audrey 0.14.0 → 0.16.0

package/src/llm.js

@@ -1,240 +1,246 @@
-/**
- * @typedef {Object} ChatMessage
- * @property {'system' | 'user' | 'assistant'} role
- * @property {string} content
- */
-
-/**
- * @typedef {Object} LLMCompletionResult
- * @property {string} content
- */
-
-/**
- * @typedef {Object} LLMCompletionOptions
- * @property {number} [maxTokens]
- */
-
-/**
- * @typedef {Object} LLMProvider
- * @property {string} modelName
- * @property {string} modelVersion
- * @property {(messages: ChatMessage[], options?: LLMCompletionOptions) => Promise<LLMCompletionResult>} complete
- * @property {(messages: ChatMessage[], options?: LLMCompletionOptions) => Promise<Object>} json
- */
-
-/**
- * @typedef {Object} MockLLMConfig
- * @property {'mock'} provider
- * @property {Record<string, Object>} [responses={}]
- */
-
-/**
- * @typedef {Object} AnthropicLLMConfig
- * @property {'anthropic'} provider
- * @property {string} [apiKey]
- * @property {string} [model='claude-sonnet-4-6']
- * @property {number} [maxTokens=1024]
- */
-
-/**
- * @typedef {Object} OpenAILLMConfig
- * @property {'openai'} provider
- * @property {string} [apiKey]
- * @property {string} [model='gpt-4o']
- * @property {number} [maxTokens=1024]
- */
-
-const PROMPT_TYPE_KEYS = [
-  'principleExtraction',
-  'contradictionDetection',
-  'causalArticulation',
-  'contextResolution',
-];
-
-/** @implements {LLMProvider} */
-export class MockLLMProvider {
-  /** @param {Partial<MockLLMConfig>} [config={}] */
-  constructor({ responses = {} } = {}) {
-    this.responses = responses;
-    this.modelName = 'mock-llm';
-    this.modelVersion = '1.0.0';
-  }
-
-  _matchPromptType(messages) {
-    const systemMsg = messages.find(m => m.role === 'system')?.content || '';
-    for (const key of PROMPT_TYPE_KEYS) {
-      if (systemMsg.includes(key)) return key;
-    }
-    return null;
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @returns {Promise<LLMCompletionResult>}
-   */
-  async complete(messages) {
-    const promptType = this._matchPromptType(messages);
-    const cannedResponse = promptType ? this.responses[promptType] : undefined;
-    return { content: cannedResponse !== undefined ? JSON.stringify(cannedResponse) : '{}' };
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @returns {Promise<Object>}
-   */
-  async json(messages) {
-    const promptType = this._matchPromptType(messages);
-    const cannedResponse = promptType ? this.responses[promptType] : undefined;
-    return cannedResponse !== undefined ? cannedResponse : {};
-  }
-}
-
-/** @implements {LLMProvider} */
-export class AnthropicLLMProvider {
-  /** @param {Partial<AnthropicLLMConfig>} [config={}] */
-  constructor({ apiKey, model = 'claude-sonnet-4-6', maxTokens = 1024, timeout = 30000 } = {}) {
-    this.apiKey = apiKey || process.env.ANTHROPIC_API_KEY;
-    this.model = model;
-    this.maxTokens = maxTokens;
-    this.timeout = timeout;
-    this.modelName = model;
-    this.modelVersion = 'latest';
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @param {LLMCompletionOptions} [options={}]
-   * @returns {Promise<LLMCompletionResult>}
-   */
-  async complete(messages, options = {}) {
-    const systemMsg = messages.find(m => m.role === 'system')?.content;
-    const nonSystemMsgs = messages.filter(m => m.role !== 'system');
-
-    const body = {
-      model: this.model,
-      max_tokens: options.maxTokens || this.maxTokens,
-      messages: nonSystemMsgs,
-    };
-    if (systemMsg) body.system = systemMsg;
-
-    const controller = new AbortController();
-    const timer = setTimeout(() => controller.abort(), this.timeout);
-    try {
-      const response = await fetch('https://api.anthropic.com/v1/messages', {
-        method: 'POST',
-        headers: {
-          'x-api-key': this.apiKey,
-          'anthropic-version': '2023-06-01',
-          'content-type': 'application/json',
-        },
-        body: JSON.stringify(body),
-        signal: controller.signal,
-      });
-
-      if (!response.ok) {
