libsql-search 0.1.5 → 0.2.0

package/dist/index.cjs

@@ -4,7 +4,15 @@ var promises = require('fs/promises');
 var path = require('path');
 var matter = require('gray-matter');
 
-const providerCache = {};
+const DEFAULT_DIMENSIONS = 768;
+const DEFAULT_MAX_LENGTH = 8e3;
+const DEFAULT_TIMEOUT_MS = 3e4;
+const LOCAL_MODEL = "Xenova/all-MiniLM-L6-v2";
+const GEMINI_MODEL = "text-embedding-004";
+const OPENAI_SMALL_MODEL = "text-embedding-3-small";
+const OPENAI_LARGE_MODEL = "text-embedding-3-large";
+const OPENAI_EMBEDDINGS_URL = "https://api.openai.com/v1/embeddings";
+const localModelCacheByModel = /* @__PURE__ */ new Map();
 function getEnvironmentVariable(name) {
   const runtime = globalThis;
   const nodeValue = runtime.process?.env?.[name];
@@ -17,80 +25,403 @@ function getEnvironmentVariable(name) {
     return void 0;
   }
 }
-async function getLocalEmbeddingModel() {
-  if (!providerCache.local) {
-    console.log("Loading local embedding model (Xenova/all-MiniLM-L6-v2)...");
-    const { pipeline } = await import('@xenova/transformers');
-    providerCache.local = await pipeline("feature-extraction", "Xenova/all-MiniLM-L6-v2");
-    console.log("Local model loaded successfully");
+async function getLocalEmbeddingModel(modelName) {
+  const cached = localModelCacheByModel.get(modelName);
+  if (cached?.model) {
+    return cached.model;
   }
-  return providerCache.local;
+  console.log(`Loading local embedding model (${modelName})...`);
+  const { pipeline } = await import('@xenova/transformers');
+  const model = await pipeline("feature-extraction", modelName);
+  localModelCacheByModel.set(modelName, { model });
+  console.log("Local model loaded successfully");
+  return model;
 }
-async function generateEmbedding(text, options = {}) {
-  const {
-    provider = "local",
-    apiKey,
-    dimensions = 768,
-    maxLength = 8e3
-  } = options;
-  const truncated = text.substring(0, maxLength);
-  switch (provider) {
+function getPositiveInteger(value, optionName) {
+  if (!Number.isInteger(value) || value <= 0) {
+    throw new Error(`Invalid ${optionName}: expected a positive integer`);
+  }
+  return value;
+}
+function getTimeoutMs(value) {
+  return getPositiveInteger(value ?? DEFAULT_TIMEOUT_MS, "timeoutMs");
+}
+function resolveProviderName(provider) {
+  switch (provider ?? "local") {
     case "local":
-      return generateLocalEmbedding(truncated, dimensions);
     case "gemini":
-      return generateGeminiEmbedding(truncated, apiKey);
     case "openai":
-      return generateOpenAIEmbedding(truncated, apiKey, dimensions);
+      return provider ?? "local";
     default:
-      throw new Error(`Unknown embedding provider: ${provider}`);
+      throw new Error(`Unknown embedding provider: ${String(provider)}`);
+  }
+}
+function resolveIntent(intent) {
+  switch (intent ?? "document") {
+    case "document":
+    case "query":
+      return intent ?? "document";
+    default:
+      throw new Error(`Unknown embedding intent: ${String(intent)}`);
+  }
+}
+function truncateTexts(texts, maxLength) {
+  return texts.map((text) => text.substring(0, maxLength));
+}
+function getOpenAIModel(dimensions) {
+  return dimensions <= 1536 ? OPENAI_SMALL_MODEL : OPENAI_LARGE_MODEL;
+}
+function hasOwnProperty(value, property) {
+  return Object.prototype.hasOwnProperty.call(value, property);
+}
+function redactErrorText(value) {
+  const raw = value instanceof Error ? value.message : String(value);
+  return raw.replace(/Bearer\s+[A-Za-z0-9._~+/=-]+/gi, "Bearer [redacted]").replace(/([?&](?:api[_-]?key|key|token)=)[^&\s]+/gi, "$1[redacted]").replace(/(authorization\s*[:=]\s*)[^\s,}]+/gi, "$1[redacted]").replace(/https?:\/\/[^\s)]+/gi, (match) => {
