
-      hint: "Pi not detected — ~/.pi/agent/sessions/ missing.",
-    };
-  }
-
-  async discover(options: DiscoverOptions = {}): Promise<ReadonlyArray<string>> {
-    if (!existsSync(this.sessionsPath)) return [];
-
-    const found: { mtime: number; path: string }[] = [];
-    await walk(this.sessionsPath, async (full, st) => {
-      if (st.size === 0) return;
-      if (extname(full) !== ".jsonl") return;
-      if (options.since && st.mtime < options.since) return;
-      found.push({ mtime: st.mtimeMs, path: full });
-    });
-    found.sort((a, b) => a.mtime - b.mtime);
-    return found.map((f) => f.path);
-  }
-
-  async parseSession(path: string): Promise<SessionChunk | null> {
-    let raw: string;
-    try {
-      raw = await fs.readFile(path, "utf8");
-    } catch {
-      return null;
-    }
-
-    const turns: Turn[] = [];
-    let sessionId = "";
-    let projectDir = "";
-    let startedAt = "";
-    let endedAt = "";
-    let totalBytes = 0;
-    let allAssistantErrors = true;
-
-    for (const line of raw.split("\n")) {
-      totalBytes += Buffer.byteLength(line, "utf8") + 1;
-      const trimmed = line.trim();
-      if (!trimmed) continue;
-
-      let evt: Record<string, unknown>;
-      try {
-        evt = JSON.parse(trimmed) as Record<string, unknown>;
-      } catch {
-        continue;
-      }
-
-      const evtType = evt["type"];
-
-      if (evtType === "session") {
-        if (typeof evt["id"] === "string") sessionId = evt["id"];
-        if (typeof evt["cwd"] === "string") projectDir = evt["cwd"];
-        continue;
-      }
-      if (
-        evtType === "model_change" ||
-        evtType === "thinking_level_change" ||
-        evtType === "custom_message"
-      ) {
-        continue;
-      }
-      if (evtType !== "message") continue;
-
-      const msg = isRecord(evt["message"]) ? (evt["message"] as Record<string, unknown>) : {};
-      const role = msg["role"];
-      if (role !== "user" && role !== "assistant") continue;
-
-      const innerTs = msg["timestamp"];
-      const outerTs = evt["timestamp"];
-      const ts = innerTs
-        ? normalizeTimestamp(innerTs)
-        : normalizeTimestamp(outerTs);
-      if (ts) {
-        if (!startedAt) startedAt = ts;
-        endedAt = ts;
-      }
-
-      const text = extractPiText(msg["content"]);
-
-      if (role === "assistant") {
-        const stop = typeof msg["stopReason"] === "string" ? msg["stopReason"] : "";
-        if (stop !== "error") allAssistantErrors = false;
-        if (!text.trim()) continue; // error turns have empty content
-      } else if (!text.trim()) {
-        continue;
-      }
-
-      turns.push({ role, text, timestamp: ts });
-    }
-
-    if (turns.length === 0) return null;
-
-    const hasUser = turns.some((t) => t.role === "user");
-    const hasSuccessfulAssistant = turns.some((t) => t.role === "assistant");
-    const isAborted =
-      hasUser && (!hasSuccessfulAssistant || allAssistantErrors);
-
-    const transcript = turns.map((t) => `[${t.role}] ${t.text}`).join("\n\n");
-    const duration = durationMinutes(startedAt, endedAt);
-    const label = provisionalLabel(turns);
-    const rawId = sessionId || path.split("/").pop()!.replace(".jsonl", "");
-
-    return {
-      id: safeSessionId("pi", rawId),
-      runtime: this.runtimeVersion,
-      runtimeSessionId: rawId,
-      sourcePath: path,
-      startedAt,
-      endedAt,
-      durationMin: duration,
-      turnCount: turns.length,
-      byteRange: [0, totalBytes],
-      projectDir,
-      gitBranch: isAborted ? "aborted" : "",
-      text: transcript,
-      label,
-    };
-  }
-}
-
-// ── helpers ──────────────────────────────────────────────────────────────
-
-function isRecord(v: unknown): v is Record<string, unknown> {
-  return !!v && typeof v === "object" && !Array.isArray(v);
-}
-
-async function walk(
-  dir: string,
-  onFile: (path: string, st: import("node:fs").Stats) => Promise<void>,
