nttp 1.4.11 → 1.4.13

package/README.md

@@ -85,17 +85,17 @@ Most queries hit L1 or L2. Only novel queries reach the LLM.
 ## Documentation
 
 - **Quick Reference:** `npx nttp docs [topic]`
-- **Full Guides:** [/docs](/docs)
+- **Full Guides:** [/docs](https://github.com/Tylerbryy/nttp_typescript/tree/main/packages/nttp/docs)
 
 ### Guides
 
-- [API Reference](docs/api.md) - Complete API documentation
-- [Caching System](docs/caching.md) - 3-layer cache deep dive
-- [Configuration](docs/configuration.md) - All config options
-- [LLM Models](docs/models.md) - Model selection guide
-- [Examples](docs/examples.md) - Usage examples
-- [Production](docs/production.md) - Deployment best practices
-- [Troubleshooting](docs/troubleshooting.md) - Common issues
+- [API Reference](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/api.md) - Complete API documentation
+- [Caching System](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/caching.md) - 3-layer cache deep dive
+- [Configuration](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/configuration.md) - All config options
+- [LLM Models](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/models.md) - Model selection guide
+- [Examples](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/examples.md) - Usage examples
+- [Production](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/production.md) - Deployment best practices
+- [Troubleshooting](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/troubleshooting.md) - Common issues
 
 ---
 
@@ -162,7 +162,7 @@ const nttp = new NTTP({
 await nttp.init();
 ```
 
-See [Configuration Guide](docs/configuration.md) for all options.
+See [Configuration Guide](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/configuration.md) for all options.
 
 ---
 
@@ -194,7 +194,7 @@ See [Configuration Guide](docs/configuration.md) for all options.
 - **Mistral** - Open-source preference
 - **Google Gemini** - Multimodal capabilities
 
-See [Model Selection Guide](docs/models.md) for detailed comparison.
+See [Model Selection Guide](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/models.md) for detailed comparison.
 
 ---
 
@@ -246,7 +246,7 @@ await nttp.pinSchema(schemaId);
 await nttp.close();
 ```
 
-See [API Reference](docs/api.md) for complete documentation.
+See [API Reference](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/api.md) for complete documentation.
 