+    try {
+      const url = new URL(match);
+      return `${url.origin}${url.pathname}`;
+    } catch {
+      return "[redacted-url]";
+    }
+  }).slice(0, 300);
+}
+function providerError(provider, message, cause) {
+  const safeCause = cause === void 0 ? "" : `: ${redactErrorText(cause).trim()}`;
+  return new Error(`${provider} embedding error: ${message}${safeCause}`);
+}
+function getResponseHeader(response, name) {
+  return response.headers.get(name) ?? void 0;
+}
+async function withTimeout(provider, operation, timeoutMs, run, parentSignal) {
+  if (parentSignal?.aborted) {
+    throw providerError(provider, `${operation} was aborted`);
+  }
+  const controller = new AbortController();
+  let abortError;
+  let rejectAbort = () => {
+  };
+  const abortPromise = new Promise((_resolve, reject) => {
+    rejectAbort = reject;
+  });
+  const abort = (error) => {
+    if (abortError) {
+      return;
+    }
+    abortError = error;
+    controller.abort();
+    rejectAbort(error);
+  };
+  const onParentAbort = () => abort(providerError(provider, `${operation} was aborted`));
+  const timeout = setTimeout(() => {
+    abort(providerError(provider, `${operation} timed out after ${timeoutMs}ms`));
+  }, timeoutMs);
+  parentSignal?.addEventListener("abort", onParentAbort, { once: true });
+  const operationPromise = Promise.resolve().then(() => run(controller.signal));
+  operationPromise.catch(() => void 0);
+  try {
+    return await Promise.race([operationPromise, abortPromise]);
+  } catch (error) {
+    if (error === abortError) {
+      throw error;
+    }
+    throw providerError(provider, `${operation} failed`, error);
+  } finally {
+    clearTimeout(timeout);
+    parentSignal?.removeEventListener("abort", onParentAbort);
+  }
+}
+function validateEmbeddingBatch(items, expectedCount, expectedDimensions, provider) {
+  if (items.length !== expectedCount) {
+    throw providerError(
+      provider,
+      `expected ${expectedCount} embedding result(s), received ${items.length}`
+    );
+  }
+  const hasIndexedItems = items.some((item) => !Array.isArray(item) && hasOwnProperty(item, "index"));
+  const ordered = hasIndexedItems ? reorderIndexedEmbeddings(items, expectedCount, provider) : items.map((item) => Array.isArray(item) ? item : item.embedding);
+  return ordered.map(
+    (embedding, itemIndex) => validateEmbeddingVector(embedding, expectedDimensions, provider, itemIndex)
+  );
+}
+function reorderIndexedEmbeddings(items, expectedCount, provider) {
+  const ordered = new Array(expectedCount);
+  const seen = /* @__PURE__ */ new Set();
+  for (const item of items) {
+    if (Array.isArray(item) || !hasOwnProperty(item, "index")) {
+      throw providerError(provider, "provider returned a partially indexed embedding batch");
+    }
+    if (typeof item.index !== "number" || !Number.isInteger(item.index) || item.index < 0 || item.index >= expectedCount) {
+      throw providerError(provider, `provider returned invalid embedding index ${String(item.index)}`);
+    }
+    if (seen.has(item.index)) {
+      throw providerError(provider, `provider returned duplicate embedding index ${item.index}`);
+    }
+    seen.add(item.index);
+    ordered[item.index] = item.embedding;
+  }
+  if (seen.size !== expectedCount) {
+    throw providerError(provider, "provider returned non-contiguous embedding indices");
+  }
+  return ordered;
+}
+function validateEmbeddingVector(embedding, expectedDimensions, provider, itemIndex) {
+  if (!Array.isArray(embedding)) {
+    throw providerError(provider, `embedding ${itemIndex} is not an array`);
+  }
+  if (embedding.length !== expectedDimensions) {
+    throw providerError(
+      provider,
