@lde/pipeline 0.6.14 → 0.6.16

package/dist/distribution/index.d.ts

@@ -1,3 +1,4 @@
 export { probe, NetworkError, SparqlProbeResult, DataDumpProbeResult, type ProbeResultType, } from './probe.js';
-export { DistributionAnalyzer, ImportSuccessful, ImportFailed, NoDistributionAvailable, type Importer, type DistributionAnalyzerOptions, } from './analyzer.js';
+export { probeResultsToQuads } from './report.js';
+export { ResolvedDistribution, NoDistributionAvailable, SparqlDistributionResolver, type DistributionResolver, type SparqlDistributionResolverOptions, } from './resolver.js';
 //# sourceMappingURL=index.d.ts.map

package/dist/distribution/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/distribution/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,KAAK,EACL,YAAY,EACZ,iBAAiB,EACjB,mBAAmB,EACnB,KAAK,eAAe,GACrB,MAAM,YAAY,CAAC;AAEpB,OAAO,EACL,oBAAoB,EACpB,gBAAgB,EAChB,YAAY,EACZ,uBAAuB,EACvB,KAAK,QAAQ,EACb,KAAK,2BAA2B,GACjC,MAAM,eAAe,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/distribution/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,KAAK,EACL,YAAY,EACZ,iBAAiB,EACjB,mBAAmB,EACnB,KAAK,eAAe,GACrB,MAAM,YAAY,CAAC;AAEpB,OAAO,EAAE,mBAAmB,EAAE,MAAM,aAAa,CAAC;AAElD,OAAO,EACL,oBAAoB,EACpB,uBAAuB,EACvB,0BAA0B,EAC1B,KAAK,oBAAoB,EACzB,KAAK,iCAAiC,GACvC,MAAM,eAAe,CAAC"}

package/dist/distribution/index.js

@@ -1,2 +1,3 @@
 export { probe, NetworkError, SparqlProbeResult, DataDumpProbeResult, } from './probe.js';
-export { DistributionAnalyzer, ImportSuccessful, ImportFailed, NoDistributionAvailable, } from './analyzer.js';
+export { probeResultsToQuads } from './report.js';
+export { ResolvedDistribution, NoDistributionAvailable, SparqlDistributionResolver, } from './resolver.js';

package/dist/distribution/probe.d.ts

@@ -40,7 +40,7 @@ export type ProbeResultType = SparqlProbeResult | DataDumpProbeResult | NetworkE
  * For SPARQL endpoints, sends a simple SELECT query.
  * For data dumps, sends HEAD (or GET if HEAD returns no Content-Length).
  *
- * Updates the distribution's isValid, lastModified, and byteSize properties.
+ * Returns pure probe results without mutating the distribution.
  */
 export declare function probe(distribution: Distribution, timeout?: number): Promise<ProbeResultType>;
 export {};

package/dist/distribution/probe.js

@@ -61,7 +61,7 @@ export class DataDumpProbeResult extends ProbeResult {
  * For SPARQL endpoints, sends a simple SELECT query.
  * For data dumps, sends HEAD (or GET if HEAD returns no Content-Length).
  *
- * Updates the distribution's isValid, lastModified, and byteSize properties.
+ * Returns pure probe results without mutating the distribution.
  */
 export async function probe(distribution, timeout = 5000) {
     try {
@@ -85,9 +85,7 @@ async function probeSparqlEndpoint(distribution, timeout) {
         },
         body: `query=${encodeURIComponent('SELECT * { ?s ?p ?o } LIMIT 1')}`,
     });
-    const result = new SparqlProbeResult(url, response);
-    distribution.isValid = result.isSuccess();
-    return result;
+    return new SparqlProbeResult(url, response);
 }
 async function probeDataDump(distribution, timeout) {
     const url = distribution.accessUrl.toString();
@@ -112,9 +110,5 @@ async function probeDataDump(distribution, timeout) {
             ...requestOptions,
         });
     }
-    const result = new DataDumpProbeResult(url, response);
-    distribution.isValid = result.isSuccess();
-    distribution.lastModified ??= result.lastModified ?? undefined;
-    distribution.byteSize ??= result.contentSize ?? undefined;
-    return result;
+    return new DataDumpProbeResult(url, response);
 }

package/dist/distribution/report.d.ts

@@ -0,0 +1,15 @@
+import type { ImportFailed } from '@lde/sparql-importer';
+import { type Quad } from 'n3';
+import { type ProbeResultType } from './probe.js';
+/**
+ * Convert probe results into RDF quads describing each probe as a `schema:Action`.
+ *
+ * Successful SPARQL probes emit `void:sparqlEndpoint`;
+ * successful data-dump probes emit `void:dataDump` with optional metadata.
+ * Failed probes emit `schema:error`.
+ *
+ * When an {@link ImportFailed} is provided its error is attached to the action
+ * whose `schema:target` matches the failed distribution's access URL.
+ */
+export declare function probeResultsToQuads(probeResults: ProbeResultType[], datasetIri: string, importResult?: ImportFailed): AsyncIterable<Quad>;
+//# sourceMappingURL=report.d.ts.map

package/dist/distribution/report.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"report.d.ts","sourceRoot":"","sources":["../../src/distribution/report.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AACzD,OAAO,EAAe,KAAK,IAAI,EAAE,MAAM,IAAI,CAAC;AAC5C,OAAO,EAIL,KAAK,eAAe,EACrB,MAAM,YAAY,CAAC;AAUpB;;;;;;;;;GASG;AACH,wBAAuB,mBAAmB,CACxC,YAAY,EAAE,eAAe,EAAE,EAC/B,UAAU,EAAE,MAAM,EAClB,YAAY,CAAC,EAAE,YAAY,GAC1B,aAAa,CAAC,IAAI,CAAC,CAkCrB"}

