chore: Standardize files

CONTRIBUTING.md

@@ -14,7 +14,7 @@ npm run compile
 npm run ld-workbench -- --config static/example
 ```
 
-The configuration of this project is validated and defined by [JSON Schema](https://json-schema.org). The schema is located in `./static/ld-workbench-schema.json`. To create the types from this schema, run `npm run util:json-schema-to-typescript`. This will regenerate `./src/types/LDWorkbenchConfiguration.d.ts`, do not modify this file by hand.
+The configuration of this project is validated and defined by [JSON Schema](https://json-schema.org). The schema is located in `./static/ld-workbench-schema.json`. To create the types from this schema, run `npm run util:json-schema-to-typescript`. This will regenerate `./src/types/configuration.d.ts`, do not modify this file by hand.
 
 ## Committing changes
 

src/cliArgs.ts → src/cli.ts

@@ -27,16 +27,16 @@ program
   )
   .version(version());
 program.parse();
-export const cliArgs: {
+export const cli: {
   config: string;
   pipeline?: string;
   stage?: string;
   silent?: boolean;
   init?: boolean;
 } = program.opts();
 
-if (cliArgs.init !== undefined) {
-  if (Object.values(cliArgs).length !== 2) {
+if (cli.init !== undefined) {
+  if (Object.values(cli).length !== 2) {
     error(
       'The --init flag can not be used in conjunction with other CLI arguments.'
     );

src/configuration.d.ts

@@ -0,0 +1,158 @@
+/* eslint-disable */
+/**
+ * This file was automatically generated by json-schema-to-typescript.
+ * DO NOT MODIFY IT BY HAND. Instead, modify the source JSONSchema file,
+ * and run json-schema-to-typescript to regenerate this file.
+ */
+
+/**
+ * JSON Schema for LDWorkbench.
+ * It helps with the writing of the configuration needed to run LDWorkbench pipelines.
+ */
+export interface Configuration {
+  /**
+   * The name of your pipeline. It must be unique over all your configurations.
+   */
+  name: string;
+  /**
+   * An optional description for your pipeline.
+   */
+  description?: string;
+  /**
+   * An optional base directory for files referenced by `file://...` paths. This defaults to the directory that contains the YAML configuration file.
+   */
+  baseDir?: string;
+  /**
+   * The file where the final result of your pipeline is saved.
+   */
+  destination?: string;
+  /**
+   * A pipeline stage consists of an iterator and one or more generators.
+   *
+   * @minItems 1
+   */
+  stages: [
+    {
+      /**
+       * The name of the stage. It must be unique within the pipeline.
+       */
+      name: string;
+      iterator: {
+        /**
+         * SPARQL SELECT query that returns a `$this` binding for each URI that will be passed to the generator(s). Either an inline string (`SELECT $this WHERE {...}`) or a reference to a file (`file://...`) that contains the query.
+         */
+        query: string;
+        /**
+         * SPARQL endpoint for the iterator. If it starts with `file://`, a local RDF file is queried. If omitted the result of the previous stage is used.
+         */
+        endpoint?: string;
+        /**
+         * Number of `$this` bindings retrieved per query. Defaults to the LIMIT value of your iterator query or 10 if no LIMIT is present.
+         */
+        batchSize?: number;
+        /**
+         * Human-readable time delay for requests to the the iterator’s SPARQL endpoint (e.g. `5ms`, `100 milliseconds`, `1s`).
+         */
+        delay?: string;
+      };
+      /**
+       * @minItems 1
+       */
+      generator: [
+        {
+          /**
+           * SPARQL CONSTRUCT query that takes a `$this` binding from the iterator and generates triples for it. Either an inline string (`CONSTRUCT $this schema:name ?name WHERE {$this ...}`) or a reference to a file (`file://...`) that contains the query.
+           */
+          query: string;
+          /**
+           * The SPARQL endpoint for the generator. If it starts with `file://`, a local RDF file is queried. If omitted, the endpoint of the iterator is used.
+           */
+          endpoint?: string;
+          /**
+           * Overrule the generator's behaviour of fetching results for 10 bindings of `$this` per request.
+           */
+          batchSize?: number;
+        },
+        ...{
+          /**
+           * SPARQL CONSTRUCT query that takes a `$this` binding from the iterator and generates triples for it. Either an inline string (`CONSTRUCT $this schema:name ?name WHERE {$this ...}`) or a reference to a file (`file://...`) that contains the query.
+           */
+          query: string;
+          /**
+           * The SPARQL endpoint for the generator. If it starts with `file://`, a local RDF file is queried. If omitted, the endpoint of the iterator is used.
+           */
+          endpoint?: string;
+          /**
+           * Overrule the generator's behaviour of fetching results for 10 bindings of `$this` per request.
+           */
+          batchSize?: number;
+        }[]
+      ];
+      /**
+       * The optional path where the results are saved. If omitted, a temporary file will be created.
+       */
+      destination?: string;
+    },
+    ...{
+      /**
+       * The name of the stage. It must be unique within the pipeline.