-): Promise<void> {
-  let entries;
-  try {
-    entries = await fs.readdir(dir, { withFileTypes: true });
-  } catch {
-    return;
-  }
-  for (const ent of entries) {
-    const full = join(dir, ent.name);
-    if (ent.isDirectory()) {
-      await walk(full, onFile);
-    } else if (ent.isFile()) {
-      try {
-        const st = await fs.stat(full);
-        await onFile(full, st);
-      } catch {
-        continue;
-      }
-    }
-  }
-}
-
-function extractPiText(content: unknown): string {
-  if (typeof content === "string") return content.trim();
-  if (!Array.isArray(content)) return "";
-
-  const parts: string[] = [];
-  for (const block of content) {
-    if (!isRecord(block)) continue;
-    const btype = block["type"];
-    if (btype === "text") {
-      const t = block["text"];
-      if (typeof t === "string" && t.trim()) parts.push(t.trim());
-    } else if (btype === "tool_use") {
-      const name = typeof block["name"] === "string" ? block["name"] : "tool";
-      parts.push(`[tool_use: ${name}]`);
-    } else if (btype === "tool_result") {
-      let res = block["content"];
-      if (Array.isArray(res)) {
-        res = res
-          .filter(isRecord)
-          .map((b) => (typeof b["text"] === "string" ? (b["text"] as string) : ""))
-          .join("\n");
-      }
-      if (typeof res === "string" && res) {
-        const preview = res.slice(0, TOOL_RESULT_PREVIEW_CHARS);
-        const ellipsis = res.length > TOOL_RESULT_PREVIEW_CHARS ? "..." : "";
-        parts.push(`[tool_result: ${preview}${ellipsis}]`);
-      }
-    }
-  }
-  return parts.filter((p) => p).join("\n");
-}
-
-function provisionalLabel(turns: ReadonlyArray<Turn>): string {
-  for (const t of turns) {
-    if (t.role !== "user") continue;
-    const firstLine = t.text.split("\n", 1)[0]?.trim();
-    if (firstLine) return firstLine.slice(0, 80);
-  }
-  return "Untitled session";
-}

package/src/core/adapters/windsurf.ts

@@ -1,386 +0,0 @@
-/**
- * WindsurfAdapter — reads Windsurf (Codeium Cascade) sessions.
- *
- * ## Storage locations (macOS; Linux uses ~/.config/)
- *
- *   Workspace DBs  ~/Library/Application Support/Windsurf/User/workspaceStorage/<hash>/state.vscdb
- *     Table: ItemTable
- *     Key:   workbench.panel.aichat.view.aichat.chatdata  — chat tabs
- *     Bubble role: type 'user' → user, type 'ai' → assistant
- *
- *   Global DB  ~/Library/Application Support/Windsurf/User/globalStorage/state.vscdb
- *     Table: cursorDiskKV (if present) — composerData:*, agentData:*, flowData:*
- *     Table: ItemTable (fallback)     — keys matching %agent%, %flow%, %cascade%
- *     Conversation format: type 1/2 (user/assistant) or role: user/assistant
- *
- * ## Session ID prefixes
- *
- *   ws_  — workspace chat tab (ItemTable chatdata)
- *   wsg_ — global DB agent/flow session (cursorDiskKV or ItemTable)
- *
- * ## pathOrUrl in source registry
- *   Path to the Windsurf User directory. The adapter discovers:
- *     <userDir>/workspaceStorage/<hash>/state.vscdb  (workspace)
- *     <userDir>/globalStorage/state.vscdb             (global)
- *
- * Env override: NLM_WINDSURF_USER_DIR
- */
-
-import { existsSync, readdirSync } from "node:fs";
-import { homedir } from "node:os";
-import { join } from "node:path";
-import Database from "better-sqlite3";
-import type {
-  DetectionResult,
-  DiscoverOptions,
-  SessionChunk,
-  TranscriptAdapter,
-} from "@ports/transcript-adapter.js";
-import { durationMinutes, normalizeTimestamp, safeSessionId } from "./common.js";
-
-export interface WindsurfAdapterOptions {
-  readonly userDir?: string;
-}
-
-// ── Types ─────────────────────────────────────────────────────────────────────
-
-interface ItemTableRow {
-  readonly key: string;
-  readonly value: string | null;
-}
-
-interface KVRow {
-  readonly key: string;
-  readonly value: string | null;
-}
-
-interface ChatTab {
-  readonly tabId?: string;