package/dist/distribution/report.js

@@ -0,0 +1,64 @@
+import { DataFactory } from 'n3';
+import { NetworkError, SparqlProbeResult, } from './probe.js';
+const { quad, namedNode, blankNode, literal } = DataFactory;
+const RDF = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#';
+const SCHEMA = 'https://schema.org/';
+const VOID = 'http://rdfs.org/ns/void#';
+const XSD = 'http://www.w3.org/2001/XMLSchema#';
+const HTTP_STATUS = 'https://www.w3.org/2011/http-statusCodes#';
+/**
+ * Convert probe results into RDF quads describing each probe as a `schema:Action`.
+ *
+ * Successful SPARQL probes emit `void:sparqlEndpoint`;
+ * successful data-dump probes emit `void:dataDump` with optional metadata.
+ * Failed probes emit `schema:error`.
+ *
+ * When an {@link ImportFailed} is provided its error is attached to the action
+ * whose `schema:target` matches the failed distribution's access URL.
+ */
+export async function* probeResultsToQuads(probeResults, datasetIri, importResult) {
+    // Track blank nodes per URL so import errors can reference the right action.
+    const actionsByUrl = new Map();
+    for (const result of probeResults) {
+        const action = blankNode();
+        actionsByUrl.set(result.url, action);
+        yield quad(action, namedNode(`${RDF}type`), namedNode(`${SCHEMA}Action`));
+        yield quad(action, namedNode(`${SCHEMA}target`), namedNode(result.url));
+        if (result instanceof NetworkError) {
+            yield quad(action, namedNode(`${SCHEMA}error`), literal(result.message));
+        }
+        else if (result.isSuccess()) {
+            yield* successQuads(action, result, datasetIri);
+        }
+        else {
+            // HTTP error
+            const statusUri = `${HTTP_STATUS}${result.statusText.replace(/ /g, '')}`;
+            yield quad(action, namedNode(`${SCHEMA}error`), namedNode(statusUri));
+        }
+    }
+    if (importResult) {
+        const action = actionsByUrl.get(importResult.distribution.accessUrl.toString());
+        if (action) {
+            yield quad(action, namedNode(`${SCHEMA}error`), literal(importResult.error));
+        }
+    }
+}
+function* successQuads(action, result, datasetIri) {
+    const distributionUrl = namedNode(result.url);
+    yield quad(action, namedNode(`${SCHEMA}result`), distributionUrl);
+    if (result.lastModified) {
+        yield quad(distributionUrl, namedNode(`${SCHEMA}dateModified`), literal(result.lastModified.toISOString(), namedNode(`${XSD}dateTime`)));
+    }
+    if (result instanceof SparqlProbeResult) {
+        yield quad(namedNode(datasetIri), namedNode(`${VOID}sparqlEndpoint`), distributionUrl);
+    }
+    else {
+        yield quad(namedNode(datasetIri), namedNode(`${VOID}dataDump`), distributionUrl);
+        if (result.contentSize) {
+            yield quad(distributionUrl, namedNode(`${SCHEMA}contentSize`), literal(result.contentSize.toString()));
+        }
+        if (result.contentType) {
+            yield quad(distributionUrl, namedNode(`${SCHEMA}encodingFormat`), literal(result.contentType));
+        }
+    }
+}

package/dist/distribution/resolver.d.ts

@@ -0,0 +1,35 @@
+import { Dataset, Distribution } from '@lde/dataset';
+import type { Importer } from '@lde/sparql-importer';
+export declare class ResolvedDistribution {
+    readonly distribution: Distribution;
+    constructor(distribution: Distribution);
+}
+export declare class NoDistributionAvailable {
+    readonly dataset: Dataset;
+    readonly message: string;
+    constructor(dataset: Dataset, message: string);
+}
+export interface DistributionResolver {
+    resolve(dataset: Dataset): Promise<ResolvedDistribution | NoDistributionAvailable>;
+}
+export interface SparqlDistributionResolverOptions {
+    importer?: Importer;
+    timeout?: number;
+}
+/**
+ * Resolves a dataset to a usable SPARQL distribution by probing its distributions.
+ *
+ * 1. Probes all distributions in parallel.
+ * 2. Returns the first valid SPARQL endpoint as a `ResolvedDistribution`.
+ * 3. If none: tries the importer (if provided) and returns the imported distribution.
+ * 4. If nothing works: returns `NoDistributionAvailable`.
+ *
+ * Does not mutate `dataset.distributions`.
+ */
+export declare class SparqlDistributionResolver implements DistributionResolver {
+    private readonly importer?;
+    private readonly timeout;
+    constructor(options?: SparqlDistributionResolverOptions);
+    resolve(dataset: Dataset): Promise<ResolvedDistribution | NoDistributionAvailable>;
+}
+//# sourceMappingURL=resolver.d.ts.map

package/dist/distribution/resolver.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"resolver.d.ts","sourceRoot":"","sources":["../../src/distribution/resolver.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AAIrD,qBAAa,oBAAoB;IACnB,QAAQ,CAAC,YAAY,EAAE,YAAY;gBAA1B,YAAY,EAAE,YAAY;CAChD;AAED,qBAAa,uBAAuB;IACtB,QAAQ,CAAC,OAAO,EAAE,OAAO;IAAE,QAAQ,CAAC,OAAO,EAAE,MAAM;gBAA1C,OAAO,EAAE,OAAO,EAAW,OAAO,EAAE,MAAM;CAChE;AAED,MAAM,WAAW,oBAAoB;IACnC,OAAO,CACL,OAAO,EAAE,OAAO,GACf,OAAO,CAAC,oBAAoB,GAAG,uBAAuB,CAAC,CAAC;CAC5D;AAED,MAAM,WAAW,iCAAiC;IAChD,QAAQ,CAAC,EAAE,QAAQ,CAAC;IACpB,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;;;;;;;;GASG;AACH,qBAAa,0BAA2B,YAAW,oBAAoB;IACrE,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAW;IACrC,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAS;gBAErB,OAAO,CAAC,EAAE,iCAAiC;IAKjD,OAAO,CACX,OAAO,EAAE,OAAO,GACf,OAAO,CAAC,oBAAoB,GAAG,uBAAuB,CAAC;CAsC3D"}

package/dist/distribution/resolver.js

@@ -0,0 +1,57 @@
+import { Distribution } from '@lde/dataset';
+import { ImportSuccessful } from '@lde/sparql-importer';
+import { probe, SparqlProbeResult } from './probe.js';
+export class ResolvedDistribution {
+    distribution;
+    constructor(distribution) {
+        this.distribution = distribution;
+    }
+}
+export class NoDistributionAvailable {
+    dataset;
+    message;
+    constructor(dataset, message) {
+        this.dataset = dataset;
+        this.message = message;
+    }
+}
+/**
+ * Resolves a dataset to a usable SPARQL distribution by probing its distributions.
+ *
+ * 1. Probes all distributions in parallel.
+ * 2. Returns the first valid SPARQL endpoint as a `ResolvedDistribution`.
+ * 3. If none: tries the importer (if provided) and returns the imported distribution.
+ * 4. If nothing works: returns `NoDistributionAvailable`.
+ *
+ * Does not mutate `dataset.distributions`.
+ */
+export class SparqlDistributionResolver {
+    importer;
+    timeout;
+    constructor(options) {
+        this.importer = options?.importer;
+        this.timeout = options?.timeout ?? 5000;
+    }
+    async resolve(dataset) {
+        const results = await Promise.all(dataset.distributions.map((distribution) => probe(distribution, this.timeout)));
+        // Find first valid SPARQL endpoint.
+        for (let i = 0; i < dataset.distributions.length; i++) {
+            const distribution = dataset.distributions[i];
+            const result = results[i];
+            if (distribution.isSparql() &&
+                result instanceof SparqlProbeResult &&
+                result.isSuccess()) {
+                return new ResolvedDistribution(distribution);
+            }
+        }
+        // No SPARQL endpoint; try importer if available.
+        if (this.importer) {
+            const importResult = await this.importer.import(dataset);
+            if (importResult instanceof ImportSuccessful) {
+                const distribution = Distribution.sparql(importResult.distribution.accessUrl, importResult.identifier);
+                return new ResolvedDistribution(distribution);
+            }
+        }
+        return new NoDistributionAvailable(dataset, 'No SPARQL endpoint or importable data dump available');
+    }
+}

