[Query API] Provide Access to Clustering

src/cli/repl/commands/repl-query.ts

@@ -21,6 +21,7 @@ import { graphToMermaidUrl } from '../../../util/mermaid/dfg';
 import { normalizedAstToMermaidUrl } from '../../../util/mermaid/ast';
 
 import { printAsMs } from '../../../util/time';
+import { textWithTooltip } from '../../../documentation/doc-util/doc-hover-over';
 
 async function getDataflow(shell: RShell, remainingLine: string) {
 	return await new PipelineExecutor(DEFAULT_DATAFLOW_PIPELINE, {
@@ -131,7 +132,7 @@ function summarizeIdsIfTooLong(ids: readonly NodeId[]) {
 	if(i < ids.length) {
 		acc += '... (see JSON below)';
 	}
-	return acc;
+	return textWithTooltip(acc, JSON.stringify(ids));
 }
 
 export function asciiSummaryOfQueryResult(formatter: OutputFormatter, totalInMs: number, results: QueryResults<SupportedQueryTypes>, processed: PipelineOutput<typeof DEFAULT_DATAFLOW_PIPELINE>): string {
@@ -161,6 +162,17 @@ export function asciiSummaryOfQueryResult(formatter: OutputFormatter, totalInMs:
 			result.push(`Query: ${bold(query, formatter)} (${printAsMs(out['.meta'].timing, 0)})`);
 			result.push(`   ╰ [Normalized AST](${normalizedAstToMermaidUrl(out.normalized.ast)})`);
 			continue;
+		} else if(query === 'dataflow-cluster') {
+			const out = queryResults as QueryResults<'dataflow-cluster'>['dataflow-cluster'];
+			result.push(`Query: ${bold(query, formatter)} (${out['.meta'].timing.toFixed(0)}ms)`);
+			result.push(`   ╰ Found ${out.clusters.length} cluster${out.clusters.length === 1 ? '': 's'}`);
+			for(const cluster of out.clusters) {
+				const unknownSideEffects = cluster.hasUnknownSideEffects ? '(has unknown side effect)' : '';
+				result.push(`      ╰ ${unknownSideEffects} {${summarizeIdsIfTooLong(cluster.members)}} ([marked](${
+					graphToMermaidUrl(processed.dataflow.graph, false, new Set(cluster.members))
+				}))`);
+			}
+			continue;
 		}
 
 		result.push(`Query: ${bold(query, formatter)}`);

src/documentation/doc-util/doc-dfg.ts

@@ -32,11 +32,12 @@ export interface PrintDataflowGraphOptions {
 	readonly codeOpen?:           boolean;
 	readonly exposeResult?:       boolean;
 	readonly switchCodeAndGraph?: boolean;
+	readonly hideEnvInMermaid?:   boolean;
 }
 
 export async function printDfGraphForCode(shell: RShell, code: string, options: PrintDataflowGraphOptions & { exposeResult: true }): Promise<[string, PipelineOutput<typeof DEFAULT_DATAFLOW_PIPELINE>]>;
 export async function printDfGraphForCode(shell: RShell, code: string, options?: PrintDataflowGraphOptions & { exposeResult?: false | undefined }): Promise<string>;
-export async function printDfGraphForCode(shell: RShell, code: string, { mark, showCode = true, codeOpen = false, exposeResult, switchCodeAndGraph = false }: PrintDataflowGraphOptions = {}): Promise<string | [string, PipelineOutput<typeof DEFAULT_DATAFLOW_PIPELINE>]> {
+export async function printDfGraphForCode(shell: RShell, code: string, { mark, showCode = true, codeOpen = false, exposeResult, switchCodeAndGraph = false, hideEnvInMermaid = false }: PrintDataflowGraphOptions = {}): Promise<string | [string, PipelineOutput<typeof DEFAULT_DATAFLOW_PIPELINE>]> {
 	const now = performance.now();
 	const result = await new PipelineExecutor(DEFAULT_DATAFLOW_PIPELINE, {
 		shell,
@@ -48,7 +49,7 @@ export async function printDfGraphForCode(shell: RShell, code: string, { mark, s
 		guard(showCode, 'can not switch code and graph if code is not shown');
 	}
 
-	const metaInfo = `The analysis required _${printAsMs(duration)}_ (including parsing and normalization) within the generation environment.`;
+	const metaInfo = `The analysis required _${printAsMs(duration)}_ (incl. parse and normalize) within the generation environment.`;
 	const dfGraph = printDfGraph(result.dataflow.graph, mark);
 	let resultText = '\n\n';
 