 ---
 
@@ -269,7 +269,7 @@ try {
 }
 ```
 
-All errors include helpful suggestions. See [Troubleshooting Guide](docs/troubleshooting.md).
+All errors include helpful suggestions. See [Troubleshooting Guide](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/troubleshooting.md).
 
 ---
 
@@ -291,19 +291,19 @@ import type {
 
 ## Examples
 
-- [Basic Usage](docs/examples.md#basic-queries)
-- [Express Integration](docs/examples.md#using-with-express)
-- [Next.js Integration](docs/examples.md#using-with-nextjs)
-- [CLI Tools](docs/examples.md#cli-integration)
+- [Basic Usage](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/examples.md#basic-queries)
+- [Express Integration](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/examples.md#using-with-express)
+- [Next.js Integration](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/examples.md#using-with-nextjs)
+- [CLI Tools](https://github.com/Tylerbryy/nttp_typescript/blob/main/packages/nttp/docs/examples.md#cli-integration)
 
 ---
 
 ## Links
 
-- [GitHub](https://github.com/tylergibbs/nttp)
+- [GitHub](https://github.com/Tylerbryy/nttp_typescript)
 - [npm](https://www.npmjs.com/package/nttp)
-- [Issues](https://github.com/tylergibbs/nttp/issues)
-- [Documentation](/docs)
+- [Issues](https://github.com/Tylerbryy/nttp_typescript/issues)
+- [Documentation](https://github.com/Tylerbryy/nttp_typescript/tree/main/packages/nttp/docs)
 
 ---
 

package/dist/cli.js

@@ -11,7 +11,7 @@ const program = new Command();
 program
     .name('nttp')
     .description('Query databases with natural language')
-    .version('1.4.11');
+    .version('1.4.13');
 program
     .command('setup')
     .description('Interactive setup wizard (or use --non-interactive for agents)')

package/docs/README.md

@@ -0,0 +1,31 @@
+# NTTP Documentation
+
+Complete documentation for Natural Text Transfer Protocol.
+
+## Quick Links
+
+- [API Reference](./api.md) - Complete API documentation
+- [Caching System](./caching.md) - 3-layer cache architecture
+- [Configuration](./configuration.md) - All configuration options
+- [LLM Models](./models.md) - Model selection guide
+- [Examples](./examples.md) - Comprehensive usage examples
+- [Production Guide](./production.md) - Best practices and deployment
+- [Troubleshooting](./troubleshooting.md) - Common issues and solutions
+
+## Getting Started
+
+If you're new to NTTP, start here:
+
+1. **Installation**: `npm install nttp`
+2. **Setup**: `npx nttp setup`
+3. **First Query**: `npx nttp query "show me 5 users"`
+
+## Documentation Versions
+
+- **CLI Docs**: Run `npx nttp docs [topic]` for searchable inline documentation
+- **Full Guides**: Browse this `/docs` folder for comprehensive guides
+- **Quick Reference**: See main [README.md](../README.md) for quick start
+
+## Contributing
+
+Found an issue or want to improve the docs? [Open an issue](https://github.com/tylergibbs/nttp/issues) or submit a PR.

package/docs/api.md

@@ -0,0 +1,559 @@
+# API Reference
+
+Complete API documentation for NTTP.
+
+## Table of Contents
+
+- [NTTP Class](#nttp-class)
+  - [Constructor](#constructor)
+  - [Static Methods](#static-methods)
+  - [Instance Methods](#instance-methods)
+- [Types](#types)
+- [Errors](#errors)
+
+---
+
+## NTTP Class
+
+### Constructor
+
+#### `new NTTP(config: NTTPConfig)`
+
+Creates a new NTTP instance with manual configuration.
+
+**Parameters:**
+
+```typescript
+interface NTTPConfig {
+  database: DatabaseConfig;
+  llm: LLMConfig;
+  cache?: CacheConfig;
+  limits?: LimitsConfig;
+}
+```
+
+**Example:**
+
+```typescript
+import { NTTP } from 'nttp';
+
+const nttp = new NTTP({
+  database: {
+    client: 'pg',
+    connection: 'postgresql://user:pass@localhost:5432/mydb'
+  },
+  llm: {
+    provider: 'anthropic',
+    model: 'claude-sonnet-4-5-20250929',
+    apiKey: process.env.ANTHROPIC_API_KEY
+  },
+  cache: {
+    l1: { enabled: true, maxSize: 1000 },
+    l2: { enabled: true, provider: 'openai', apiKey: process.env.OPENAI_API_KEY },
+    redis: { url: 'redis://localhost:6379' }