package/dist/index.js

@@ -7,3 +7,4 @@ export * from './builder.js';
 export * from './config.js';
 export * from './sparql/index.js';
 export * from './distribution/index.js';
+// first test

package/dist/pipeline.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"pipeline.d.ts","sourceRoot":"","sources":["../src/pipeline.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAc,IAAI,EAAE,MAAM,WAAW,CAAC;AAI7C,qBAAa,QAAQ;IACP,OAAO,CAAC,QAAQ,CAAC,MAAM;gBAAN,MAAM,EAAE;QAAE,QAAQ,EAAE,QAAQ,CAAC;QAAC,KAAK,EAAE,IAAI,EAAE,CAAA;KAAE;IAE7D,GAAG;CAgCjB"}
+{"version":3,"file":"pipeline.d.ts","sourceRoot":"","sources":["../src/pipeline.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,EAAE,MAAM,eAAe,CAAC;AACzC,OAAO,EAAc,IAAI,EAAE,MAAM,WAAW,CAAC;AAI7C,qBAAa,QAAQ;IACP,OAAO,CAAC,QAAQ,CAAC,MAAM;gBAAN,MAAM,EAAE;QAAE,QAAQ,EAAE,QAAQ,CAAC;QAAC,KAAK,EAAE,IAAI,EAAE,CAAA;KAAE;IAE7D,GAAG;CAiCjB"}

package/dist/pipeline.js

@@ -8,8 +8,9 @@ export class Pipeline {
     async run() {
         const datasets = await this.config.selector.select();
         for await (const dataset of datasets) {
+            const distribution = dataset.getSparqlDistribution() ?? undefined;
             for (const step of this.config.steps) {
-                const result = await step.execute(dataset);
+                const result = await step.execute(dataset, distribution);
                 if (result instanceof NotSupported) {
                     console.error(result);
                 }

package/dist/sparql/executor.d.ts

@@ -6,24 +6,21 @@ import { NotSupported } from '../step.js';
 export { NotSupported } from '../step.js';
 /** A single row of variable bindings (variable name → NamedNode). */
 export type VariableBindings = Record<string, NamedNode>;
+export interface ExecuteOptions {
+    /**
+     * Variable bindings to inject as a VALUES clause into the query.
+     * When non-empty, a VALUES block is prepended to the WHERE clause.
+     */
+    bindings?: VariableBindings[];
+}
 export interface Executor {
-    execute(dataset: ExecutableDataset, options?: SparqlConstructExecuteOptions): Promise<AsyncIterable<Quad> | NotSupported>;
+    execute(dataset: Dataset, distribution: Distribution, options?: ExecuteOptions): Promise<AsyncIterable<Quad> | NotSupported>;
 }
 /**
  * A quad stream that is both an RDFJS Stream and Node.js Readable (async iterable).
  * This is the actual return type from SparqlEndpointFetcher.fetchTriples().
  */
 export type QuadStream = Readable & Stream<Quad>;
-/**
- * Extended dataset with optional SPARQL filtering options.
- */
-export interface ExecutableDataset extends Dataset {
-    /**
-     * Optional SPARQL filter clause to restrict analysis to a subset of the data.
-     * This is substituted for `#subjectFilter#` in queries.
-     */
-    subjectFilter?: string;
-}
 /**
  * Options for SparqlConstructExecutor.
  */
@@ -42,20 +39,6 @@ export interface SparqlConstructExecutorOptions {
      */
     fetcher?: SparqlEndpointFetcher;
 }
-/**
- * Options for `execute()`.
- */
-export interface SparqlConstructExecuteOptions {
-    /**
-     * Explicit SPARQL endpoint URL. If not provided, uses the dataset's SPARQL distribution.
-     */
-    endpoint?: URL;
-    /**
-     * Variable bindings to inject as a VALUES clause into the query.
-     * When non-empty, a VALUES block is prepended to the WHERE clause.
-     */
-    bindings?: VariableBindings[];
-}
 /**
  * A streaming SPARQL CONSTRUCT executor that parses the query once (in the
  * constructor) and operates on the AST for graph and VALUES injection.
@@ -69,7 +52,7 @@ export interface SparqlConstructExecuteOptions {
  * const executor = new SparqlConstructExecutor({
  *   query: 'CONSTRUCT { ?dataset ?p ?o } WHERE { ?s ?p ?o }',
  * });
- * const result = await executor.execute(dataset);
+ * const result = await executor.execute(dataset, distribution);
  * if (result instanceof NotSupported) {
  *   console.log(result.message);
  * } else {
@@ -85,13 +68,14 @@ export declare class SparqlConstructExecutor implements Executor {
     private readonly generator;
     constructor(options: SparqlConstructExecutorOptions);
     /**
-     * Execute the SPARQL CONSTRUCT query against the dataset's SPARQL endpoint.
+     * Execute the SPARQL CONSTRUCT query against the distribution's endpoint.
      *
      * @param dataset The dataset to execute against.
-     * @param options Optional endpoint override.
-     * @returns AsyncIterable<Quad> stream of results, or NotSupported if no SPARQL endpoint available.
+     * @param distribution The distribution providing the SPARQL endpoint.
+     * @param options Optional execution options (bindings).
+     * @returns AsyncIterable<Quad> stream of results.
      */
-    execute(dataset: ExecutableDataset, options?: SparqlConstructExecuteOptions): Promise<QuadStream | NotSupported>;
+    execute(dataset: Dataset, distribution: Distribution, options?: ExecuteOptions): Promise<QuadStream>;
     /**
      * Create an executor from a query file.
      *
@@ -103,11 +87,11 @@ export declare class SparqlConstructExecutor implements Executor {
 /**
  * Substitute template variables in a SPARQL query.
  *
- * - `#subjectFilter#` — replaced with the distribution's or dataset's subject filter
+ * - `#subjectFilter#` — replaced with the distribution's subject filter
  * - `#namedGraph#` — replaced with `FROM <graph>` clause if the distribution has a named graph
  * - `?dataset` — replaced with the dataset IRI
  */
-export declare function substituteQueryTemplates(query: string, distribution: Distribution | null, dataset: ExecutableDataset): string;
+export declare function substituteQueryTemplates(query: string, distribution: Distribution | null, dataset: Dataset): string;
 /**
  * Read a SPARQL query from a file.
  */