@@ -73,8 +74,9 @@ ${switchCodeAndGraph ? dfGraph : codeText}
 
 \`\`\`
 ${graphToMermaid({
-		graph:  result.dataflow.graph,
-		prefix: 'flowchart LR'
+		graph:               result.dataflow.graph,
+		prefix:              'flowchart LR', 
+		includeEnvironments: !hideEnvInMermaid
 	}).string}
 \`\`\`
 

src/documentation/doc-util/doc-query.ts

@@ -103,7 +103,7 @@ function linkify(name: string) {
 }
 
 export function tocForQueryType(type: 'active' | 'virtual') {
-	const queries = RegisteredQueries[type];
+	const queries = [...RegisteredQueries[type].entries()].sort(([,{ name: a }], [, { name: b }]) => a.localeCompare(b));
 	const result: string[] = [];
 	for(const [id, { name, shortDescription }] of queries) {
 		result.push(`1. [${name}](#${linkify(name)}) (\`${id}\`):\\\n    ${shortDescription}`);

src/documentation/print-query-wiki.ts

@@ -18,6 +18,7 @@ import { codeBlock } from './doc-util/doc-code';
 import { executeDataflowQuery } from '../queries/catalog/dataflow-query/dataflow-query-executor';
 import { executeIdMapQuery } from '../queries/catalog/id-map-query/id-map-query-executor';
 import { executeNormalizedAstQuery } from '../queries/catalog/normalized-ast-query/normalized-ast-query-executor';
+import { executeDataflowClusterQuery } from '../queries/catalog/cluster-query/cluster-query-executor';
 
 
 registerQueryDocumentation('call-context', {
@@ -130,6 +131,27 @@ ${
 	}
 });
 
+registerQueryDocumentation('dataflow-cluster', {
+	name:             'Dataflow Cluster Query',
+	type:             'active',
+	shortDescription: 'Calculates and returns all the clusters present in the dataflow graph.',
+	functionName:     executeDataflowClusterQuery.name,
+	functionFile:     '../queries/catalog/cluster-query/cluster-query-executor.ts',
+	buildExplanation: async(shell: RShell) => {
+		return `
+This query automatically calculates clusters in flowR's dataflow graph and returns a list of all clusters
+found.
+
+Using the example code from above, the following query returns all clusters:
+${
+	await showQuery(shell, exampleQueryCode, [{
+		type: 'dataflow-cluster'
+	}], { showCode: false })
+}
+		`;
+	}
+});
+
 registerQueryDocumentation('id-map', {
 	name:             'Id-Map Query',
 	type:             'active',

src/queries/catalog/cluster-query/cluster-query-executor.ts

@@ -0,0 +1,20 @@
+import type { BasicQueryData } from '../../query';
+import { log } from '../../../util/log';
+import type { DataflowClusterQuery, DataflowClusterQueryResult } from './cluster-query-format';
+import { findAllClusters } from '../../../dataflow/cluster';
+
+
+export function executeDataflowClusterQuery({ graph }: BasicQueryData, queries: readonly DataflowClusterQuery[]): DataflowClusterQueryResult {
+	if(queries.length !== 1) {
+		log.warn('The dataflow cluster query expects only up to one query, but got', queries.length);
+	}
+
+	const start = Date.now();
+	const clusters = findAllClusters(graph);
+	return {
+		'.meta': {
+			timing: Date.now() - start
+		},
+		clusters
+	};
+}

src/queries/catalog/cluster-query/cluster-query-format.ts

@@ -0,0 +1,14 @@
+import type { BaseQueryFormat, BaseQueryResult } from '../../base-query-format';
+import type { DataflowGraphClusters } from '../../../dataflow/cluster';
+
+/**
+ * Calculates and returns all clusters encountered in the dataflow graph.
+ */
+export interface DataflowClusterQuery extends BaseQueryFormat {
+	readonly type: 'dataflow-cluster';
+}
+
+export interface DataflowClusterQueryResult extends BaseQueryResult {
+	/** All clusters found in the respective dataflow */
+	readonly clusters: DataflowGraphClusters;
+}

src/queries/query-schema.ts

@@ -27,11 +27,17 @@ export const NormalizedAstQuerySchema = Joi.object({
 	type: Joi.string().valid('normalized-ast').required().description('The type of the query.'),
 }).description('The normalized AST query simply returns the normalized AST, there is no need to pass it multiple times!');
 
