@voidwire/lore 1.0.6 → 1.0.8

package/lib/config.ts

@@ -33,6 +33,10 @@ export interface LoreConfig {
     sqlite: string;
     custom_sqlite?: string;
   };
+  embedding: {
+    model: string;
+    dimensions: number;
+  };
 }
 
 let cachedConfig: LoreConfig | null = null;
@@ -81,9 +85,27 @@ export function getConfig(): LoreConfig {
       "Invalid config: missing [database] section in config.toml",
     );
   }
+  if (!parsed.embedding || typeof parsed.embedding !== "object") {
+    throw new Error(
+      "Invalid config: missing [embedding] section in config.toml",
+    );
+  }
 
   const paths = parsed.paths as Record<string, unknown>;
   const database = parsed.database as Record<string, unknown>;
+  const embedding = parsed.embedding as Record<string, unknown>;
+
+  // Validate required embedding fields
+  if (typeof embedding.model !== "string") {
+    throw new Error(
+      "Invalid config: embedding.model is missing or not a string",
+    );
+  }
+  if (typeof embedding.dimensions !== "number") {
+    throw new Error(
+      "Invalid config: embedding.dimensions is missing or not a number",
+    );
+  }
 
   // Validate required path fields
   const requiredPaths = [
@@ -140,7 +162,11 @@ export function getConfig(): LoreConfig {
           ? resolvePath(database.custom_sqlite)
           : undefined,
     },
+    embedding: {
+      model: embedding.model as string,
+      dimensions: embedding.dimensions as number,
+    },
   };
 
-  return cachedConfig;
+  return cachedConfig!;
 }

package/lib/indexer.ts

@@ -18,6 +18,7 @@
 
 import { Database } from "bun:sqlite";
 import { createHash } from "crypto";
+import { existsSync } from "fs";
 import { getConfig, type LoreConfig } from "./config";
 
 export interface IndexEntry {
@@ -39,6 +40,35 @@ export interface IndexerContext {
 
 export type IndexerFunction = (ctx: IndexerContext) => Promise<void>;
 
+/**
+ * Check if a path is configured and exists on disk.
+ * Logs a specific reason when the check fails:
+ *   - "not configured" when path is undefined
+ *   - "not found: /path" or "not found — hint" when path doesn't exist
+ *
+ * Returns true (with type narrowing) if path exists.
+ */
+export function checkPath(
+  source: string,
+  name: string,
+  path: string | undefined,
+  hint?: string,
+): path is string {
+  if (!path) {
+    console.log(`${source}: ${name} not configured`);
+    return false;
+  }
+  if (!existsSync(path)) {
+    if (hint) {
+      console.log(`${source}: ${name} not found — ${hint}`);
+    } else {
+      console.log(`${source}: ${name} not found: ${path}`);
+    }
+    return false;
+  }
+  return true;
+}
+
 /**
  * Content chunking with overlap.
  * Splits content at sentence boundaries when possible.

package/lib/indexers/blogs.ts

@@ -13,7 +13,7 @@
 
 import { readdirSync, readFileSync, statSync, existsSync } from "fs";
 import { join, basename } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 function walkMarkdownFiles(dir: string, files: string[] = []): string[] {
   if (!existsSync(dir)) return files;
@@ -37,10 +37,7 @@ export async function indexBlogs(ctx: IndexerContext): Promise<void> {
   const blogsDir = ctx.config.paths.blogs;
   const postsDir = join(blogsDir, "content", "posts");
 
-  if (!existsSync(postsDir)) {
-    console.log(`Blog posts directory not found: ${postsDir}`);
-    return;
-  }
+  if (!checkPath("blogs", "content/posts", postsDir)) return;
 
   if (!ctx.config.paths.blog_url) {
     console.warn(

package/lib/indexers/captures.ts

@@ -10,15 +10,20 @@
  * Timestamp: event timestamp
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexCaptures(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping captures");
+  if (
+    !checkPath(
+      "captures",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
 

package/lib/indexers/commits.ts

@@ -13,15 +13,11 @@
 import { readdirSync, existsSync } from "fs";
 import { join } from "path";
 import { spawnSync } from "child_process";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexCommits(ctx: IndexerContext): Promise<void> {
   const projectsDir = ctx.config.paths.projects;
-
-  if (!existsSync(projectsDir)) {
-    console.log(`Projects directory not found: ${projectsDir}`);
-    return;
-  }
+  if (!checkPath("commits", "paths.projects", projectsDir)) return;
 
   const projects = readdirSync(projectsDir, { withFileTypes: true })
     .filter((dirent) => dirent.isDirectory())