package/dist/sparql/executor.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"executor.d.ts","sourceRoot":"","sources":["../../src/sparql/executor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAC9D,OAAO,KAAK,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAC5D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAI5C,OAAO,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAK1C,OAAO,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAE1C,qEAAqE;AACrE,MAAM,MAAM,gBAAgB,GAAG,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;AAEzD,MAAM,WAAW,QAAQ;IACvB,OAAO,CACL,OAAO,EAAE,iBAAiB,EAC1B,OAAO,CAAC,EAAE,6BAA6B,GACtC,OAAO,CAAC,aAAa,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;CAChD;AAED;;;GAGG;AACH,MAAM,MAAM,UAAU,GAAG,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;AAEjD;;GAEG;AACH,MAAM,WAAW,iBAAkB,SAAQ,OAAO;IAChD;;;OAGG;IACH,aAAa,CAAC,EAAE,MAAM,CAAC;CACxB;AAED;;GAEG;AACH,MAAM,WAAW,8BAA8B;IAC7C;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IAEd;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;IAEjB;;OAEG;IACH,OAAO,CAAC,EAAE,qBAAqB,CAAC;CACjC;AAED;;GAEG;AACH,MAAM,WAAW,6BAA6B;IAC5C;;OAEG;IACH,QAAQ,CAAC,EAAE,GAAG,CAAC;IAEf;;;OAGG;IACH,QAAQ,CAAC,EAAE,gBAAgB,EAAE,CAAC;CAC/B;AAED;;;;;;;;;;;;;;;;;;;;;;GAsBG;AACH,qBAAa,uBAAwB,YAAW,QAAQ;IACtD,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAiB;IACvC,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAwB;IAChD,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;gBAEjC,OAAO,EAAE,8BAA8B;IAcnD;;;;;;OAMG;IACG,OAAO,CACX,OAAO,EAAE,iBAAiB,EAC1B,OAAO,CAAC,EAAE,6BAA6B,GACtC,OAAO,CAAC,UAAU,GAAG,YAAY,CAAC;IA4BrC;;;;;OAKG;WACiB,QAAQ,CAC1B,QAAQ,EAAE,MAAM,EAChB,OAAO,CAAC,EAAE,IAAI,CAAC,8BAA8B,EAAE,OAAO,CAAC,GACtD,OAAO,CAAC,uBAAuB,CAAC;CAIpC;AAED;;;;;;GAMG;AACH,wBAAgB,wBAAwB,CACtC,KAAK,EAAE,MAAM,EACb,YAAY,EAAE,YAAY,GAAG,IAAI,EACjC,OAAO,EAAE,iBAAiB,GACzB,MAAM,CAYR;AAED;;GAEG;AACH,wBAAsB,aAAa,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAErE"}
+{"version":3,"file":"executor.d.ts","sourceRoot":"","sources":["../../src/sparql/executor.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAC9D,OAAO,KAAK,EAAE,SAAS,EAAE,IAAI,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAC5D,OAAO,KAAK,EAAE,QAAQ,EAAE,MAAM,aAAa,CAAC;AAI5C,OAAO,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAK1C,OAAO,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAE1C,qEAAqE;AACrE,MAAM,MAAM,gBAAgB,GAAG,MAAM,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;AAEzD,MAAM,WAAW,cAAc;IAC7B;;;OAGG;IACH,QAAQ,CAAC,EAAE,gBAAgB,EAAE,CAAC;CAC/B;AAED,MAAM,WAAW,QAAQ;IACvB,OAAO,CACL,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY,EAC1B,OAAO,CAAC,EAAE,cAAc,GACvB,OAAO,CAAC,aAAa,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC,CAAC;CAChD;AAED;;;GAGG;AACH,MAAM,MAAM,UAAU,GAAG,QAAQ,GAAG,MAAM,CAAC,IAAI,CAAC,CAAC;AAEjD;;GAEG;AACH,MAAM,WAAW,8BAA8B;IAC7C;;OAEG;IACH,KAAK,EAAE,MAAM,CAAC;IAEd;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;IAEjB;;OAEG;IACH,OAAO,CAAC,EAAE,qBAAqB,CAAC;CACjC;AAED;;;;;;;;;;;;;;;;;;;;;;GAsBG;AACH,qBAAa,uBAAwB,YAAW,QAAQ;IACtD,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAiB;IACvC,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAwB;IAChD,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAmB;gBAEjC,OAAO,EAAE,8BAA8B;IAcnD;;;;;;;OAOG;IACG,OAAO,CACX,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY,EAC1B,OAAO,CAAC,EAAE,cAAc,GACvB,OAAO,CAAC,UAAU,CAAC;IAoBtB;;;;;OAKG;WACiB,QAAQ,CAC1B,QAAQ,EAAE,MAAM,EAChB,OAAO,CAAC,EAAE,IAAI,CAAC,8BAA8B,EAAE,OAAO,CAAC,GACtD,OAAO,CAAC,uBAAuB,CAAC;CAIpC;AAED;;;;;;GAMG;AACH,wBAAgB,wBAAwB,CACtC,KAAK,EAAE,MAAM,EACb,YAAY,EAAE,YAAY,GAAG,IAAI,EACjC,OAAO,EAAE,OAAO,GACf,MAAM,CAWR;AAED;;GAEG;AACH,wBAAsB,aAAa,CAAC,QAAQ,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAErE"}

package/dist/sparql/executor.js

@@ -2,7 +2,6 @@ import { SparqlEndpointFetcher } from 'fetch-sparql-endpoint';
 import { readFile } from 'node:fs/promises';
 import { resolve } from 'node:path';
 import { Generator, Parser } from 'sparqljs';
-import { NotSupported } from '../step.js';
 import { withDefaultGraph } from './graph.js';
 import { injectValues } from './values.js';
 // Re-export for convenience
@@ -20,7 +19,7 @@ export { NotSupported } from '../step.js';
  * const executor = new SparqlConstructExecutor({
  *   query: 'CONSTRUCT { ?dataset ?p ?o } WHERE { ?s ?p ?o }',
  * });
- * const result = await executor.execute(dataset);
+ * const result = await executor.execute(dataset, distribution);
  * if (result instanceof NotSupported) {
  *   console.log(result.message);
  * } else {
@@ -48,23 +47,17 @@ export class SparqlConstructExecutor {
                 });
     }
     /**
-     * Execute the SPARQL CONSTRUCT query against the dataset's SPARQL endpoint.
+     * Execute the SPARQL CONSTRUCT query against the distribution's endpoint.
      *
      * @param dataset The dataset to execute against.
-     * @param options Optional endpoint override.
-     * @returns AsyncIterable<Quad> stream of results, or NotSupported if no SPARQL endpoint available.
+     * @param distribution The distribution providing the SPARQL endpoint.
+     * @param options Optional execution options (bindings).
+     * @returns AsyncIterable<Quad> stream of results.
      */
-    async execute(dataset, options) {
-        const distribution = dataset.getSparqlDistribution();
-        let endpoint = options?.endpoint;
-        if (endpoint === undefined) {
-            if (distribution === null || !distribution.isValid) {
-                return new NotSupported('No SPARQL distribution available');
-            }
-            endpoint = distribution.accessUrl;
-        }
+    async execute(dataset, distribution, options) {
+        const endpoint = distribution.accessUrl;
         let ast = structuredClone(this.query);
-        if (distribution?.namedGraph) {
+        if (distribution.namedGraph) {
             withDefaultGraph(ast, distribution.namedGraph);
         }
         const bindings = options?.bindings;
@@ -89,12 +82,12 @@ export class SparqlConstructExecutor {
 /**
  * Substitute template variables in a SPARQL query.
  *
- * - `#subjectFilter#` — replaced with the distribution's or dataset's subject filter
+ * - `#subjectFilter#` — replaced with the distribution's subject filter
  * - `#namedGraph#` — replaced with `FROM <graph>` clause if the distribution has a named graph
  * - `?dataset` — replaced with the dataset IRI
  */
 export function substituteQueryTemplates(query, distribution, dataset) {
-    const subjectFilter = distribution?.subjectFilter ?? dataset.subjectFilter ?? '';
+    const subjectFilter = distribution?.subjectFilter ?? '';
     const namedGraph = distribution?.namedGraph
         ? `FROM <${distribution.namedGraph}>`
         : '';