+  }
+});
+```
+
+See [Configuration](./configuration.md) for all config options.
+
+---
+
+### Static Methods
+
+#### `NTTP.fromEnv(): Promise<NTTP>`
+
+Creates an NTTP instance from environment variables. This is the recommended way to initialize NTTP.
+
+**Environment Variables Required:**
+
+```bash
+DATABASE_TYPE=pg
+DATABASE_URL=postgresql://user:pass@localhost:5432/mydb
+LLM_PROVIDER=anthropic
+LLM_MODEL=claude-sonnet-4-5-20250929
+ANTHROPIC_API_KEY=sk-ant-...
+```
+
+**Optional Environment Variables:**
+
+```bash
+REDIS_URL=redis://localhost:6379
+OPENAI_API_KEY=sk-...  # For L2 semantic cache
+```
+
+**Example:**
+
+```typescript
+import { NTTP } from 'nttp';
+
+// Automatically reads from .env file (using dotenv)
+const nttp = await NTTP.fromEnv();
+const result = await nttp.query("show me users");
+await nttp.close();
+```
+
+**Returns:** `Promise<NTTP>` - Initialized NTTP instance (already called `init()`)
+
+**Throws:** `Error` if required environment variables are missing
+
+See [Configuration](./configuration.md#environment-variables) for complete env var reference.
+
+---
+
+### Instance Methods
+
+#### `init(): Promise<void>`
+
+Initializes the NTTP instance by connecting to the database and building the schema cache.
+
+**Required:** Must be called before any queries (unless using `fromEnv()`, which calls this automatically).
+
+**Example:**
+
+```typescript
+const nttp = new NTTP({ /* config */ });
+await nttp.init();  // Connect to database, build schema cache
+```
+
+**Throws:** `Error` if database connection fails
+
+---
+
+#### `query(query: string, options?: QueryOptions): Promise<QueryResult>`
+
+Execute a natural language query against the database.
+
+**Parameters:**
+
+- `query` (string): Natural language query (e.g., "show me active users")
+- `options` (optional):
+  ```typescript
+  interface QueryOptions {
+    useCache?: boolean;       // Default: true
+    forceNewSchema?: boolean; // Default: false
+  }
+  ```
+
+**Returns:**
+
+```typescript
+interface QueryResult {
+  query: string;           // Original natural language query
+  data: any[];            // Query results (array of objects)
+  schemaId: string;       // Cache key (16-char hash)
+  cacheHit: boolean;      // true if result came from cache
+  intent: Intent;         // Parsed intent structure
+  sql?: string;           // Generated SQL (for debugging)
+  params?: any[];         // SQL parameters (for debugging)
+  meta?: {                // Cache metadata (v2 only)
+    cacheLayer: 1 | 2 | 3;  // Which cache layer was hit
+    cost: number;           // Estimated cost in USD
+    latency: number;        // Query latency in ms
+    similarity?: number;    // Semantic similarity (L2 only)
+  };
+}
+```
+
+**Examples:**
+
+```typescript
+// Basic query
+const result = await nttp.query("show me 5 users");
+console.log(result.data); // Array of user objects
+
+// Filtered query
+const active = await nttp.query("active users from California");
+
+// Aggregation
+const count = await nttp.query("count pending orders");
+
+// With options
+const fresh = await nttp.query("show products", {
+  useCache: false,        // Skip cache, always generate new SQL
+  forceNewSchema: true    // Force new schema generation
+});
+
+// Check cache performance
+const result = await nttp.query("show users");
+if (result.meta) {
+  console.log(`Cache: L${result.meta.cacheLayer}, Cost: $${result.meta.cost}, Latency: ${result.meta.latency}ms`);
+}
+```
+
+**Throws:**
+- `IntentParseError` - Failed to parse natural language query
+- `SQLGenerationError` - Failed to generate SQL from intent
+- `SQLExecutionError` - Database query execution failed
+
+See [Examples](./examples.md) for comprehensive query examples.
+
+---
+
+#### `explain(query: string): Promise<ExplanationResult>`
+
+Explains what SQL would be generated for a query without executing it. Useful for debugging and understanding how NTTP interprets queries.
+
+**Parameters:**
+
+- `query` (string): Natural language query to explain