package/lib/indexers/development.ts

@@ -10,15 +10,11 @@
 
 import { readdirSync, readFileSync, statSync, existsSync } from "fs";
 import { join } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexDevelopment(ctx: IndexerContext): Promise<void> {
   const projectsDir = ctx.config.paths.projects;
-
-  if (!existsSync(projectsDir)) {
-    console.log(`Projects directory not found: ${projectsDir}`);
-    return;
-  }
+  if (!checkPath("development", "paths.projects", projectsDir)) return;
 
   const projects = readdirSync(projectsDir, { withFileTypes: true })
     .filter((dirent) => dirent.isDirectory())

package/lib/indexers/events.ts

@@ -10,16 +10,20 @@
  * Timestamp: last event timestamp per project
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexEvents(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping events");
+  if (
+    !checkPath(
+      "events",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
   const projectData = new Map<

package/lib/indexers/explorations.ts

@@ -10,9 +10,9 @@
  * Timestamp: file mtime as ISO 8601
  */
 
-import { readdirSync, readFileSync, statSync, existsSync } from "fs";
+import { readdirSync, readFileSync, statSync } from "fs";
 import { join, basename, dirname } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 function walkMarkdownFiles(dir: string, files: string[] = []): string[] {
   const entries = readdirSync(dir, { withFileTypes: true });
@@ -33,10 +33,7 @@ function walkMarkdownFiles(dir: string, files: string[] = []): string[] {
 export async function indexExplorations(ctx: IndexerContext): Promise<void> {
   const explorationsDir = ctx.config.paths.explorations;
 
-  if (!existsSync(explorationsDir)) {
-    console.log(`Explorations directory not found: ${explorationsDir}`);
-    return;
-  }
+  if (!checkPath("explorations", "paths.explorations", explorationsDir)) return;
 
   const files = walkMarkdownFiles(explorationsDir);
 

package/lib/indexers/flux.ts

@@ -14,7 +14,7 @@
 
 import { readdirSync, readFileSync, existsSync, statSync } from "fs";
 import { join, basename } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexFlux(ctx: IndexerContext): Promise<void> {
   const fluxDir = ctx.config.paths.flux;
@@ -22,7 +22,7 @@ export async function indexFlux(ctx: IndexerContext): Promise<void> {
   let found = false;
 
   // Pass 1: General flux files (no project association)
-  if (fluxDir && existsSync(fluxDir)) {
+  if (checkPath("flux", "paths.flux", fluxDir)) {
     found = true;
     const files = readdirSync(fluxDir).filter((f) => f.endsWith(".md"));
     for (const file of files) {
@@ -37,7 +37,7 @@ export async function indexFlux(ctx: IndexerContext): Promise<void> {
   }
 
   // Pass 2: Per-project flux files (active.md, later.md)
-  if (fluxProjectsDir && existsSync(fluxProjectsDir)) {
+  if (checkPath("flux", "paths.flux_projects", fluxProjectsDir)) {
     found = true;
     const projects = readdirSync(fluxProjectsDir, { withFileTypes: true })
       .filter((d) => d.isDirectory())
@@ -57,9 +57,7 @@ export async function indexFlux(ctx: IndexerContext): Promise<void> {
     }
   }
 
-  if (!found) {
-    console.log("No flux directories found, skipping flux");
-  }
+  if (!found) return;
 }
 
 function statusFromFilename(name: string): string {

package/lib/indexers/insights.ts

@@ -10,15 +10,20 @@
  * Timestamp: event timestamp
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexInsights(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping insights");
+  if (
+    !checkPath(
+      "insights",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
 

package/lib/indexers/learnings.ts

@@ -10,16 +10,20 @@
  * Timestamp: event timestamp
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexLearnings(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping learnings");
+  if (
+    !checkPath(
+      "learnings",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
 

package/lib/indexers/observations.ts

@@ -10,15 +10,20 @@
  * Timestamp: event timestamp
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexObservations(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping observations");
+  if (
+    !checkPath(
+      "observations",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
 

package/lib/indexers/obsidian.ts

@@ -11,9 +11,9 @@
  * Timestamp: file mtime as ISO 8601
  */
 
-import { readdirSync, readFileSync, statSync, existsSync } from "fs";
+import { readdirSync, readFileSync, statSync } from "fs";
 import { join, basename, dirname } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 function walkMarkdownFiles(
   dir: string,
@@ -41,10 +41,7 @@ function walkMarkdownFiles(
 export async function indexObsidian(ctx: IndexerContext): Promise<void> {
   const obsidianDir = ctx.config.paths.obsidian;
 