-  readonly chatTitle?: string;
-  readonly lastSendTime?: number;
-  readonly bubbles?: ChatBubble[];
-}
-
-interface ChatBubble {
-  readonly type?: "user" | "ai" | string;
-  readonly text?: string;
-  readonly rawText?: string;
-}
-
-interface AgentComposer {
-  readonly composerId?: string;
-  readonly name?: string;
-  readonly createdAt?: unknown;
-  readonly lastUpdatedAt?: unknown;
-  readonly conversation?: AgentBubble[];
-  readonly status?: string;
-}
-
-interface AgentBubble {
-  readonly type?: number | string;
-  readonly role?: string;
-  readonly text?: string;
-}
-
-type Turn = { role: "user" | "assistant"; text: string };
-
-const CHAT_KEY = "workbench.panel.aichat.view.aichat.chatdata";
-
-// ── Path helpers ──────────────────────────────────────────────────────────────
-
-export function defaultUserDir(): string {
-  if (process.env["NLM_WINDSURF_USER_DIR"]) return process.env["NLM_WINDSURF_USER_DIR"];
-  const home = homedir();
-  if (process.platform === "darwin") {
-    return join(home, "Library/Application Support/Windsurf/User");
-  }
-  return join(home, ".config/Windsurf/User");
-}
-
-function workspaceStorageDir(userDir: string): string {
-  return join(userDir, "workspaceStorage");
-}
-
-function globalDbPath(userDir: string): string {
-  return join(userDir, "globalStorage", "state.vscdb");
-}
-
-function listWorkspaceDbs(userDir: string): string[] {
-  const wsDir = workspaceStorageDir(userDir);
-  if (!existsSync(wsDir)) return [];
-  try {
-    return readdirSync(wsDir, { withFileTypes: true })
-      .filter((e) => e.isDirectory())
-      .map((e) => join(wsDir, e.name, "state.vscdb"))
-      .filter((p) => existsSync(p));
-  } catch {
-    return [];
-  }
-}
-
-// ── Turn extraction ───────────────────────────────────────────────────────────
-
-function extractChatTurns(bubbles: ChatBubble[]): Turn[] {
-  const turns: Turn[] = [];
-  for (const b of bubbles) {
-    const text = (b.rawText ?? b.text ?? "").trim();
-    if (!text) continue;
-    turns.push({ role: b.type === "user" ? "user" : "assistant", text });
-  }
-  return turns;
-}
-
-function extractAgentTurns(bubbles: AgentBubble[]): Turn[] {
-  const turns: Turn[] = [];
-  for (const b of bubbles) {
-    const text = (b.text ?? "").trim();
-    if (!text) continue;
-    // Accept numeric type (1/2) or string role
-    const isUser = b.type === 1 || b.role === "user";
-    const isAssistant = b.type === 2 || b.role === "assistant";
-    if (!isUser && !isAssistant) continue;
-    turns.push({ role: isUser ? "user" : "assistant", text });
-  }
-  return turns;
-}
-
-function provisionalLabel(turns: ReadonlyArray<Turn>): string {
-  for (const t of turns) {
-    if (t.role !== "user") continue;
-    const first = t.text.split("\n", 1)[0]?.trim();
-    if (first) return first.slice(0, 80);
-  }
-  return "Untitled session";
-}
-
-function buildChunk(
-  id: string,
-  runtimeSessionId: string,
-  sourcePath: string,
-  turns: Turn[],
-  startedAt: string,
-  endedAt: string,
-  label: string,
-): SessionChunk {
-  const text = turns.map((t) => `${t.role}: ${t.text}`).join("\n\n");
-  return {
-    id,
-    runtime: "windsurf/1.0",
-    runtimeSessionId,
-    sourcePath,
-    startedAt,
-    endedAt,
-    durationMin: durationMinutes(startedAt, endedAt),
-    turnCount: turns.length,
-    byteRange: [0, Buffer.byteLength(text, "utf8")],
-    projectDir: "",
-    gitBranch: "",
-    text,
-    label,
-  };
-}
-
-// ── Workspace chat helpers ────────────────────────────────────────────────────
-
-function parseTabsFromDb(dbPath: string): ChatTab[] {
-  let db: Database.Database | undefined;
-  try {
-    db = new Database(dbPath, { readonly: true });
-    const row = db
-      .prepare<[string], ItemTableRow>(`SELECT key, value FROM ItemTable WHERE key = ?`)
-      .get(CHAT_KEY);
-    if (!row?.value) return [];
-    const data = JSON.parse(row.value) as { tabs?: ChatTab[] };