-        throw new Error(`Anthropic API error: ${response.status}`);
-      }
-
-      const data = await response.json();
-      const text = data.content?.[0]?.text || '';
-      return { content: text };
-    } finally {
-      clearTimeout(timer);
-    }
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @param {LLMCompletionOptions} [options={}]
-   * @returns {Promise<Object>}
-   */
-  async json(messages, options = {}) {
-    const result = await this.complete(messages, options);
-    try {
-      return JSON.parse(result.content);
-    } catch {
-      throw new Error(`Failed to parse LLM response as JSON: ${result.content.slice(0, 200)}`);
-    }
-  }
-}
-
-/** @implements {LLMProvider} */
-export class OpenAILLMProvider {
-  /** @param {Partial<OpenAILLMConfig>} [config={}] */
-  constructor({ apiKey, model = 'gpt-4o', maxTokens = 1024, timeout = 30000 } = {}) {
-    this.apiKey = apiKey || process.env.OPENAI_API_KEY;
-    this.model = model;
-    this.maxTokens = maxTokens;
-    this.timeout = timeout;
-    this.modelName = model;
-    this.modelVersion = 'latest';
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @param {LLMCompletionOptions} [options={}]
-   * @returns {Promise<LLMCompletionResult>}
-   */
-  async complete(messages, options = {}) {
-    const body = {
-      model: this.model,
-      max_tokens: options.maxTokens || this.maxTokens,
-      messages,
-    };
-
-    const controller = new AbortController();
-    const timer = setTimeout(() => controller.abort(), this.timeout);
-    try {
-      const response = await fetch('https://api.openai.com/v1/chat/completions', {
-        method: 'POST',
-        headers: {
-          'Authorization': `Bearer ${this.apiKey}`,
-          'Content-Type': 'application/json',
-        },
-        body: JSON.stringify(body),
-        signal: controller.signal,
-      });
-
-      if (!response.ok) {
-        throw new Error(`OpenAI API error: ${response.status}`);
-      }
-
-      const data = await response.json();
-      const text = data.choices?.[0]?.message?.content || '';
-      return { content: text };
-    } finally {
-      clearTimeout(timer);
-    }
-  }
-
-  /**
-   * @param {ChatMessage[]} messages
-   * @param {LLMCompletionOptions} [options={}]
-   * @returns {Promise<Object>}
-   */
-  async json(messages, options = {}) {
-    const result = await this.complete(messages, options);
-    try {
-      return JSON.parse(result.content);
-    } catch {
-      throw new Error(`Failed to parse LLM response as JSON: ${result.content.slice(0, 200)}`);
-    }
-  }
-}
-
-/**
- * @param {MockLLMConfig | AnthropicLLMConfig | OpenAILLMConfig} config
- * @returns {MockLLMProvider | AnthropicLLMProvider | OpenAILLMProvider}
- */
-export function createLLMProvider(config) {
-  switch (config.provider) {
-    case 'mock':
-      return new MockLLMProvider(config);
-    case 'anthropic':
-      return new AnthropicLLMProvider(config);
-    case 'openai':
-      return new OpenAILLMProvider(config);
-    default:
-      throw new Error(`Unknown LLM provider: ${config.provider}. Valid: mock, anthropic, openai`);
-  }
-}
+/**
+ * @typedef {Object} ChatMessage
+ * @property {'system' | 'user' | 'assistant'} role
+ * @property {string} content
+ */
+
+function extractJSON(text) {
+  const fenced = text.match(/```(?:json)?\s*\n?([\s\S]*?)```/);
+  return fenced ? fenced[1].trim() : text.trim();
+}
+
+/**
+ * @typedef {Object} LLMCompletionResult
+ * @property {string} content
+ */
+
+/**
+ * @typedef {Object} LLMCompletionOptions
+ * @property {number} [maxTokens]
+ */
+
+/**
+ * @typedef {Object} LLMProvider
+ * @property {string} modelName
+ * @property {string} modelVersion
+ * @property {(messages: ChatMessage[], options?: LLMCompletionOptions) => Promise<LLMCompletionResult>} complete
+ * @property {(messages: ChatMessage[], options?: LLMCompletionOptions) => Promise<Object>} json
+ */
+
+/**
+ * @typedef {Object} MockLLMConfig
+ * @property {'mock'} provider
+ * @property {Record<string, Object>} [responses={}]