+      `embedding ${itemIndex} has ${embedding.length} dimensions, expected ${expectedDimensions}`
+    );
+  }
+  for (let i = 0; i < embedding.length; i++) {
+    const value = embedding[i];
+    if (typeof value !== "number" || !Number.isFinite(value)) {
+      throw providerError(provider, `embedding ${itemIndex} contains a non-finite value at dimension ${i}`);
+    }
+  }
+  return embedding;
+}
+function createProviderMetadata(provider, dimensions) {
+  switch (provider) {
+    case "local":
+      return Object.freeze({
+        name: "local",
+        model: LOCAL_MODEL,
+        dimensions,
+        batch: Object.freeze({ mode: "sequential" })
+      });
+    case "gemini":
+      return Object.freeze({
+        name: "gemini",
+        model: GEMINI_MODEL,
+        dimensions: DEFAULT_DIMENSIONS,
+        batch: Object.freeze({ mode: "sequential" })
+      });
+    case "openai":
+      return Object.freeze({
+        name: "openai",
+        model: getOpenAIModel(dimensions),
+        dimensions,
+        batch: Object.freeze({ mode: "native", maxSize: 2048 })
+      });
   }
 }
-async function generateLocalEmbedding(text, targetDimensions) {
-  const model = await getLocalEmbeddingModel();
-  const output = await model(text, {
-    pooling: "mean",
-    normalize: true
+function getEffectiveDimensions(provider, dimensions) {
+  if (provider === "gemini") {
+    return DEFAULT_DIMENSIONS;
+  }
+  return getPositiveInteger(dimensions ?? DEFAULT_DIMENSIONS, "dimensions");
+}
+function assertBatchSize(metadata, count) {
+  if (metadata.batch.maxSize !== void 0 && count > metadata.batch.maxSize) {
+    throw providerError(
+      metadata.name,
+      `batch size ${count} exceeds maximum ${metadata.batch.maxSize}`
+    );
+  }
+}
+function createEmbeddingBatchResult(metadata, intent, embeddings) {
+  return Object.freeze({
+    embeddings,
+    provider: metadata.name,
+    model: metadata.model,
+    dimensions: metadata.dimensions,
+    intent
   });
-  const embedding = Array.from(output.data);
-  return padEmbedding(embedding, targetDimensions);
-}
-async function generateGeminiEmbedding(text, apiKey) {
-  const key = apiKey || getEnvironmentVariable("GEMINI_API_KEY");
-  if (!key) {
-    throw new Error("GEMINI_API_KEY is required for Gemini embeddings");
-  }
-  if (!providerCache.gemini) {
-    const { GoogleGenerativeAI } = await import('@google/generative-ai');
-    const genAI = new GoogleGenerativeAI(key);
-    providerCache.gemini = genAI.getGenerativeModel({ model: "text-embedding-004" });
-  }
-  const result = await providerCache.gemini.embedContent(text);
-  return result.embedding.values;
-}
-async function generateOpenAIEmbedding(text, apiKey, dimensions = 1536) {
-  const key = apiKey || getEnvironmentVariable("OPENAI_API_KEY");
-  if (!key) {
-    throw new Error("OPENAI_API_KEY is required for OpenAI embeddings");
-  }
-  const model = dimensions <= 1536 ? "text-embedding-3-small" : "text-embedding-3-large";
-  const response = await fetch("https://api.openai.com/v1/embeddings", {
-    method: "POST",
-    headers: {
-      "Content-Type": "application/json",
-      "Authorization": `Bearer ${key}`
-    },
-    body: JSON.stringify({
-      input: text,
-      model,
-      dimensions
-    })
+}
+class LocalEmbeddingProvider {
+  metadata;
+  #timeoutMs;
+  constructor(metadata, timeoutMs) {
+    this.metadata = metadata;
+    this.#timeoutMs = timeoutMs;
+  }
+  async embed(texts, options = {}) {
+    const intent = resolveIntent(options.intent);
+    if (texts.length === 0) {
+      return createEmbeddingBatchResult(this.metadata, intent, []);
+    }
+    assertBatchSize(this.metadata, texts.length);
