wrekenfile-converter
v2.2.5
Published
Convert OpenAPI and Postman specs into Wrekenfiles, with chunking for vector database storage
Downloads
408
Maintainers
Readme
Wrekenfile Converter
A comprehensive TypeScript/JavaScript library for converting OpenAPI specifications (v2 and v3) and Postman collections into Wrekenfiles, declarative YAML artifacts that act as the single source of truth for API methods, workflows, headers, and responses. Generated Wrekenfiles are compliant with the Wreken Specification v2.0.2 and support advanced mini-chunking for vector database storage and AI context management.
Features
- Multi-format Support: Convert OpenAPI v2 (Swagger), OpenAPI v3, and Postman collections
- Wrekenfile v2.0.2 Compliant: Generates Wrekenfiles compliant with the Wreken Specification v2.0.2 (latest)
- Complete Response Handling: All response types (success and error) included in
RETURNSarrays withSTATUScodes - Proper Parameter Structure: Path parameters in
ENDPOINT, header parameters inHTTP.HEADERS, query and body parameters inINPUTSwithLOCATIONfield - HTTP Details: Includes
HTTP.CONTENT_TYPE,HTTP.ACCEPT,HTTP.BODY.TYPE, andHTTP.HEADERSfor complete HTTP execution context - Execution Metadata: Includes
EXECUTION.KIND(http/sdk/hybrid) andEXECUTION.MODE(sync/async/fire_and_forget) - Authentication Handling: Auth headers (Authorization, X-API-Key, etc.) are properly mapped to
HEADERSwith placeholder values - Comprehensive Error Handling: Detailed error messages with context and error codes for invalid inputs, all errors include
STATUScodes - AI-Optimized: Response structs explicitly referenced for easy AI consumption
- Standalone Mini Wrekenfiles: Generate execution-complete, standalone mini-wrekenfiles (one per method) for vector DB storage and LLM code generation
- TypeScript Support: Full TypeScript definitions and exports
- Subproject Ready: Designed to work as a dependency in larger projects
Request-body struct invariants (v2)
For OpenAPI v2/v3 and Postman converters in v2:
- Any schema used as a JSON request body is always emitted as a top-level entry in
STRUCTS, keyed by the schema name. - The corresponding method’s
INPUTSandHTTP.BODY.TYPEalways useTYPE: STRUCT(<SchemaName>)with exactly the same name as theSTRUCTSkey. - The
STRUCTSmap is always flat (no grouping by tag/path/namespace), so downstream tools (mini-wrekenfile generator, CLI resolvers, SDKs) can rely purely onSTRUCT(...),[]STRUCT(...), andmap[...]STRUCT(...)references without special casing.
Installation
As a Dependency (Recommended)
npm install wrekenfile-converteror
yarn add wrekenfile-converterUsage
Version Support
This library supports two Wrekenfile versions:
- v1 based on Wreken Specification 1.2 (legacy)
- v2 based on Wreken Specification 2.0.2 (latest)
Importing the Library
Default import (v1, for backward compatibility):
import {
generateWrekenfile, // OpenAPI v3
generateWrekenfileV2, // OpenAPI v2 (Swagger)
generateWrekenfileFromPostman, // Postman collections
validateWrekenfile,
generateMiniWrekenfiles,
MiniWrekenfile,
ValidationResult
} from 'wrekenfile-converter';Recommended: Use v2 (latest Wrekenfile spec 2.0.2):
import {
generateWrekenfile, // OpenAPI v3
generateWrekenfileV2, // OpenAPI v2 (Swagger)
generateWrekenfileFromPostman, // Postman collections
generateMiniWrekenfiles,
MiniWrekenfile
} from 'wrekenfile-converter/v2';Explicit version imports:
// Import v1 (Wrekenfile spec 1.2)
import { generateWrekenfile } from 'wrekenfile-converter/v1';
// Import v2 (Wrekenfile spec 2.0.2) - Recommended