-    return Array.isArray(data.tabs) ? data.tabs : [];
-  } catch {
-    return [];
-  } finally {
-    db?.close();
-  }
-}
-
-// ── Global DB helpers ─────────────────────────────────────────────────────────
-
-interface GlobalSession {
-  readonly id: string;       // raw composerId/session key
-  readonly meta: AgentComposer;
-  readonly dbPath: string;
-}
-
-function parseGlobalSessions(globalPath: string): GlobalSession[] {
-  if (!existsSync(globalPath)) return [];
-  let db: Database.Database | undefined;
-  const results: GlobalSession[] = [];
-  try {
-    db = new Database(globalPath, { readonly: true });
-    const tables = db
-      .prepare<[], { name: string }>(`SELECT name FROM sqlite_master WHERE type='table'`)
-      .all()
-      .map((r) => r.name);
-
-    if (tables.includes("cursorDiskKV")) {
-      const rows = db
-        .prepare<[], KVRow>(
-          `SELECT key, value FROM cursorDiskKV
-           WHERE key LIKE 'composerData:%' OR key LIKE 'agentData:%' OR key LIKE 'flowData:%'
-           ORDER BY rowid ASC`,
-        )
-        .all();
-      for (const row of rows) {
-        if (!row.value) continue;
-        try {
-          const meta = JSON.parse(row.value) as AgentComposer;
-          const rawId = meta.composerId ?? row.key.split(":").slice(1).join(":");
-          if (rawId) results.push({ id: rawId, meta, dbPath: globalPath });
-        } catch { /* skip */ }
-      }
-    } else if (tables.includes("ItemTable")) {
-      // Fallback: probe for agent/flow/cascade keys
-      const rows = db
-        .prepare<[], ItemTableRow>(
-          `SELECT key, value FROM ItemTable
-           WHERE key LIKE '%agent%' OR key LIKE '%flow%' OR key LIKE '%cascade%'`,
-        )
-        .all();
-      for (const row of rows) {
-        if (!row.value) continue;
-        try {
-          const data = JSON.parse(row.value);
-          if (typeof data !== "object" || !data) continue;
-          // Accept any object with a conversation array
-          const conv = (data as AgentComposer).conversation;
-          if (!Array.isArray(conv) || conv.length === 0) continue;
-          const id = (data as AgentComposer).composerId ?? row.key;
-          results.push({ id, meta: data as AgentComposer, dbPath: globalPath });
-        } catch { /* skip */ }
-      }
-    }
-  } catch { /* inaccessible */ } finally {
-    db?.close();
-  }
-  return results;
-}
-
-// ── Adapter ───────────────────────────────────────────────────────────────────
-
-export class WindsurfAdapter implements TranscriptAdapter {
-  readonly name = "windsurf";
-  readonly runtimeVersion = "windsurf/1.0";
-  readonly transcriptKind = "windsurf-sqlite";
-
-  private readonly userDir: string;
-
-  constructor(opts: WindsurfAdapterOptions = {}) {
-    this.userDir = opts.userDir ?? defaultUserDir();
-  }
-
-  detect(): DetectionResult {
-    if (existsSync(this.userDir)) {
-      return { adapterName: this.name, enabled: true, path: this.userDir, hint: null };
-    }
-    return {
-      adapterName: this.name,
-      enabled: false,
-      path: null,
-      hint: "Windsurf User directory not found — install Windsurf or set NLM_WINDSURF_USER_DIR.",
-    };
-  }
-
-  async discover(options?: DiscoverOptions): Promise<ReadonlyArray<string>> {
-    const seen = new Set<string>();
-    const ids: string[] = [];
-    const add = (id: string) => { if (!seen.has(id)) { seen.add(id); ids.push(id); } };
-    const cutoff = options?.since?.getTime();
-
-    // ── Workspace chat tabs ───────────────────────────────────────────────
-    for (const dbPath of listWorkspaceDbs(this.userDir)) {
-      for (const tab of parseTabsFromDb(dbPath)) {
-        if (!tab.tabId) continue;
-        if (!(tab.bubbles && tab.bubbles.length > 0)) continue;
-        if (cutoff !== undefined) {
-          const ts = tab.lastSendTime;
-          // Only skip when we have a real non-zero timestamp older than the cutoff.