+ */
+
+/**
+ * @typedef {Object} AnthropicLLMConfig
+ * @property {'anthropic'} provider
+ * @property {string} [apiKey]
+ * @property {string} [model='claude-sonnet-4-6']
+ * @property {number} [maxTokens=1024]
+ */
+
+/**
+ * @typedef {Object} OpenAILLMConfig
+ * @property {'openai'} provider
+ * @property {string} [apiKey]
+ * @property {string} [model='gpt-4o']
+ * @property {number} [maxTokens=1024]
+ */
+
+const PROMPT_TYPE_KEYS = [
+  'principleExtraction',
+  'contradictionDetection',
+  'causalArticulation',
+  'contextResolution',
+];
+
+/** @implements {LLMProvider} */
+export class MockLLMProvider {
+  /** @param {Partial<MockLLMConfig>} [config={}] */
+  constructor({ responses = {} } = {}) {
+    this.responses = responses;
+    this.modelName = 'mock-llm';
+    this.modelVersion = '1.0.0';
+  }
+
+  _matchPromptType(messages) {
+    const systemMsg = messages.find(m => m.role === 'system')?.content || '';
+    for (const key of PROMPT_TYPE_KEYS) {
+      if (systemMsg.includes(key)) return key;
+    }
+    return null;
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @returns {Promise<LLMCompletionResult>}
+   */
+  async complete(messages) {
+    const promptType = this._matchPromptType(messages);
+    const cannedResponse = promptType ? this.responses[promptType] : undefined;
+    return { content: cannedResponse !== undefined ? JSON.stringify(cannedResponse) : '{}' };
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @returns {Promise<Object>}
+   */
+  async json(messages) {
+    const promptType = this._matchPromptType(messages);
+    const cannedResponse = promptType ? this.responses[promptType] : undefined;
+    return cannedResponse !== undefined ? cannedResponse : {};
+  }
+}
+
+/** @implements {LLMProvider} */
+export class AnthropicLLMProvider {
+  /** @param {Partial<AnthropicLLMConfig>} [config={}] */
+  constructor({ apiKey, model = 'claude-sonnet-4-6', maxTokens = 1024, timeout = 30000 } = {}) {
+    this.apiKey = apiKey || process.env.ANTHROPIC_API_KEY;
+    this.model = model;
+    this.maxTokens = maxTokens;
+    this.timeout = timeout;
+    this.modelName = model;
+    this.modelVersion = 'latest';
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @param {LLMCompletionOptions} [options={}]
+   * @returns {Promise<LLMCompletionResult>}
+   */
+  async complete(messages, options = {}) {
+    const systemMsg = messages.find(m => m.role === 'system')?.content;
+    const nonSystemMsgs = messages.filter(m => m.role !== 'system');
+
+    const body = {
+      model: this.model,
+      max_tokens: options.maxTokens || this.maxTokens,
+      messages: nonSystemMsgs,
+    };
+    if (systemMsg) body.system = systemMsg;
+
+    const controller = new AbortController();
+    const timer = setTimeout(() => controller.abort(), this.timeout);
+    try {
+      const response = await fetch('https://api.anthropic.com/v1/messages', {
+        method: 'POST',
+        headers: {
+          'x-api-key': this.apiKey,
+          'anthropic-version': '2023-06-01',
+          'content-type': 'application/json',
+        },
+        body: JSON.stringify(body),
+        signal: controller.signal,
+      });
+
+      if (!response.ok) {
+        const errorBody = await response.text().catch(() => '');
+        throw new Error(`Anthropic API error: ${response.status} ${errorBody}`);
+      }
+
+      const data = await response.json();
+      const text = data.content?.[0]?.text || '';
+      return { content: text };
+    } finally {
+      clearTimeout(timer);
+    }
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @param {LLMCompletionOptions} [options={}]
+   * @returns {Promise<Object>}
+   */
+  async json(messages, options = {}) {
+    const result = await this.complete(messages, options);
+    try {
+      return JSON.parse(extractJSON(result.content));
+    } catch {
+      throw new Error(`Failed to parse LLM response as JSON: ${result.content.slice(0, 200)}`);
+    }
+  }
+}
+
+/** @implements {LLMProvider} */
+export class OpenAILLMProvider {