import { generateWrekenfile } from 'wrekenfile-converter/v2';Convert OpenAPI v3 to Wrekenfile
import fs from 'fs';
import yaml from 'js-yaml';
import { generateWrekenfile } from 'wrekenfile-converter/v2';
const fileContent = fs.readFileSync('./openapi.yaml', 'utf8');
const openapiSpec = yaml.load(fileContent);
const wrekenfileYaml = generateWrekenfile(openapiSpec, './');Convert OpenAPI v2 (Swagger) to Wrekenfile
import fs from 'fs';
import yaml from 'js-yaml';
import { generateWrekenfile } from 'wrekenfile-converter/v2';
const fileContent = fs.readFileSync('./swagger.yaml', 'utf8');
const swaggerSpec = yaml.load(fileContent);
const wrekenfileYaml = generateWrekenfile(swaggerSpec, './');Convert Postman Collection to Wrekenfile
import fs from 'fs';
import { generateWrekenfileFromPostman } from 'wrekenfile-converter/v2';
const collection = JSON.parse(fs.readFileSync('./collection.json', 'utf8'));
const variables = {}; // Optionally provide Postman environment variables
const wrekenfileYaml = generateWrekenfileFromPostman(collection, variables);Validate a Wrekenfile
import { validateWrekenfile } from 'wrekenfile-converter';
const result = validateWrekenfile('./Wrekenfile.yaml');
console.log(result.isValid ? 'Valid' : 'Invalid');
console.log(result.errors, result.warnings);Generate Mini Wrekenfiles
import { generateMiniWrekenfiles, MiniWrekenfile } from 'wrekenfile-converter/v2';
const wrekenfileYaml = generateWrekenfile(openapiSpec, './');
const miniFiles: MiniWrekenfile[] = generateMiniWrekenfiles(wrekenfileYaml);
// Each miniFile contains { content, metadata }
// One mini-wrekenfile is generated per method (standalone, execution-complete)Note: Mini-wrekenfiles are now standalone and execution-complete, meaning each mini-wrekenfile contains all necessary information (HTTP details, SDK details, structs, inputs with LOCATION, returns, errors) for an LLM to generate execution code without external references. They follow the Unified Mini-Wrekenfile Specification v2.0. Only structs that are actually referenced via STRUCT(...), []STRUCT(...), or map[...]STRUCT(...) in the main Wrekenfile are included; no additional schema-name inference is performed at mini-generation time.
CANONICAL_ID (Stable Method Identifier)
Generated Wrekenfiles use CANONICAL_ID as the method key in the METHODS section, providing a stable, semantic identifier for each API method.
- Method Keys: Methods are keyed by their
CANONICAL_ID(e.g.,METHODS.api.cluster.get) - Format:
<namespace>.<resource>.<action>(example:api.cluster.get) - Deterministic: derived from
HTTP.METHOD+HTTP.ENDPOINT(no LLM) - No path params:
{id}etc. are excluded from the identifier - Collision-safe: if two methods would produce the same canonical ID, the generator deterministically appends a short hash to keep IDs unique
- RETURNVAR: Return variable names are derived from
CANONICAL_ID(dots replaced with underscores, e.g.,api_cluster_getfor status 200,api_cluster_get_404for error responses) - Struct Names: Inline request/response struct names are aligned with
CANONICAL_ID(e.g.,api.cluster.getRequest,api.cluster.getResponse200) - Struct Filtering: Unused struct definitions are automatically removed from the generated Wrekenfile
Example:
METHODS:
api.cluster.get:
CANONICAL_ID: api.cluster.get
HTTP:
METHOD: GET
ENDPOINT: /api/clusters/{id}
RETURNS:
- TYPE: STRUCT(api.cluster.getResponse200)
RETURNVAR: api_cluster_get
STATUS: 200The full deterministic spec (for re-implementing in SDKs/other languages) is documented in the source at src/v2/utils/canonical-id.ts.