+export const DataflowClusterQuerySchema = Joi.object({
+	type: Joi.string().valid('dataflow-cluster').required().description('The type of the query.'),
+}).description('The cluster query calculates and returns all clusters in the dataflow graph.');
+
+
 export const SupportedQueriesSchema = Joi.alternatives(
 	CallContextQuerySchema,
 	DataflowQuerySchema,
 	IdMapQuerySchema,
-	NormalizedAstQuerySchema
+	NormalizedAstQuerySchema,
+	DataflowClusterQuerySchema
 ).description('Supported queries');
 
 export const CompoundQuerySchema = Joi.object({

src/queries/query.ts

@@ -14,8 +14,10 @@ import { executeIdMapQuery } from './catalog/id-map-query/id-map-query-executor'
 import type { IdMapQuery } from './catalog/id-map-query/id-map-query-format';
 import { executeNormalizedAstQuery } from './catalog/normalized-ast-query/normalized-ast-query-executor';
 import type {	NormalizedAstQuery } from './catalog/normalized-ast-query/normalized-ast-query-format';
+import type { DataflowClusterQuery } from './catalog/cluster-query/cluster-query-format';
+import { executeDataflowClusterQuery } from './catalog/cluster-query/cluster-query-executor';
 
-export type Query = CallContextQuery | DataflowQuery | NormalizedAstQuery | IdMapQuery;
+export type Query = CallContextQuery | DataflowQuery | NormalizedAstQuery | IdMapQuery | DataflowClusterQuery;
 
 export type QueryArgumentsWithType<QueryType extends BaseQueryFormat['type']> = Query & { type: QueryType };
 
@@ -32,10 +34,11 @@ type SupportedQueries = {
 }
 
 export const SupportedQueries = {
-	'call-context':   executeCallContextQueries,
-	'dataflow':       executeDataflowQuery,
-	'id-map':         executeIdMapQuery,
-	'normalized-ast': executeNormalizedAstQuery
+	'call-context':     executeCallContextQueries,
+	'dataflow':         executeDataflowQuery,
+	'id-map':           executeIdMapQuery,
+	'normalized-ast':   executeNormalizedAstQuery,
+	'dataflow-cluster': executeDataflowClusterQuery,
 } as const satisfies SupportedQueries;
 
 export type SupportedQueryTypes = keyof typeof SupportedQueries;

src/statistics/summarizer/post-process/clusterer.ts

@@ -19,7 +19,7 @@ export type ClusterValueInfoMap = DefaultMap<string, ContextsWithCount>
 export interface ClusterReport extends MergeableRecord {
 	/** The input file which has been clustered */
 	readonly filepath: string
-	/** Maps each context encountered (i.e., every file which contains something associated with the feature) to a unique id, used in the {@link ClusterReport#valueInfoMap | valueInfoMap}. */
+	/** Maps each context encountered (i.e., every file that contains something associated with the feature) to a unique id, used in the {@link ClusterReport#valueInfoMap|valueInfoMap}. */
 	contextIdMap:      ClusterContextIdMap
 	/**
    * Counts which contexts contained which values of a feature.

src/util/ansi.ts

@@ -69,7 +69,15 @@ export const markdownFormatter: OutputFormatter = new class implements OutputFor
 		}
 
 		input = input.replaceAll(/\\/g, '\\\\');
-		return input.replaceAll(/\n/g, '\\\n').replaceAll(/ /g, ' ');
+		let source = input.replaceAll(/\n/g, '\\\n');
+		/* repeatedly replace all spaces but only at the beginning of a line */
+		let target = source;
+		do{
+			source = target;
+			/* or replace back to front */
+			target = source.replace(/^(?<leading>(&nbsp;)*) /m, '$<leading>&nbsp;');
+		} while(target !== source);
+		return target;
 	}
 
 	public getFormatString(_options?: FormatOptions): string {

test/functionality/dataflow/query/dataflow-cluster-query-tests.ts

@@ -0,0 +1,14 @@
+import { assertQuery } from '../../_helper/query';
+import { label } from '../../_helper/label';
+import { withShell } from '../../_helper/shell';
+import { findAllClusters } from '../../../../src/dataflow/cluster';
+import type { DataflowClusterQuery } from '../../../../src/queries/catalog/cluster-query/cluster-query-format';
+
+describe('Dataflow Cluster Query', withShell(shell => {
+	function testQuery(name: string, code: string, query: readonly DataflowClusterQuery[]) {
+		assertQuery(label(name), shell, code, query, ({ dataflow }) => ({ 'dataflow-cluster': { clusters: findAllClusters(dataflow.graph) } }));
+	}
+
+	testQuery('Single Expression', 'x + 1', [{ type: 'dataflow-cluster' }]);
+	testQuery('Multiple Queries', 'x + 1', [{ type: 'dataflow-cluster' }, { type: 'dataflow-cluster' }, { type: 'dataflow-cluster' }]);
+}));