+    const vectors = await withTimeout(
+      "local",
+      "model inference",
+      this.#timeoutMs,
+      async () => {
+        const model = await getLocalEmbeddingModel(this.metadata.model);
+        const results = [];
+        for (const text of texts) {
+          const output = await model(text, {
+            pooling: "mean",
+            normalize: true
+          });
+          const embedding = Array.from(output.data);
+          results.push(padEmbedding(embedding, this.metadata.dimensions));
+        }
+        return results;
+      },
+      options.signal
+    );
+    return createEmbeddingBatchResult(
+      this.metadata,
+      intent,
+      validateEmbeddingBatch(vectors, texts.length, this.metadata.dimensions, "local")
+    );
+  }
+}
+class GeminiEmbeddingProvider {
+  metadata;
+  #apiKey;
+  #timeoutMs;
+  constructor(metadata, apiKey, timeoutMs) {
+    this.metadata = metadata;
+    this.#apiKey = apiKey;
+    this.#timeoutMs = timeoutMs;
+  }
+  async embed(texts, options = {}) {
+    const intent = resolveIntent(options.intent);
+    if (texts.length === 0) {
+      return createEmbeddingBatchResult(this.metadata, intent, []);
+    }
+    assertBatchSize(this.metadata, texts.length);
+    const vectors = await withTimeout(
+      "gemini",
+      "API request",
+      this.#timeoutMs,
+      async () => {
+        const { GoogleGenerativeAI } = await import('@google/generative-ai');
+        const genAI = new GoogleGenerativeAI(this.#apiKey);
+        const model = genAI.getGenerativeModel({ model: this.metadata.model });
+        const results = [];
+        for (const text of texts) {
+          const result = await model.embedContent(text);
+          const values = result.embedding?.values;
+          if (!Array.isArray(values)) {
+            throw new Error("Gemini response did not include embedding values");
+          }
+          results.push(values);
+        }
+        return results;
+      },
+      options.signal
+    );
+    return createEmbeddingBatchResult(
+      this.metadata,
+      intent,
+      validateEmbeddingBatch(vectors, texts.length, this.metadata.dimensions, "gemini")
+    );
+  }
+}
+class OpenAIEmbeddingProvider {
+  metadata;
+  #apiKey;
+  #timeoutMs;
+  constructor(metadata, apiKey, timeoutMs) {
+    this.metadata = metadata;
+    this.#apiKey = apiKey;
+    this.#timeoutMs = timeoutMs;
+  }
+  async embed(texts, options = {}) {
+    const intent = resolveIntent(options.intent);
+    if (texts.length === 0) {
+      return createEmbeddingBatchResult(this.metadata, intent, []);
+    }
+    assertBatchSize(this.metadata, texts.length);
+    const items = await withTimeout(
+      "openai",
+      "API request",
+      this.#timeoutMs,
+      async (signal) => {
+        const response = await fetch(OPENAI_EMBEDDINGS_URL, {
+          method: "POST",
+          headers: {
+            "Content-Type": "application/json",
+            "Authorization": `Bearer ${this.#apiKey}`
+          },
+          signal,
+          body: JSON.stringify({
+            input: texts,
+            model: this.metadata.model,
+            dimensions: this.metadata.dimensions
+          })
+        });
+        if (!response.ok) {
+          const requestId = getResponseHeader(response, "x-request-id");
+          const context = requestId ? ` status ${response.status}, request ${requestId}` : ` status ${response.status}`;
+          throw new Error(context);
+        }
+        const data = await response.json();
+        if (!Array.isArray(data.data)) {
+          throw new Error("OpenAI response did not include a data array");
+        }
+        return data.data.map((item) => {
+          const typed = item;