+       */
+      name: string;
+      iterator: {
+        /**
+         * SPARQL SELECT query that returns a `$this` binding for each URI that will be passed to the generator(s). Either an inline string (`SELECT $this WHERE {...}`) or a reference to a file (`file://...`) that contains the query.
+         */
+        query: string;
+        /**
+         * SPARQL endpoint for the iterator. If it starts with `file://`, a local RDF file is queried. If omitted the result of the previous stage is used.
+         */
+        endpoint?: string;
+        /**
+         * Number of `$this` bindings retrieved per query. Defaults to the LIMIT value of your iterator query or 10 if no LIMIT is present.
+         */
+        batchSize?: number;
+        /**
+         * Human-readable time delay for requests to the the iterator’s SPARQL endpoint (e.g. `5ms`, `100 milliseconds`, `1s`).
+         */
+        delay?: string;
+      };
+      /**
+       * @minItems 1
+       */
+      generator: [
+        {
+          /**
+           * SPARQL CONSTRUCT query that takes a `$this` binding from the iterator and generates triples for it. Either an inline string (`CONSTRUCT $this schema:name ?name WHERE {$this ...}`) or a reference to a file (`file://...`) that contains the query.
+           */
+          query: string;
+          /**
+           * The SPARQL endpoint for the generator. If it starts with `file://`, a local RDF file is queried. If omitted, the endpoint of the iterator is used.
+           */
+          endpoint?: string;
+          /**
+           * Overrule the generator's behaviour of fetching results for 10 bindings of `$this` per request.
+           */
+          batchSize?: number;
+        },
+        ...{
+          /**
+           * SPARQL CONSTRUCT query that takes a `$this` binding from the iterator and generates triples for it. Either an inline string (`CONSTRUCT $this schema:name ?name WHERE {$this ...}`) or a reference to a file (`file://...`) that contains the query.
+           */
+          query: string;
+          /**
+           * The SPARQL endpoint for the generator. If it starts with `file://`, a local RDF file is queried. If omitted, the endpoint of the iterator is used.
+           */
+          endpoint?: string;
+          /**
+           * Overrule the generator's behaviour of fetching results for 10 bindings of `$this` per request.
+           */
+          batchSize?: number;
+        }[]
+      ];
+      /**
+       * The optional path where the results are saved. If omitted, a temporary file will be created.
+       */
+      destination?: string;
+    }[]
+  ];
+}

src/lib/File.class.ts → src/file.ts

@@ -6,11 +6,11 @@ import {
   mkdirSync,
   createReadStream,
 } from 'fs';
-import {isFile, isFilePathString} from '../utils/guards.js';
+import {isFile, isFilePathString} from './utils/guards.js';
 import {dirname} from 'path';
 import chalk from 'chalk';
 import {type Ora} from 'ora';
-import type Pipeline from './Pipeline.class.js';
+import type Pipeline from './pipeline.js';
 import {pipeline as streamPipeline} from 'stream/promises';
 
 export default class File {

src/lib/Generator.class.ts → src/generator.ts

@@ -1,12 +1,12 @@
 import type {ConstructQuery} from 'sparqljs';
-import type Stage from './Stage.class.js';
-import getSPARQLQuery from '../utils/getSPARQLQuery.js';
+import type Stage from './stage.js';
+import getSPARQLQuery from './utils/getSPARQLQuery.js';
 import type {Quad, NamedNode} from '@rdfjs/types';
-import getSPARQLQueryString from '../utils/getSPARQLQueryString.js';
-import getEndpoint from '../utils/getEndpoint.js';
+import getSPARQLQueryString from './utils/getSPARQLQueryString.js';
+import getEndpoint from './utils/getEndpoint.js';
 import type {Endpoint, QueryEngine, QuerySource} from './types.js';
-import getEngine from '../utils/getEngine.js';
-import getEngineSource from '../utils/getEngineSource.js';
+import getEngine from './utils/getEngine.js';
+import getEngineSource from './utils/getEngineSource.js';
 import EventEmitter from 'node:events';
 
 const DEFAULT_BATCH_SIZE = 10;

src/lib/Iterator.class.ts → src/iterator.ts

@@ -1,14 +1,14 @@
 import EventEmitter from 'node:events';
 import type {SelectQuery} from 'sparqljs';
-import type Stage from './Stage.class.js';
+import type Stage from './stage.js';
 import type {NamedNode} from '@rdfjs/types';
-import getSPARQLQuery from '../utils/getSPARQLQuery.js';
+import getSPARQLQuery from './utils/getSPARQLQuery.js';
 import {type Bindings} from '@comunica/types';
-import getSPARQLQueryString from '../utils/getSPARQLQueryString.js';
-import getEndpoint from '../utils/getEndpoint.js';
+import getSPARQLQueryString from './utils/getSPARQLQueryString.js';
+import getEndpoint from './utils/getEndpoint.js';
 import type {Endpoint, QueryEngine, QuerySource} from './types.js';
-import getEngine from '../utils/getEngine.js';
-import getEngineSource from '../utils/getEngineSource.js';
+import getEngine from './utils/getEngine.js';
+import getEngineSource from './utils/getEngineSource.js';
 import parse from 'parse-duration';
 
 const DEFAULT_LIMIT = 10;