Parameter Structure
The converter properly structures parameters according to the Wrekenfile v2.0.2 specification:
- Path Parameters (e.g.,
/users/{userId}): Included in theENDPOINTfield only, not inINPUTS - Header Parameters (e.g.,
Authorization,X-Request-Id): Included inHTTP.HEADERSonly, not inINPUTS - Query Parameters: Included in
INPUTSsection withLOCATION: query - Body Parameters: Included in
INPUTSsection withLOCATION: body(e.g.,body: STRUCT(RequestType))
All input parameters include a LOCATION field to clearly indicate where they should be placed in the HTTP request. The HTTP.BODY.TYPE field indicates the content type for body parameters (e.g., application/json, application/x-www-form-urlencoded).
This ensures proper separation of concerns and follows the Wrekenfile specification correctly.
Error Handling
The converters include comprehensive error handling with detailed error messages:
- Validation Errors: Invalid specifications are caught early with descriptive error messages
- Error Codes: Structured error codes (e.g.,
INVALID_SPEC_TYPE,MISSING_OPENAPI_VERSION) for programmatic handling - Context Logging: Errors include full context (spec title, version, file paths, etc.) for debugging
- Stack Traces: Full stack traces are logged for development and debugging
- Status Codes: All errors in the generated Wrekenfile include
STATUScodes (e.g.,400,404,500)
All errors are logged to the console with timestamps and context information. The generated Wrekenfiles include STATUS codes in both RETURNS and ERRORS sections for complete HTTP response handling.
CLI Tools
The CLI tools are available for both v1 and v2. The examples below use v2 (latest), which generates Wrekenfile spec 2.0.2.
Convert OpenAPI v3 to Wrekenfile
Generate a Wrekenfile YAML from an OpenAPI v3 (YAML or JSON) spec:
npx ts-node src/v2/cli/cli-openapi-to-wrekenfile.ts --input <openapi.yaml|json> [--output <wrekenfile.yaml>] [--cwd <dir>]Options:
--inputor-i: Path to your OpenAPI v3 YAML or JSON file (required)--outputor-o: Path to output Wrekenfile YAML (optional, defaults tooutput_wrekenfile.yaml)--cwd: Working directory for resolving $refs (optional, defaults to the input file's directory)
Example:
npx ts-node src/v2/cli/cli-openapi-to-wrekenfile.ts --input examples/3n.yaml --output 3n_wrekenfile_v2.yamlConvert OpenAPI v2 (Swagger) to Wrekenfile
Generate a Wrekenfile YAML from an OpenAPI v2/Swagger (YAML or JSON) spec:
npx ts-node src/v2/cli/cli-openapi-v2-to-wrekenfile.ts --input <swagger.yaml|json> [--output <wrekenfile.yaml>] [--cwd <dir>]Options:
--inputor-i: Path to your OpenAPI v2/Swagger YAML or JSON file (required)--outputor-o: Path to output Wrekenfile YAML (optional, defaults tooutput_wrekenfile.yaml)--cwd: Working directory for resolving $refs (optional, defaults to the input file's directory)
Example:
npx ts-node src/v2/cli/cli-openapi-v2-to-wrekenfile.ts --input examples/5n_v2.yaml --output 5n_v2_wrekenfile.yamlConvert Postman Collection to Wrekenfile
Convert a Postman collection JSON to a Wrekenfile YAML file:
npx ts-node src/v2/cli/cli-postman-to-wrekenfile.ts <postman_collection.json> <output_wrekenfile.yaml> [postman_environment.json]Arguments:
postman_collection.json: Path to your Postman collection JSON file (required)output_wrekenfile.yaml: Path to output Wrekenfile YAML (required)postman_environment.json: Path to Postman environment file (optional)
Example:
npx ts-node src/v2/cli/cli-postman-to-wrekenfile.ts examples/Nium\ APIpostman_collection.json nium_wrekenfile_v2.yamlGenerate Mini Wrekenfiles
Generate standalone, execution-complete mini Wrekenfiles (one per method) from a main Wrekenfile YAML:
npx ts-node src/v2/cli/cli-mini-wrekenfile-generator.ts --input <wrekenfile.yaml> [--output <dir>]Options:
--inputor-i: Path to your main Wrekenfile YAML (required)--outputor-o: Output directory for mini Wrekenfiles (optional, defaults to./mini-wrekenfiles-v2)
Example:
npx ts-node src/v2/cli/cli-mini-wrekenfile-generator.ts --input wrekenfile.yaml --output ./mini-wrekenfiles-v2This will generate one standalone mini Wrekenfile per method in the specified output directory. Each mini-wrekenfile is execution-complete and includes all necessary details (HTTP, SDK, INPUTS with LOCATION, RETURNS, ERRORS, STRUCTS) for LLM code generation without external references.