+          const embedding = typed.embedding;
+          if (!Array.isArray(embedding)) {
+            throw new Error("OpenAI response item did not include an embedding array");
+          }
+          const parsed = { embedding };
+          if (hasOwnProperty(typed, "index")) {
+            parsed.index = typed.index;
+          }
+          return parsed;
+        });
+      },
+      options.signal
+    );
+    return createEmbeddingBatchResult(
+      this.metadata,
+      intent,
+      validateEmbeddingBatch(items, texts.length, this.metadata.dimensions, "openai")
+    );
+  }
+}
+function createEmbeddingProvider(options = {}) {
+  const provider = resolveProviderName(options.provider);
+  const metadata = getEmbeddingProviderMetadata(options);
+  const timeoutMs = getTimeoutMs(options.timeoutMs);
+  switch (provider) {
+    case "local":
+      return new LocalEmbeddingProvider(metadata, timeoutMs);
+    case "gemini": {
+      const key = options.apiKey || getEnvironmentVariable("GEMINI_API_KEY");
+      if (!key) {
+        throw new Error("GEMINI_API_KEY is required for Gemini embeddings");
+      }
+      return new GeminiEmbeddingProvider(metadata, key, timeoutMs);
+    }
+    case "openai": {
+      const key = options.apiKey || getEnvironmentVariable("OPENAI_API_KEY");
+      if (!key) {
+        throw new Error("OPENAI_API_KEY is required for OpenAI embeddings");
+      }
+      return new OpenAIEmbeddingProvider(metadata, key, timeoutMs);
+    }
+  }
+}
+function getEmbeddingProviderMetadata(options = {}) {
+  const provider = resolveProviderName(options.provider);
+  return createProviderMetadata(provider, getEffectiveDimensions(provider, options.dimensions));
+}
+async function generateEmbeddings(texts, options = {}) {
+  const maxLength = getPositiveInteger(options.maxLength ?? DEFAULT_MAX_LENGTH, "maxLength");
+  const intent = resolveIntent(options.intent);
+  if (texts.length === 0) {
+    return [];
+  }
+  const provider = createEmbeddingProvider(options);
+  const result = await provider.embed(truncateTexts(texts, maxLength), {
+    intent,
+    ...options.signal ? { signal: options.signal } : {}
   });
-  if (!response.ok) {
-    const error = await response.text();
-    throw new Error(`OpenAI API error: ${error}`);
+  return result.embeddings;
+}
+async function generateEmbedding(text, options = {}) {
+  const [embedding] = await generateEmbeddings([text], options);
+  if (!embedding) {
+    throw new Error("Embedding provider returned no embedding");
   }
-  const data = await response.json();
-  return data.data[0].embedding;
+  return embedding;
 }
 function padEmbedding(embedding, targetDimensions) {
   if (embedding.length === targetDimensions) {
@@ -114,6 +445,34 @@ function prepareTextForEmbedding(fields) {
   return parts.filter(Boolean).join("\n\n");
 }
 
+const SQL_IDENTIFIER_PATTERN = /^[A-Za-z_][A-Za-z0-9_]*$/;
+const DEFAULT_SEARCH_LIMIT = 10;
+const MAX_SEARCH_LIMIT = 100;
+function validateSqlIdentifier(identifier, name = "identifier") {
+  if (typeof identifier !== "string" || !SQL_IDENTIFIER_PATTERN.test(identifier)) {
+    throw new Error(
+      `Invalid SQL ${name}: expected an ASCII identifier matching ${SQL_IDENTIFIER_PATTERN.toString()}`
+    );
+  }
+  return identifier;
+}
+function quoteSqlIdentifier(identifier, name) {
+  const validated = validateSqlIdentifier(identifier, name);
+  return `"${validated.replace(/"/g, '""')}"`;
+}
+function normalizeSearchLimit(limit = DEFAULT_SEARCH_LIMIT) {
+  if (typeof limit !== "number" || !Number.isFinite(limit) || !Number.isInteger(limit) || limit < 1 || limit > MAX_SEARCH_LIMIT) {
+    throw new Error(`Invalid search limit: expected an integer from 1 to ${MAX_SEARCH_LIMIT}`);