-          // Missing (undefined) or zero timestamps are treated as "unknown age" — include.
-          if (ts !== undefined && ts > 0 && ts < cutoff) continue;
-        }
-        add(`ws_${tab.tabId}`);
-      }
-    }
-
-    // ── Global agent/flow sessions ────────────────────────────────────────
-    for (const gs of parseGlobalSessions(globalDbPath(this.userDir))) {
-      if (cutoff !== undefined) {
-        const ts = gs.meta.lastUpdatedAt ?? gs.meta.createdAt;
-        if (ts !== undefined && ts !== null) {
-          const normalized = normalizeTimestamp(ts);
-          if (normalized && Date.parse(normalized) < cutoff) continue;
-        }
-      }
-      add(`wsg_${gs.id}`);
-    }
-
-    return ids;
-  }
-
-  async parseSession(id: string): Promise<SessionChunk | null> {
-    if (id.startsWith("wsg_")) {
-      return this._parseGlobalSession(id.slice("wsg_".length));
-    }
-    // ws_ prefix or legacy plain tabId
-    const tabId = id.startsWith("ws_") ? id.slice("ws_".length) : id;
-    return this._parseWorkspaceChatTab(tabId);
-  }
-
-  private _parseWorkspaceChatTab(tabId: string): SessionChunk | null {
-    for (const dbPath of listWorkspaceDbs(this.userDir)) {
-      const tabs = parseTabsFromDb(dbPath);
-      const tab = tabs.find((t) => t.tabId === tabId);
-      if (!tab) continue;
-
-      const turns = extractChatTurns(tab.bubbles ?? []);
-      if (turns.length === 0) return null;
-
-      const endedAtMs = tab.lastSendTime ?? 0;
-      const endedAt = endedAtMs > 0 ? new Date(endedAtMs).toISOString() : "";
-      const label = tab.chatTitle?.trim()
-        ? tab.chatTitle.trim().slice(0, 80)
-        : provisionalLabel(turns);
-
-      return buildChunk(
-        safeSessionId("ws", tabId),
-        tabId,
-        `${dbPath}::${tabId}`,
-        turns,
-        endedAt,
-        endedAt,
-        label,
-      );
-    }
-    return null;
-  }
-
-  private _parseGlobalSession(rawId: string): SessionChunk | null {
-    for (const gs of parseGlobalSessions(globalDbPath(this.userDir))) {
-      if (gs.id !== rawId) continue;
-      const turns = extractAgentTurns(gs.meta.conversation ?? []);
-      if (turns.length === 0) return null;
-
-      const startedAt = normalizeTimestamp(gs.meta.createdAt ?? gs.meta.lastUpdatedAt ?? "");
-      const endedAt = normalizeTimestamp(gs.meta.lastUpdatedAt ?? gs.meta.createdAt ?? "");
-      const label = gs.meta.name?.trim()
-        ? gs.meta.name.trim().slice(0, 80)
-        : provisionalLabel(turns);
-
-      return buildChunk(
-        safeSessionId("wsg", rawId),
-        rawId,
-        `${gs.dbPath}::${rawId}`,
-        turns,
-        startedAt,
-        endedAt,
-        label,
-      );
-    }
-    return null;
-  }
-}