package/dist/sparql/index.d.ts

@@ -1,4 +1,4 @@
-export { SparqlConstructExecutor, substituteQueryTemplates, NotSupported, readQueryFile, type ExecutableDataset, type Executor, type SparqlConstructExecuteOptions, type SparqlConstructExecutorOptions, type QuadStream, type VariableBindings, } from './executor.js';
+export { SparqlConstructExecutor, substituteQueryTemplates, NotSupported, readQueryFile, type ExecuteOptions, type Executor, type SparqlConstructExecutorOptions, type QuadStream, type VariableBindings, } from './executor.js';
 export { collect } from './collect.js';
 export { SparqlSelector, type SparqlSelectorOptions } from './selector.js';
 export { injectValues } from './values.js';

package/dist/sparql/index.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/sparql/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,uBAAuB,EACvB,wBAAwB,EACxB,YAAY,EACZ,aAAa,EACb,KAAK,iBAAiB,EACtB,KAAK,QAAQ,EACb,KAAK,6BAA6B,EAClC,KAAK,8BAA8B,EACnC,KAAK,UAAU,EACf,KAAK,gBAAgB,GACtB,MAAM,eAAe,CAAC;AAEvB,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AAEvC,OAAO,EAAE,cAAc,EAAE,KAAK,qBAAqB,EAAE,MAAM,eAAe,CAAC;AAE3E,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAE3C,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC"}
+{"version":3,"file":"index.d.ts","sourceRoot":"","sources":["../../src/sparql/index.ts"],"names":[],"mappings":"AAAA,OAAO,EACL,uBAAuB,EACvB,wBAAwB,EACxB,YAAY,EACZ,aAAa,EACb,KAAK,cAAc,EACnB,KAAK,QAAQ,EACb,KAAK,8BAA8B,EACnC,KAAK,UAAU,EACf,KAAK,gBAAgB,GACtB,MAAM,eAAe,CAAC;AAEvB,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AAEvC,OAAO,EAAE,cAAc,EAAE,KAAK,qBAAqB,EAAE,MAAM,eAAe,CAAC;AAE3E,OAAO,EAAE,YAAY,EAAE,MAAM,aAAa,CAAC;AAE3C,OAAO,EAAE,gBAAgB,EAAE,MAAM,YAAY,CAAC"}

package/dist/stage.d.ts

@@ -1,5 +1,6 @@
+import { Dataset, Distribution } from '@lde/dataset';
 import type { Quad } from '@rdfjs/types';
-import type { ExecutableDataset, Executor, VariableBindings } from './sparql/executor.js';
+import type { Executor, VariableBindings } from './sparql/executor.js';
 import { NotSupported } from './sparql/executor.js';
 export interface StageOptions {
     name: string;
@@ -11,7 +12,7 @@ export declare class Stage {
     private readonly executors;
     private readonly selector?;
     constructor(options: StageOptions);
-    run(dataset: ExecutableDataset): Promise<AsyncIterable<Quad> | NotSupported>;
+    run(dataset: Dataset, distribution: Distribution): Promise<AsyncIterable<Quad> | NotSupported>;
     private collectBindings;
 }
 /** Stage-level selector that yields variable bindings for use in executor queries. Pagination is an implementation detail. */

package/dist/stage.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"stage.d.ts","sourceRoot":"","sources":["../src/stage.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,KAAK,EACV,iBAAiB,EACjB,QAAQ,EACR,gBAAgB,EACjB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAEpD,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,QAAQ,GAAG,QAAQ,EAAE,CAAC;IACjC,QAAQ,CAAC,EAAE,aAAa,CAAC;CAC1B;AAED,qBAAa,KAAK;IAChB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAa;IACvC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAgB;gBAE9B,OAAO,EAAE,YAAY;IAQ3B,GAAG,CACP,OAAO,EAAE,iBAAiB,GACzB,OAAO,CAAC,aAAa,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC;YAoBhC,eAAe;CAW9B;AAUD,8HAA8H;AAE9H,MAAM,WAAW,aAAc,SAAQ,aAAa,CAAC,gBAAgB,CAAC;CAAG"}
+{"version":3,"file":"stage.d.ts","sourceRoot":"","sources":["../src/stage.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,KAAK,EAAE,IAAI,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,KAAK,EACV,QAAQ,EAER,gBAAgB,EACjB,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAE,YAAY,EAAE,MAAM,sBAAsB,CAAC;AAEpD,MAAM,WAAW,YAAY;IAC3B,IAAI,EAAE,MAAM,CAAC;IACb,SAAS,EAAE,QAAQ,GAAG,QAAQ,EAAE,CAAC;IACjC,QAAQ,CAAC,EAAE,aAAa,CAAC;CAC1B;AAED,qBAAa,KAAK;IAChB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,OAAO,CAAC,QAAQ,CAAC,SAAS,CAAa;IACvC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAgB;gBAE9B,OAAO,EAAE,YAAY;IAQ3B,GAAG,CACP,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY,GACzB,OAAO,CAAC,aAAa,CAAC,IAAI,CAAC,GAAG,YAAY,CAAC;YAyBhC,eAAe;CAW9B;AAUD,8HAA8H;AAE9H,MAAM,WAAW,aAAc,SAAQ,aAAa,CAAC,gBAAgB,CAAC;CAAG"}

package/dist/stage.js

@@ -10,12 +10,12 @@ export class Stage {
             : [options.executors];
         this.selector = options.selector;
     }
-    async run(dataset) {
+    async run(dataset, distribution) {
         const bindings = await this.collectBindings();
         const executeOptions = bindings.length > 0 ? { bindings } : undefined;
         const streams = [];
         for (const executor of this.executors) {
-            const result = await executor.execute(dataset, executeOptions);
+            const result = await executor.execute(dataset, distribution, executeOptions);
             if (!(result instanceof NotSupported)) {
                 streams.push(result);
             }

package/dist/step/sparqlQuery.d.ts

@@ -1,5 +1,5 @@
-import { DataEmittingStep, NotSupported } from './../step.js';
-import { Dataset } from '@lde/dataset';
+import { DataEmittingStep } from './../step.js';
+import { Dataset, Distribution } from '@lde/dataset';
 import { SparqlEndpointFetcher } from 'fetch-sparql-endpoint';
 /**
  * Arguments for the SparqlQuery step.
@@ -25,7 +25,7 @@ export declare class SparqlQuery implements DataEmittingStep {
     private readonly query;
     private readonly fetcher?;
     constructor({ identifier, query, fetcher }: Args);
-    execute(dataset: Dataset): Promise<NotSupported | import("../sparql/executor.js").QuadStream>;
+    execute(dataset: Dataset, distribution: Distribution): Promise<import("../sparql/executor.js").QuadStream>;
     static fromFile(filename: string): Promise<SparqlQuery>;
 }
 /**