+  /** @param {Partial<OpenAILLMConfig>} [config={}] */
+  constructor({ apiKey, model = 'gpt-4o', maxTokens = 1024, timeout = 30000 } = {}) {
+    this.apiKey = apiKey || process.env.OPENAI_API_KEY;
+    this.model = model;
+    this.maxTokens = maxTokens;
+    this.timeout = timeout;
+    this.modelName = model;
+    this.modelVersion = 'latest';
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @param {LLMCompletionOptions} [options={}]
+   * @returns {Promise<LLMCompletionResult>}
+   */
+  async complete(messages, options = {}) {
+    const body = {
+      model: this.model,
+      max_tokens: options.maxTokens || this.maxTokens,
+      messages,
+    };
+
+    const controller = new AbortController();
+    const timer = setTimeout(() => controller.abort(), this.timeout);
+    try {
+      const response = await fetch('https://api.openai.com/v1/chat/completions', {
+        method: 'POST',
+        headers: {
+          'Authorization': `Bearer ${this.apiKey}`,
+          'Content-Type': 'application/json',
+        },
+        body: JSON.stringify(body),
+        signal: controller.signal,
+      });
+
+      if (!response.ok) {
+        throw new Error(`OpenAI API error: ${response.status}`);
+      }
+
+      const data = await response.json();
+      const text = data.choices?.[0]?.message?.content || '';
+      return { content: text };
+    } finally {
+      clearTimeout(timer);
+    }
+  }
+
+  /**
+   * @param {ChatMessage[]} messages
+   * @param {LLMCompletionOptions} [options={}]
+   * @returns {Promise<Object>}
+   */
+  async json(messages, options = {}) {
+    const result = await this.complete(messages, options);
+    try {
+      return JSON.parse(extractJSON(result.content));
+    } catch {
+      throw new Error(`Failed to parse LLM response as JSON: ${result.content.slice(0, 200)}`);
+    }
+  }
+}
+
+/**
+ * @param {MockLLMConfig | AnthropicLLMConfig | OpenAILLMConfig} config
+ * @returns {MockLLMProvider | AnthropicLLMProvider | OpenAILLMProvider}
+ */
+export function createLLMProvider(config) {
+  switch (config.provider) {
+    case 'mock':
+      return new MockLLMProvider(config);
+    case 'anthropic':
+      return new AnthropicLLMProvider(config);
+    case 'openai':
+      return new OpenAILLMProvider(config);
+    default:
+      throw new Error(`Unknown LLM provider: ${config.provider}. Valid: mock, anthropic, openai`);
+  }
+}

package/src/migrate.js

@@ -1,58 +1,58 @@
-import { dropVec0Tables, createVec0Tables } from './db.js';
-
-export async function reembedAll(db, embeddingProvider, { dropAndRecreate = false } = {}) {
-  if (dropAndRecreate) {
-    dropVec0Tables(db);
-    createVec0Tables(db, embeddingProvider.dimensions);
-  }
-
-  const episodes = db.prepare('SELECT id, content, source FROM episodes').all();
-  const semantics = db.prepare('SELECT id, content, state FROM semantics').all();
-  const procedures = db.prepare('SELECT id, content, state FROM procedures').all();
-
-  const episodeVectors = episodes.length > 0
-    ? await embeddingProvider.embedBatch(episodes.map(ep => ep.content))
-    : [];
-  const semanticVectors = semantics.length > 0
-    ? await embeddingProvider.embedBatch(semantics.map(s => s.content))
-    : [];
-  const procedureVectors = procedures.length > 0
-    ? await embeddingProvider.embedBatch(procedures.map(p => p.content))
-    : [];
-
-  const updateEpLegacy = db.prepare('UPDATE episodes SET embedding = ? WHERE id = ?');
-  const deleteVecEp = db.prepare('DELETE FROM vec_episodes WHERE id = ?');
-  const insertVecEp = db.prepare('INSERT INTO vec_episodes(id, embedding, source, consolidated) VALUES (?, ?, ?, ?)');
-
-  const updateSemLegacy = db.prepare('UPDATE semantics SET embedding = ? WHERE id = ?');
-  const deleteVecSem = db.prepare('DELETE FROM vec_semantics WHERE id = ?');
-  const insertVecSem = db.prepare('INSERT INTO vec_semantics(id, embedding, state) VALUES (?, ?, ?)');
-
-  const updateProcLegacy = db.prepare('UPDATE procedures SET embedding = ? WHERE id = ?');