+  }
+  return limit;
+}
+function normalizeVectorDimensions(dimensions) {
+  if (typeof dimensions !== "number" || !Number.isFinite(dimensions) || !Number.isInteger(dimensions) || dimensions < 1) {
+    throw new Error("Invalid vector dimensions: expected a positive integer");
+  }
+  return dimensions;
+}
+
 async function indexContent(options) {
   const {
     client,
@@ -124,12 +483,13 @@ async function indexContent(options) {
     tableName = "articles",
     onProgress
   } = options;
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
   const files = await findFiles(contentPath, contentPath, fileExtensions, exclude);
   if (files.length === 0) {
     console.warn(`No files found in ${contentPath}`);
     return { success: 0, failed: 0, total: 0 };
   }
-  await client.execute(`DELETE FROM ${tableName}`);
+  await client.execute(`DELETE FROM ${quotedTableName}`);
   let success = 0;
   let failed = 0;
   for (let i = 0; i < files.length; i++) {
@@ -139,7 +499,7 @@ async function indexContent(options) {
     }
     try {
       const document = await processFile(file, embeddingOptions);
-      await insertDocument(client, document, tableName);
+      await insertDocument(client, document, quotedTableName);
       success++;
     } catch (error) {
       console.error(`Failed to index ${file.relativePath}:`, error);
@@ -182,7 +542,10 @@ async function processFile(file, embeddingOptions) {
     content: markdown,
     tags
   });
-  const embedding = await generateEmbedding(embeddingText, embeddingOptions);
+  const embedding = await generateEmbedding(embeddingText, {
+    ...embeddingOptions,
+    intent: embeddingOptions.intent ?? "document"
+  });
   return {
     slug,
     title,
@@ -193,9 +556,9 @@ async function processFile(file, embeddingOptions) {
     metadata: frontMatter
   };
 }
-async function insertDocument(client, document, tableName) {
+async function insertDocument(client, document, quotedTableName) {
   await client.execute({
-    sql: `INSERT INTO ${tableName}
+    sql: `INSERT INTO ${quotedTableName}
           (slug, title, content, folder, tags, embedding, created_at, updated_at)
           VALUES (?, ?, ?, ?, ?, vector(?), datetime('now'), datetime('now'))`,
     args: [
@@ -209,30 +572,35 @@ async function insertDocument(client, document, tableName) {
   });
 }
 async function createTable(client, tableName = "articles", dimensions = 768) {
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
+  const vectorDimensions = normalizeVectorDimensions(dimensions);
+  const quotedEmbeddingIndexName = quoteSqlIdentifier(`${tableName}_embedding_idx`, "embedding index name");
+  const quotedFolderIndexName = quoteSqlIdentifier(`${tableName}_folder_idx`, "folder index name");
+  const quotedSlugIndexName = quoteSqlIdentifier(`${tableName}_slug_idx`, "slug index name");
   await client.execute(`
-    CREATE TABLE IF NOT EXISTS ${tableName} (
+    CREATE TABLE IF NOT EXISTS ${quotedTableName} (
       id INTEGER PRIMARY KEY AUTOINCREMENT,
       slug TEXT UNIQUE NOT NULL,
       title TEXT NOT NULL,
       content TEXT NOT NULL,
       folder TEXT NOT NULL DEFAULT 'root',
       tags TEXT DEFAULT '[]',
-      embedding F32_BLOB(${dimensions}),
+      embedding F32_BLOB(${vectorDimensions}),
       created_at TEXT NOT NULL,
       updated_at TEXT NOT NULL
     )
   `);
   await client.execute(`
-    CREATE INDEX IF NOT EXISTS ${tableName}_embedding_idx
-    ON ${tableName}(libsql_vector_idx(embedding))
+    CREATE INDEX IF NOT EXISTS ${quotedEmbeddingIndexName}
+    ON ${quotedTableName}(libsql_vector_idx(embedding))
   `);
   await client.execute(`
-    CREATE INDEX IF NOT EXISTS ${tableName}_folder_idx