API Reference
Core Functions
generateWrekenfile(spec: any, baseDir: string): string
Convert OpenAPI v3 specification to Wrekenfile format.
generateWrekenfileV2(spec: any, baseDir: string): string
Convert OpenAPI v2 (Swagger) specification to Wrekenfile format.
generateWrekenfileFromPostman(collection: any, variables: Record<string, string>): string
Convert Postman collection to Wrekenfile format.
validateWrekenfile(filePath: string): ValidationResult
Validate a Wrekenfile and return detailed results.
generateMiniWrekenfiles(wrekenfileContent: string): MiniWrekenfile[]
Generate standalone, execution-complete mini Wrekenfiles (one per method). Takes the full Wrekenfile YAML content as a string and returns an array of mini-wrekenfiles.
Types
MiniWrekenfile
interface MiniWrekenfile {
content: string; // Complete YAML content (standalone, execution-complete)
metadata: {
endpoint?: string; // API endpoint path (if applicable)
interface?: string; // SDK interface name (if applicable)
source?: string; // Source identifier
methods: string[]; // HTTP methods (GET, POST, etc.)
structs: string[]; // Required struct names (included in mini-wrekenfile)
filename: string; // Generated filename
};
}Note: Each mini-wrekenfile is standalone and execution-complete, containing:
METHODsection withID,SUMMARY, andDESCEXECUTIONsection withKIND,MODE, andEXECUTION_LEVEL: standaloneHTTPsection withMETHOD,ENDPOINT,CONTENT_TYPE,ACCEPT, andHEADERSSDKsection (if applicable) withINTERFACE.NAMEandINVOCATIONINPUTSwithLOCATIONfield for each parameterRETURNSas a single object withTYPEandDESCERRORSarray withSTATUScodesSTRUCTSsection with full struct definitions (recursively collected)
ValidationResult
interface ValidationResult {
isValid: boolean;
errors: string[];
warnings: string[];
}Development
Build Commands
# Build TypeScript to dist/
npm run build
# Clean dist folder
npm run clean
# Watch mode for development
npm run dev
# Run example usage
npm run exampleProject Structure
src/
├── index.ts # Main exports (defaults to v1)
├── example-usage.ts # Usage examples
├── versions.ts # Version constants
├── v1/ # Wrekenfile spec v1.2
│ ├── index.ts
│ ├── openapi-to-wreken.ts
│ ├── openapi-v2-to-wrekenfile.ts
│ ├── postman-to-wrekenfile.ts
│ ├── mini-wrekenfile-generator.ts
│ ├── wrekenfile-validator.ts
│ └── cli/ # CLI tools for v1
└── v2/ # Wreken spec v2.0.2
├── index.ts
├── openapi-to-wreken.ts
├── openapi-v2-to-wrekenfile.ts
├── postman-to-wrekenfile.ts
├── mini-wrekenfile-generator.ts
├── unified_mini_wrekenfile_spec_v_2.md # Mini-wrekenfile specification
├── utils/ # Utility functions
│ ├── yaml-utils.ts # YAML processing utilities
│ └── error-utils.ts # Error handling utilities
└── cli/ # CLI tools for v2
dist/ # Compiled JavaScript + types
├── index.js
├── v1/
└── v2/License
MIT