package/dist/step/sparqlQuery.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"sparqlQuery.d.ts","sourceRoot":"","sources":["../../src/step/sparqlQuery.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAC9D,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AACvC,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAO9D;;;;;;GAMG;AACH,MAAM,WAAW,IAAI;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,CAAC,EAAE,qBAAqB,CAAC;CACjC;AAED;;;;;;GAMG;AACH,qBAAa,WAAY,YAAW,gBAAgB;IAClD,SAAgB,UAAU,SAAC;IAC3B,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAS;IAC/B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAwB;gBAErC,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE,IAAI;IAM1C,OAAO,CAAC,OAAO,EAAE,OAAO;WAiBV,QAAQ,CAAC,QAAQ,EAAE,MAAM;CAM9C;AAED;;GAEG;AACH,wBAAsB,QAAQ,CAAC,QAAQ,EAAE,MAAM,mBAE9C"}
+{"version":3,"file":"sparqlQuery.d.ts","sourceRoot":"","sources":["../../src/step/sparqlQuery.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,gBAAgB,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,EAAE,qBAAqB,EAAE,MAAM,uBAAuB,CAAC;AAO9D;;;;;;GAMG;AACH,MAAM,WAAW,IAAI;IACnB,UAAU,EAAE,MAAM,CAAC;IACnB,KAAK,EAAE,MAAM,CAAC;IACd,OAAO,CAAC,EAAE,qBAAqB,CAAC;CACjC;AAED;;;;;;GAMG;AACH,qBAAa,WAAY,YAAW,gBAAgB;IAClD,SAAgB,UAAU,SAAC;IAC3B,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAS;IAC/B,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAC,CAAwB;gBAErC,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,EAAE,EAAE,IAAI;IAM1C,OAAO,CAAC,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,YAAY;WAatC,QAAQ,CAAC,QAAQ,EAAE,MAAM;CAM9C;AAED;;GAEG;AACH,wBAAsB,QAAQ,CAAC,QAAQ,EAAE,MAAM,mBAE9C"}

package/dist/step/sparqlQuery.js

@@ -1,4 +1,3 @@
-import { NotSupported } from './../step.js';
 import { SparqlConstructExecutor, substituteQueryTemplates, readQueryFile, } from '../sparql/index.js';
 /**
  * Executes a SPARQL CONSTRUCT query and emits the resulting quads.
@@ -16,17 +15,13 @@ export class SparqlQuery {
         this.query = query;
         this.fetcher = fetcher;
     }
-    async execute(dataset) {
-        const distribution = dataset.getSparqlDistribution();
-        if (distribution === null || !distribution.isValid) {
-            return new NotSupported('No SPARQL distribution available');
-        }
+    async execute(dataset, distribution) {
         const substituted = substituteQueryTemplates(this.query, distribution, dataset);
         const executor = new SparqlConstructExecutor({
             query: substituted,
             fetcher: this.fetcher,
         });
-        return await executor.execute(dataset);
+        return await executor.execute(dataset, distribution);
     }
     static async fromFile(filename) {
         return new this({

package/dist/step.d.ts

@@ -9,13 +9,13 @@ export type Step = DataEmittingStep | SingleStep;
  * Failure is expressed by emitting an error event; success by the end event.
  */
 export interface DataEmittingStep extends AbstractStep {
-    execute(dataset: Dataset): Promise<Stream | NotSupported>;
+    execute(dataset: Dataset, distribution: Distribution): Promise<Stream | NotSupported>;
 }
 /**
  * A pipeline step that executes an operation without emitting data.
  */
 export interface SingleStep extends AbstractStep {
-    execute(dataset: Dataset): Promise<NotSupported | Failure | Success>;
+    execute(dataset: Dataset, distribution?: Distribution): Promise<NotSupported | Failure | Success>;
 }
 export interface Finishable {
     finish(): Promise<void>;

package/dist/step.d.ts.map

@@ -1 +1 @@
-{"version":3,"file":"step.d.ts","sourceRoot":"","sources":["../src/step.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAE3C,UAAU,YAAY;IACpB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;CAC7B;AAED,MAAM,MAAM,IAAI,GAAG,gBAAgB,GAAG,UAAU,CAAC;AAEjD;;;GAGG;AACH,MAAM,WAAW,gBAAiB,SAAQ,YAAY;IACpD,OAAO,CAAC,OAAO,EAAE,OAAO,GAAG,OAAO,CAAC,MAAM,GAAG,YAAY,CAAC,CAAC;CAC3D;AAED;;GAEG;AACH,MAAM,WAAW,UAAW,SAAQ,YAAY;IAC9C,OAAO,CAAC,OAAO,EAAE,OAAO,GAAG,OAAO,CAAC,YAAY,GAAG,OAAO,GAAG,OAAO,CAAC,CAAC;CACtE;AAED,MAAM,WAAW,UAAU;IACzB,MAAM,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CACzB;AAED;;;;;GAKG;AACH,qBAAa,OAAO;aAEA,YAAY,EAAE,YAAY;aAC1B,OAAO,CAAC,EAAE,MAAM;gBADhB,YAAY,EAAE,YAAY,EAC1B,OAAO,CAAC,EAAE,MAAM,YAAA;CAEnC;AAED;;;;;GAKG;AACH,qBAAa,OAAO;aAEA,OAAO,EAAE,OAAO;aAChB,YAAY,EAAE,YAAY;gBAD1B,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY;CAE7C;AAED;;;;GAIG;AACH,qBAAa,YAAY;aACK,OAAO,EAAE,MAAM;gBAAf,OAAO,EAAE,MAAM;CAC5C"}
+{"version":3,"file":"step.d.ts","sourceRoot":"","sources":["../src/step.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AACrD,OAAO,KAAK,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAE3C,UAAU,YAAY;IACpB,QAAQ,CAAC,UAAU,EAAE,MAAM,CAAC;CAC7B;AAED,MAAM,MAAM,IAAI,GAAG,gBAAgB,GAAG,UAAU,CAAC;AAEjD;;;GAGG;AACH,MAAM,WAAW,gBAAiB,SAAQ,YAAY;IACpD,OAAO,CACL,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY,GACzB,OAAO,CAAC,MAAM,GAAG,YAAY,CAAC,CAAC;CACnC;AAED;;GAEG;AACH,MAAM,WAAW,UAAW,SAAQ,YAAY;IAC9C,OAAO,CACL,OAAO,EAAE,OAAO,EAChB,YAAY,CAAC,EAAE,YAAY,GAC1B,OAAO,CAAC,YAAY,GAAG,OAAO,GAAG,OAAO,CAAC,CAAC;CAC9C;AAED,MAAM,WAAW,UAAU;IACzB,MAAM,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CACzB;AAED;;;;;GAKG;AACH,qBAAa,OAAO;aAEA,YAAY,EAAE,YAAY;aAC1B,OAAO,CAAC,EAAE,MAAM;gBADhB,YAAY,EAAE,YAAY,EAC1B,OAAO,CAAC,EAAE,MAAM,YAAA;CAEnC;AAED;;;;;GAKG;AACH,qBAAa,OAAO;aAEA,OAAO,EAAE,OAAO;aAChB,YAAY,EAAE,YAAY;gBAD1B,OAAO,EAAE,OAAO,EAChB,YAAY,EAAE,YAAY;CAE7C;AAED;;;;GAIG;AACH,qBAAa,YAAY;aACK,OAAO,EAAE,MAAM;gBAAf,OAAO,EAAE,MAAM;CAC5C"}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@lde/pipeline",
-  "version": "0.6.14",
+  "version": "0.6.16",
   "repository": {
     "url": "https://github.com/ldengine/lde",
     "directory": "packages/pipeline"