-    ON ${tableName}(folder)
+    CREATE INDEX IF NOT EXISTS ${quotedFolderIndexName}
+    ON ${quotedTableName}(folder)
   `);
   await client.execute(`
-    CREATE INDEX IF NOT EXISTS ${tableName}_slug_idx
-    ON ${tableName}(slug)
+    CREATE INDEX IF NOT EXISTS ${quotedSlugIndexName}
+    ON ${quotedTableName}(slug)
   `);
 }
 
@@ -244,7 +612,12 @@ async function search(options) {
     tableName = "articles",
     embeddingOptions = {}
   } = options;
-  const queryEmbedding = await generateEmbedding(query, embeddingOptions);
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
+  const resultLimit = normalizeSearchLimit(limit);
+  const queryEmbedding = await generateEmbedding(query, {
+    ...embeddingOptions,
+    intent: embeddingOptions.intent ?? "query"
+  });
   const results = await client.execute({
     sql: `
       SELECT
@@ -256,12 +629,12 @@ async function search(options) {
         tags,
         created_at,
         vector_distance_cos(embedding, vector(?)) as distance
-      FROM ${tableName}
+      FROM ${quotedTableName}
       WHERE embedding IS NOT NULL
       ORDER BY distance
       LIMIT ?
     `,
-    args: [JSON.stringify(queryEmbedding), limit]
+    args: [JSON.stringify(queryEmbedding), resultLimit]
   });
   return results.rows.map((row) => ({
     id: row.id,
@@ -275,9 +648,10 @@ async function search(options) {
   }));
 }
 async function getAllArticles(client, tableName = "articles") {
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
   const results = await client.execute(`
     SELECT id, slug, title, folder, tags, created_at, updated_at
-    FROM ${tableName}
+    FROM ${quotedTableName}
     ORDER BY title
   `);
   return results.rows.map((row) => ({
@@ -291,10 +665,11 @@ async function getAllArticles(client, tableName = "articles") {
   }));
 }
 async function getArticleBySlug(client, slug, tableName = "articles") {
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
   const results = await client.execute({
     sql: `
       SELECT id, slug, title, content, folder, tags, created_at, updated_at
-      FROM ${tableName}
+      FROM ${quotedTableName}
       WHERE slug = ?
       LIMIT 1
     `,
@@ -316,10 +691,11 @@ async function getArticleBySlug(client, slug, tableName = "articles") {
   };
 }
 async function getArticlesByFolder(client, folder, tableName = "articles") {
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
   const results = await client.execute({
     sql: `
       SELECT id, slug, title, folder, tags
-      FROM ${tableName}
+      FROM ${quotedTableName}
       WHERE folder = ?
       ORDER BY title
     `,
@@ -334,21 +710,26 @@ async function getArticlesByFolder(client, folder, tableName = "articles") {
   }));
 }
 async function getFolders(client, tableName = "articles") {
+  const quotedTableName = quoteSqlIdentifier(tableName, "tableName");
   const results = await client.execute(`
     SELECT DISTINCT folder
-    FROM ${tableName}
+    FROM ${quotedTableName}
     ORDER BY folder
   `);
   return results.rows.map((row) => row.folder);
 }
 
+exports.createEmbeddingProvider = createEmbeddingProvider;
 exports.createTable = createTable;
 exports.generateEmbedding = generateEmbedding;
+exports.generateEmbeddings = generateEmbeddings;
 exports.getAllArticles = getAllArticles;
 exports.getArticleBySlug = getArticleBySlug;
 exports.getArticlesByFolder = getArticlesByFolder;
+exports.getEmbeddingProviderMetadata = getEmbeddingProviderMetadata;
 exports.getFolders = getFolders;
 exports.indexContent = indexContent;
 exports.padEmbedding = padEmbedding;
 exports.prepareTextForEmbedding = prepareTextForEmbedding;
 exports.search = search;
+exports.validateEmbeddingBatch = validateEmbeddingBatch;