-  const deleteVecProc = db.prepare('DELETE FROM vec_procedures WHERE id = ?');
-  const insertVecProc = db.prepare('INSERT INTO vec_procedures(id, embedding, state) VALUES (?, ?, ?)');
-
-  const writeTx = db.transaction(() => {
-    for (let i = 0; i < episodes.length; i++) {
-      const buf = embeddingProvider.vectorToBuffer(episodeVectors[i]);
-      updateEpLegacy.run(buf, episodes[i].id);
-      deleteVecEp.run(episodes[i].id);
-      insertVecEp.run(episodes[i].id, buf, episodes[i].source, BigInt(0));
-    }
-    for (let i = 0; i < semantics.length; i++) {
-      const buf = embeddingProvider.vectorToBuffer(semanticVectors[i]);
-      updateSemLegacy.run(buf, semantics[i].id);
-      deleteVecSem.run(semantics[i].id);
-      insertVecSem.run(semantics[i].id, buf, semantics[i].state);
-    }
-    for (let i = 0; i < procedures.length; i++) {
-      const buf = embeddingProvider.vectorToBuffer(procedureVectors[i]);
-      updateProcLegacy.run(buf, procedures[i].id);
-      deleteVecProc.run(procedures[i].id);
-      insertVecProc.run(procedures[i].id, buf, procedures[i].state);
-    }
-  });
-  writeTx();
-
-  return { episodes: episodes.length, semantics: semantics.length, procedures: procedures.length };
-}
+import { dropVec0Tables, createVec0Tables } from './db.js';
+
+export async function reembedAll(db, embeddingProvider, { dropAndRecreate = false } = {}) {
+  if (dropAndRecreate) {
+    dropVec0Tables(db);
+    createVec0Tables(db, embeddingProvider.dimensions);
+  }
+
+  const episodes = db.prepare('SELECT id, content, source FROM episodes').all();
+  const semantics = db.prepare('SELECT id, content, state FROM semantics').all();
+  const procedures = db.prepare('SELECT id, content, state FROM procedures').all();
+
+  const episodeVectors = episodes.length > 0
+    ? await embeddingProvider.embedBatch(episodes.map(ep => ep.content))
+    : [];
+  const semanticVectors = semantics.length > 0
+    ? await embeddingProvider.embedBatch(semantics.map(s => s.content))
+    : [];
+  const procedureVectors = procedures.length > 0
+    ? await embeddingProvider.embedBatch(procedures.map(p => p.content))
+    : [];
+
+  const updateEpLegacy = db.prepare('UPDATE episodes SET embedding = ? WHERE id = ?');
+  const deleteVecEp = db.prepare('DELETE FROM vec_episodes WHERE id = ?');
+  const insertVecEp = db.prepare('INSERT INTO vec_episodes(id, embedding, source, consolidated) VALUES (?, ?, ?, ?)');
+
+  const updateSemLegacy = db.prepare('UPDATE semantics SET embedding = ? WHERE id = ?');
+  const deleteVecSem = db.prepare('DELETE FROM vec_semantics WHERE id = ?');
+  const insertVecSem = db.prepare('INSERT INTO vec_semantics(id, embedding, state) VALUES (?, ?, ?)');
+
+  const updateProcLegacy = db.prepare('UPDATE procedures SET embedding = ? WHERE id = ?');
+  const deleteVecProc = db.prepare('DELETE FROM vec_procedures WHERE id = ?');
+  const insertVecProc = db.prepare('INSERT INTO vec_procedures(id, embedding, state) VALUES (?, ?, ?)');
+
+  const writeTx = db.transaction(() => {
+    for (let i = 0; i < episodes.length; i++) {
+      const buf = embeddingProvider.vectorToBuffer(episodeVectors[i]);
+      updateEpLegacy.run(buf, episodes[i].id);
+      deleteVecEp.run(episodes[i].id);
+      insertVecEp.run(episodes[i].id, buf, episodes[i].source, BigInt(0));
+    }
+    for (let i = 0; i < semantics.length; i++) {
+      const buf = embeddingProvider.vectorToBuffer(semanticVectors[i]);
+      updateSemLegacy.run(buf, semantics[i].id);
+      deleteVecSem.run(semantics[i].id);
+      insertVecSem.run(semantics[i].id, buf, semantics[i].state);
+    }
+    for (let i = 0; i < procedures.length; i++) {
+      const buf = embeddingProvider.vectorToBuffer(procedureVectors[i]);
+      updateProcLegacy.run(buf, procedures[i].id);
+      deleteVecProc.run(procedures[i].id);
+      insertVecProc.run(procedures[i].id, buf, procedures[i].state);
+    }
+  });
+  writeTx();
+
+  return { episodes: episodes.length, semantics: semantics.length, procedures: procedures.length };
+}