package/dist/analyzer.d.ts

@@ -1,36 +0,0 @@
-import { Dataset } from '@lde/dataset';
-import type { DatasetCore } from '@rdfjs/types';
-import { NotSupported } from './step.js';
-export { NotSupported } from './step.js';
-/**
- * Result of a successful analysis.
- */
-export declare class Success {
-    readonly data: DatasetCore;
-    constructor(data: DatasetCore);
-}
-/**
- * Analysis failed.
- */
-export declare class Failure {
-    readonly endpoint: URL;
-    readonly message?: string | undefined;
-    constructor(endpoint: URL, message?: string | undefined);
-}
-/**
- * Interface for analyzers.
- */
-export interface Analyzer {
-    readonly name: string;
-    execute(dataset: Dataset): Promise<Success | Failure | NotSupported>;
-    finish?(): Promise<void>;
-}
-/**
- * Base class for analyzers with default implementations.
- */
-export declare abstract class BaseAnalyzer implements Analyzer {
-    abstract readonly name: string;
-    abstract execute(dataset: Dataset): Promise<Success | Failure | NotSupported>;
-    finish(): Promise<void>;
-}
-//# sourceMappingURL=analyzer.d.ts.map

package/dist/analyzer.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"analyzer.d.ts","sourceRoot":"","sources":["../src/analyzer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,cAAc,CAAC;AACvC,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,cAAc,CAAC;AAChD,OAAO,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AAEzC,OAAO,EAAE,YAAY,EAAE,MAAM,WAAW,CAAC;AAEzC;;GAEG;AACH,qBAAa,OAAO;aACU,IAAI,EAAE,WAAW;gBAAjB,IAAI,EAAE,WAAW;CAC9C;AAED;;GAEG;AACH,qBAAa,OAAO;aAEA,QAAQ,EAAE,GAAG;aACb,OAAO,CAAC,EAAE,MAAM;gBADhB,QAAQ,EAAE,GAAG,EACb,OAAO,CAAC,EAAE,MAAM,YAAA;CAEnC;AAED;;GAEG;AACH,MAAM,WAAW,QAAQ;IACvB,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IACtB,OAAO,CAAC,OAAO,EAAE,OAAO,GAAG,OAAO,CAAC,OAAO,GAAG,OAAO,GAAG,YAAY,CAAC,CAAC;IACrE,MAAM,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CAC1B;AAED;;GAEG;AACH,8BAAsB,YAAa,YAAW,QAAQ;IACpD,QAAQ,CAAC,QAAQ,CAAC,IAAI,EAAE,MAAM,CAAC;IAC/B,QAAQ,CAAC,OAAO,CAAC,OAAO,EAAE,OAAO,GAAG,OAAO,CAAC,OAAO,GAAG,OAAO,GAAG,YAAY,CAAC;IAEvE,MAAM,IAAI,OAAO,CAAC,IAAI,CAAC;CAG9B"}

package/dist/analyzer.js

@@ -1,29 +0,0 @@
-export { NotSupported } from './step.js';
-/**
- * Result of a successful analysis.
- */
-export class Success {
-    data;
-    constructor(data) {
-        this.data = data;
-    }
-}
-/**
- * Analysis failed.
- */
-export class Failure {
-    endpoint;
-    message;
-    constructor(endpoint, message) {
-        this.endpoint = endpoint;
-        this.message = message;
-    }
-}
-/**
- * Base class for analyzers with default implementations.
- */
-export class BaseAnalyzer {
-    async finish() {
-        // Default no-op implementation.
-    }
-}

package/dist/distribution/analyzer.d.ts

@@ -1,58 +0,0 @@
-import { Dataset } from '@lde/dataset';
-import { Importer, ImportFailed, ImportSuccessful, NotSupported } from '@lde/sparql-importer';
-import { Store } from 'n3';
-export type { Importer };
-export { ImportFailed, ImportSuccessful, NotSupported };
-/**
- * Extended importer interface with optional cleanup method.
- */
-export interface ImporterWithFinish extends Importer {
-    finish?(): Promise<void>;
-}
-export interface DistributionAnalyzerOptions {
-    /**
-     * Optional importer for loading data dumps when no SPARQL endpoint is available.
-     */
-    importer?: ImporterWithFinish;
-    /**
-     * Timeout for probe requests in milliseconds.
-     * @default 5000
-     */
-    timeout?: number;
-}
-/**
- * Result indicating the analyzer could not find a usable distribution.
- */
-export declare class NoDistributionAvailable {
-    readonly message: string;
-    constructor(message: string);
-}
-/**
- * Analyzes dataset distributions by probing their availability.
- *
- * - Probes SPARQL endpoints with a simple SELECT query
- * - Probes data dumps with HEAD/GET requests
- * - Records probe results as RDF (schema:Action)
- * - Updates distribution metadata (isValid, lastModified, byteSize)
- * - Optionally imports data dumps if no SPARQL endpoint is available
- */
-export declare class DistributionAnalyzer {
-    readonly name = "distribution";
-    private readonly importer?;
-    private readonly timeout;
-    constructor(options?: DistributionAnalyzerOptions);
-    /**
-     * Analyze all distributions of a dataset.
-     *
-     * @returns Store with probe results as RDF, or NoDistributionAvailable if no usable distribution found
-     */
-    execute(dataset: Dataset): Promise<Store | NoDistributionAvailable>;
-    /**
-     * Cleanup resources (e.g., importer connections).
-     */
-    finish(): Promise<void>;
-    private buildProbeResultsRdf;
-    private addSuccessTriples;
-    private addImportError;
-}
-//# sourceMappingURL=analyzer.d.ts.map