+
+**Returns:**
+
+```typescript
+interface ExplanationResult {
+  query: string;                    // Original query
+  intent: Intent;                   // Parsed intent
+  sql: string;                      // Generated SQL
+  params: any[];                    // SQL parameters
+  schemaId: string;                 // Cache key
+  cachedSchema: SchemaDefinition | null;  // Cached schema if exists
+}
+```
+
+**Example:**
+
+```typescript
+const explanation = await nttp.explain("top 10 expensive products");
+
+console.log('Intent:', explanation.intent);
+// { entity: 'products', operation: 'list', sort: 'price:desc', limit: 10, ... }
+
+console.log('SQL:', explanation.sql);
+// SELECT * FROM products ORDER BY price DESC LIMIT ?
+
+console.log('Params:', explanation.params);
+// [10]
+```
+
+---
+
+#### `close(): Promise<void>`
+
+Closes the database connection and cleans up resources. Always call this when done using NTTP.
+
+**Example:**
+
+```typescript
+const nttp = await NTTP.fromEnv();
+try {
+  const result = await nttp.query("show users");
+  console.log(result.data);
+} finally {
+  await nttp.close();  // Clean up
+}
+```
+
+---
+
+### Schema Management
+
+#### `listSchemas(): Promise<SchemaDefinition[]>`
+
+Lists all cached schemas (query patterns).
+
+**Returns:** Array of schema definitions with metadata
+
+**Example:**
+
+```typescript
+const schemas = await nttp.listSchemas();
+schemas.forEach(schema => {
+  console.log(`${schema.schema_id}: ${schema.intent_pattern}`);
+  console.log(`  Used ${schema.use_count} times`);
+  console.log(`  Example: ${schema.example_queries[0]}`);
+});
+```
+
+---
+
+#### `getSchema(schemaId: string): Promise<SchemaDefinition | null>`
+
+Retrieves a specific cached schema by ID.
+
+**Parameters:**
+
+- `schemaId` (string): 16-character schema hash
+
+**Returns:** Schema definition or null if not found
+
+**Example:**
+
+```typescript
+const schema = await nttp.getSchema('a1b2c3d4e5f6g7h8');
+if (schema) {
+  console.log('SQL:', schema.generated_sql);
+  console.log('Used:', schema.use_count);
+}
+```
+
+---
+
+#### `deleteSchema(schemaId: string): Promise<void>`
+
+Deletes a cached schema.
+
+**Parameters:**
+
+- `schemaId` (string): Schema to delete
+
+**Example:**
+
+```typescript
+await nttp.deleteSchema('a1b2c3d4e5f6g7h8');
+```
+
+---
+
+#### `pinSchema(schemaId: string): Promise<void>`
+
+Pins a schema to prevent it from being evicted from cache.
+
+**Parameters:**
+
+- `schemaId` (string): Schema to pin
+
+**Example:**
+
+```typescript
+// Pin frequently used query pattern
+await nttp.pinSchema('a1b2c3d4e5f6g7h8');
+```
+
+---
+
+#### `unpinSchema(schemaId: string): Promise<void>`
+
+Unpins a previously pinned schema.
+
+**Parameters:**
+
+- `schemaId` (string): Schema to unpin
+
+---
+
+#### `getCacheStats(): Promise<CacheStats>`
+
+Gets cache statistics and performance metrics.
+
+**Returns:**
+
+```typescript
+interface CacheStats {
+  totalSchemas: number;
+  pinnedSchemas: number;
+  totalUseCount: number;
+  averageUseCount: number;
+  oldestSchema: Date;
+  newestSchema: Date;
+}
+```
+
+**Example:**
+
+```typescript
+const stats = await nttp.getCacheStats();
+console.log(`Total cached patterns: ${stats.totalSchemas}`);
+console.log(`Average uses per pattern: ${stats.averageUseCount}`);
+```
+
+---
+
+### Database Inspection
+
+#### `getTables(): Promise<string[]>`
+
+Gets list of all tables in the database.
+
+**Returns:** Array of table names
+
+**Example:**
+
+```typescript
+const tables = await nttp.getTables();
+console.log('Available tables:', tables);
+// ['users', 'products', 'orders', ...]
+```
+
+---
+
+#### `getTableSchema(tableName: string): Promise<TableSchema>`
+
+Gets the schema (columns, types, etc.) for a specific table.
+
+**Parameters:**
+
+- `tableName` (string): Name of table to inspect
+
+**Returns:** Table schema with column definitions
+
+**Example:**
+
+```typescript
+const schema = await nttp.getTableSchema('users');
+console.log('Columns:', schema.columns);