-  if (!existsSync(obsidianDir)) {
-    console.log(`Obsidian directory not found: ${obsidianDir}`);
-    return;
-  }
+  if (!checkPath("obsidian", "paths.obsidian", obsidianDir)) return;
 
   const files = walkMarkdownFiles(obsidianDir, obsidianDir);
 

package/lib/indexers/personal.ts

@@ -12,7 +12,7 @@
 
 import { readFileSync, statSync, existsSync } from "fs";
 import { join } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 function fileMtime(path: string): string {
   return statSync(path).mtime.toISOString();
@@ -27,10 +27,7 @@ function toISO(dateStr: string, fallback: string): string {
 export async function indexPersonal(ctx: IndexerContext): Promise<void> {
   const personalDir = ctx.config.paths.personal;
 
-  if (!existsSync(personalDir)) {
-    console.log(`Personal data directory not found: ${personalDir}`);
-    return;
-  }
+  if (!checkPath("personal", "paths.personal", personalDir)) return;
 
   // Books
   const booksPath = join(personalDir, "books.json");

package/lib/indexers/readmes.ts

@@ -10,15 +10,11 @@
 
 import { readdirSync, readFileSync, statSync, existsSync } from "fs";
 import { join } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexReadmes(ctx: IndexerContext): Promise<void> {
   const projectsDir = ctx.config.paths.projects;
-
-  if (!existsSync(projectsDir)) {
-    console.log(`Projects directory not found: ${projectsDir}`);
-    return;
-  }
+  if (!checkPath("readmes", "paths.projects", projectsDir)) return;
 
   const projects = readdirSync(projectsDir, { withFileTypes: true })
     .filter((dirent) => dirent.isDirectory())

package/lib/indexers/sessions.ts

@@ -10,9 +10,9 @@
  * Timestamp: first event timestamp per session
  */
 
-import { readdirSync, readFileSync, existsSync } from "fs";
+import { readdirSync, readFileSync } from "fs";
 import { join } from "path";
-import type { IndexerContext } from "../indexer";
+import { checkPath, type IndexerContext } from "../indexer";
 
 interface SessionData {
   project: string;
@@ -27,17 +27,19 @@ interface SessionData {
 }
 
 export async function indexSessions(ctx: IndexerContext): Promise<void> {
-  // Collect event files from all configured directories
-  const eventDirs = [
-    ctx.config.paths.session_events,
-    ctx.config.paths.sable_events,
-  ].filter((d): d is string => !!d && existsSync(d));
-
-  if (eventDirs.length === 0) {
-    console.log("No session events directories found, skipping sessions");
-    return;
+  // Check each event directory individually for clear diagnostics
+  const eventDirs: string[] = [];
+  for (const { name, path } of [
+    { name: "session_events", path: ctx.config.paths.session_events },
+    { name: "sable_events", path: ctx.config.paths.sable_events },
+  ]) {
+    if (checkPath("sessions", name, path)) {
+      eventDirs.push(path);
+    }
   }
 
+  if (eventDirs.length === 0) return;
+
   const eventFiles: string[] = [];
   for (const dir of eventDirs) {
     const files = readdirSync(dir)

package/lib/indexers/teachings.ts

@@ -10,15 +10,20 @@
  * Timestamp: event timestamp
  */
 
-import { readFileSync, existsSync } from "fs";
-import type { IndexerContext } from "../indexer";
+import { readFileSync } from "fs";
+import { checkPath, type IndexerContext } from "../indexer";
 
 export async function indexTeachings(ctx: IndexerContext): Promise<void> {
   const logPath = `${ctx.config.paths.data}/log.jsonl`;
-  if (!existsSync(logPath)) {
-    console.log("No log.jsonl found, skipping teachings");
+  if (
+    !checkPath(
+      "teachings",
+      "log.jsonl",
+      logPath,
+      "populated by Sable session hooks",
+    )
+  )
     return;
-  }
 
   const lines = readFileSync(logPath, "utf-8").split("\n").filter(Boolean);
 

package/lib/semantic.ts

@@ -11,6 +11,7 @@ import { existsSync } from "fs";
 import { pipeline } from "@huggingface/transformers";
 import { getDatabasePath, openDatabase } from "./db.js";
 import { search as keywordSearch, type SearchResult } from "./search.js";
+import { getConfig } from "./config.js";
 
 export interface SemanticResult {
   rowid: number;
@@ -30,8 +31,7 @@ export interface SemanticSearchOptions {
   since?: string;
 }
 
-const MODEL_NAME = "nomic-ai/nomic-embed-text-v1.5";
-const EMBEDDING_DIM = 768;
+const { model: MODEL_NAME, dimensions: EMBEDDING_DIM } = getConfig().embedding;
 
 interface EmbeddingPipeline {
   (

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@voidwire/lore",
-  "version": "1.0.6",
+  "version": "1.0.8",
   "description": "Unified knowledge CLI - Search, list, and capture your indexed knowledge",
   "type": "module",
   "main": "./index.ts",