package/dist/distribution/analyzer.d.ts.map

@@ -1 +0,0 @@
-{"version":3,"file":"analyzer.d.ts","sourceRoot":"","sources":["../../src/distribution/analyzer.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAgB,MAAM,cAAc,CAAC;AACrD,OAAO,EACL,QAAQ,EACR,YAAY,EACZ,gBAAgB,EAChB,YAAY,EACb,MAAM,sBAAsB,CAAC;AAC9B,OAAO,EAAe,KAAK,EAAE,MAAM,IAAI,CAAC;AASxC,YAAY,EAAE,QAAQ,EAAE,CAAC;AACzB,OAAO,EAAE,YAAY,EAAE,gBAAgB,EAAE,YAAY,EAAE,CAAC;AAWxD;;GAEG;AACH,MAAM,WAAW,kBAAmB,SAAQ,QAAQ;IAClD,MAAM,CAAC,IAAI,OAAO,CAAC,IAAI,CAAC,CAAC;CAC1B;AAED,MAAM,WAAW,2BAA2B;IAC1C;;OAEG;IACH,QAAQ,CAAC,EAAE,kBAAkB,CAAC;IAE9B;;;OAGG;IACH,OAAO,CAAC,EAAE,MAAM,CAAC;CAClB;AAED;;GAEG;AACH,qBAAa,uBAAuB;aACN,OAAO,EAAE,MAAM;gBAAf,OAAO,EAAE,MAAM;CAC5C;AAED;;;;;;;;GAQG;AACH,qBAAa,oBAAoB;IAC/B,SAAgB,IAAI,kBAAkB;IACtC,OAAO,CAAC,QAAQ,CAAC,QAAQ,CAAC,CAAqB;IAC/C,OAAO,CAAC,QAAQ,CAAC,OAAO,CAAS;gBAErB,OAAO,CAAC,EAAE,2BAA2B;IAKjD;;;;OAIG;IACG,OAAO,CAAC,OAAO,EAAE,OAAO,GAAG,OAAO,CAAC,KAAK,GAAG,uBAAuB,CAAC;IAmCzE;;OAEG;IACG,MAAM,IAAI,OAAO,CAAC,IAAI,CAAC;IAI7B,OAAO,CAAC,oBAAoB;IAwC5B,OAAO,CAAC,iBAAiB;IAiDzB,OAAO,CAAC,cAAc;CAevB"}

package/dist/distribution/analyzer.js

@@ -1,120 +0,0 @@
-import { Distribution } from '@lde/dataset';
-import { ImportFailed, ImportSuccessful, NotSupported, } from '@lde/sparql-importer';
-import { DataFactory, Store } from 'n3';
-import { probe, NetworkError, SparqlProbeResult, } from './probe.js';
-export { ImportFailed, ImportSuccessful, NotSupported };
-const { quad, namedNode, blankNode, literal } = DataFactory;
-// Namespace prefixes
-const RDF = 'http://www.w3.org/1999/02/22-rdf-syntax-ns#';
-const SCHEMA = 'https://schema.org/';
-const VOID = 'http://rdfs.org/ns/void#';
-const XSD = 'http://www.w3.org/2001/XMLSchema#';
-const HTTP_STATUS = 'https://www.w3.org/2011/http-statusCodes#';
-/**
- * Result indicating the analyzer could not find a usable distribution.
- */
-export class NoDistributionAvailable {
-    message;
-    constructor(message) {
-        this.message = message;
-    }
-}
-/**
- * Analyzes dataset distributions by probing their availability.
- *
- * - Probes SPARQL endpoints with a simple SELECT query
- * - Probes data dumps with HEAD/GET requests
- * - Records probe results as RDF (schema:Action)
- * - Updates distribution metadata (isValid, lastModified, byteSize)
- * - Optionally imports data dumps if no SPARQL endpoint is available
- */
-export class DistributionAnalyzer {
-    name = 'distribution';
-    importer;
-    timeout;
-    constructor(options) {
-        this.importer = options?.importer;
-        this.timeout = options?.timeout ?? 5000;
-    }
-    /**
-     * Analyze all distributions of a dataset.
-     *
-     * @returns Store with probe results as RDF, or NoDistributionAvailable if no usable distribution found
-     */
-    async execute(dataset) {
-        const results = await Promise.all(dataset.distributions.map((distribution) => probe(distribution, this.timeout)));
-        const store = this.buildProbeResultsRdf(results, dataset);
-        // If no SPARQL endpoint available, try to import a data dump
-        if (dataset.getSparqlDistribution() === null && this.importer) {
-            const importResult = await this.importer.import(dataset);
-            if (importResult instanceof ImportSuccessful) {
-                // Add imported SPARQL distribution to dataset so subsequent steps can use it
-                const distribution = Distribution.sparql(importResult.distribution.accessUrl, importResult.identifier);
-                dataset.distributions.push(distribution);
-            }
-            else if (importResult instanceof ImportFailed) {
-                // Record import error in the store
-                this.addImportError(store, importResult);
-            }
-        }
-        if (dataset.getSparqlDistribution() === null) {
-            return new NoDistributionAvailable('No SPARQL endpoint or importable data dump available');
-        }
-        return store;
-    }
-    /**
-     * Cleanup resources (e.g., importer connections).
-     */
-    async finish() {
-        await this.importer?.finish?.();
-    }
-    buildProbeResultsRdf(results, dataset) {
-        const store = new Store();
-        for (const result of results) {
-            const action = blankNode();
-            // Base action triples
-            store.addQuads([
-                quad(action, namedNode(`${RDF}type`), namedNode(`${SCHEMA}Action`)),
-                quad(action, namedNode(`${SCHEMA}target`), namedNode(result.url)),
-            ]);
-            if (result instanceof NetworkError) {
-                store.addQuad(action, namedNode(`${SCHEMA}error`), literal(result.message));
-            }
-            else if (result.isSuccess()) {
-                this.addSuccessTriples(store, action, result, dataset);
-            }
-            else {
-                // HTTP error
-                const statusUri = `${HTTP_STATUS}${result.statusText.replace(/ /g, '')}`;
-                store.addQuad(action, namedNode(`${SCHEMA}error`), namedNode(statusUri));
-            }
-        }
-        return store;
-    }
-    addSuccessTriples(store, action, result, dataset) {
-        const distributionUrl = namedNode(result.url);
-        store.addQuad(action, namedNode(`${SCHEMA}result`), distributionUrl);
-        if (result.lastModified) {
-            store.addQuad(distributionUrl, namedNode(`${SCHEMA}dateModified`), literal(result.lastModified.toISOString(), namedNode(`${XSD}dateTime`)));
-        }
-        if (result instanceof SparqlProbeResult) {
-            store.addQuad(namedNode(dataset.iri.toString()), namedNode(`${VOID}sparqlEndpoint`), distributionUrl);
-        }
-        else {
-            store.addQuad(namedNode(dataset.iri.toString()), namedNode(`${VOID}dataDump`), distributionUrl);
-            if (result.contentSize) {
-                store.addQuad(distributionUrl, namedNode(`${SCHEMA}contentSize`), literal(result.contentSize));
-            }
-            if (result.contentType) {
-                store.addQuad(distributionUrl, namedNode(`${SCHEMA}encodingFormat`), literal(result.contentType));
-            }
-        }
-    }
-    addImportError(store, importResult) {
-        // Find the action for this download URL and add the error
-        const matches = store.match(null, namedNode(`${SCHEMA}target`), namedNode(importResult.distribution.accessUrl.toString()));
-        for (const match of matches) {
-            store.addQuad(match.subject, namedNode(`${SCHEMA}error`), literal(importResult.error));
-        }
-    }
-}