+// [{ name: 'id', type: 'integer' }, { name: 'email', type: 'string' }, ...]
+```
+
+---
+
+#### `getSchemaDescription(): string`
+
+Gets a human-readable description of the database schema. This is what's sent to the LLM for context.
+
+**Returns:** Formatted schema description string
+
+**Example:**
+
+```typescript
+const description = nttp.getSchemaDescription();
+console.log(description);
+// Tables:
+// - users (id: integer, email: string, name: string, ...)
+// - products (id: integer, name: string, price: decimal, ...)
+```
+
+---
+
+## Types
+
+### Intent
+
+Structured representation of parsed natural language query.
+
+```typescript
+interface Intent {
+  entity: string;                    // Table name
+  operation: OperationType;          // 'list' | 'count' | 'aggregate' | 'filter'
+  filters: Record<string, any>;      // Filter conditions
+  limit?: number | null;             // Result limit
+  fields?: string[] | null;          // Specific fields to select
+  sort?: string | null;              // Sort specification (e.g., 'price:desc')
+  normalized_text: string;           // Normalized representation for caching
+}
+```
+
+---
+
+### SchemaDefinition
+
+Cached query pattern.
+
+```typescript
+interface SchemaDefinition {
+  schema_id: string;                 // 16-char hash
+  intent_pattern: string;            // Normalized intent
+  generated_sql: string;             // Cached SQL
+  sql_params: any[];                 // SQL parameters
+  result_schema: Record<string, any>; // Result structure
+  use_count: number;                 // Times this pattern was used
+  created_at: Date;                  // When cached
+  last_used_at: Date;                // Last access time
+  example_queries: string[];         // Example natural language queries
+  pinned: boolean;                   // Prevent eviction?
+}
+```
+
+---
+
+## Errors
+
+All NTTP errors include helpful suggestions for resolving issues.
+
+### IntentParseError
+
+Thrown when the LLM cannot parse the natural language query.
+
+```typescript
+try {
+  await nttp.query("ambiguous query");
+} catch (error) {
+  if (error instanceof IntentParseError) {
+    console.error(error.message);
+    console.log('Suggestions:', error.suggestions);
+  }
+}
+```
+
+**Common causes:**
+- Query references unknown tables/fields
+- Query is too vague or ambiguous
+- LLM API unavailable or quota exceeded
+
+---
+
+### SQLGenerationError
+
+Thrown when SQL generation fails or violates safety rules.
+
+```typescript
+try {
+  await nttp.query("complex query");
+} catch (error) {
+  if (error instanceof SQLGenerationError) {
+    console.error(error.message);
+    console.log('Suggestions:', error.suggestions);
+  }
+}
+```
+
+**Common causes:**
+- Complex query requires missing table relationships
+- Generated SQL violates safety rules
+- Schema description incomplete
+
+---
+
+### SQLExecutionError
+
+Thrown when the database rejects the generated SQL.
+
+```typescript
+try {
+  await nttp.query("show users");
+} catch (error) {
+  if (error instanceof SQLExecutionError) {
+    console.error(error.message);
+    console.log('Generated SQL:', error.sql);
+    console.log('Suggestions:', error.suggestions);
+  }
+}
+```
+
+**Common causes:**
+- Database connection issues
+- Table/column doesn't exist (schema mismatch)
+- Type mismatch in WHERE clause
+- Insufficient permissions
+
+---
+
+### LLMError
+
+Thrown when LLM API calls fail.
+
+**Common causes:**
+- Invalid or expired API key
+- Rate limit or quota exceeded
+- Network connectivity issues
+- Provider service outage
+
+---
+
+### CacheError
+
+Thrown when cache operations fail.
+
+**Common causes:**
+- Redis connection failed
+- Redis authentication error
+- Embedding API failure (L2 cache)
+- Out of memory
+
+---
+
+## See Also
+
+- [Configuration](./configuration.md) - Complete configuration reference
+- [Examples](./examples.md) - Comprehensive usage examples
+- [Troubleshooting](./troubleshooting.md) - Common issues and solutions