@graphty/graph-samples 0.0.1 → 0.1.2

package/dist/src/random/trig.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"trig.js","sourceRoot":"","sources":["../../../src/random/trig.ts"],"names":[],"mappings":"AAAA;;;;;;;;GAQG;AAEH,MAAM,IAAI,GAAG,yBAAyB,CAAC;AAEvC,MAAM,EAAE,GAAG,CAAC,yBAAyB,CAAC;AACtC,MAAM,EAAE,GAAG,yBAAyB,CAAC;AACrC,MAAM,EAAE,GAAG,CAAC,yBAAyB,CAAC;AACtC,MAAM,EAAE,GAAG,yBAAyB,CAAC;AACrC,MAAM,EAAE,GAAG,CAAC,yBAAyB,CAAC;AACtC,MAAM,EAAE,GAAG,0BAA0B,CAAC;AAEtC,MAAM,EAAE,GAAG,yBAAyB,CAAC;AACrC,MAAM,EAAE,GAAG,CAAC,yBAAyB,CAAC;AACtC,MAAM,EAAE,GAAG,yBAAyB,CAAC;AACrC,MAAM,EAAE,GAAG,CAAC,yBAAyB,CAAC;AACtC,MAAM,EAAE,GAAG,yBAAyB,CAAC;AACrC,MAAM,EAAE,GAAG,CAAC,0BAA0B,CAAC;AAEvC;;;;GAIG;AACH,SAAS,SAAS,CAAC,CAAS;IACxB,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAChB,MAAM,CAAC,GAAG,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC,CAAC;IACvD,OAAO,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;AACpC,CAAC;AAED;;;;GAIG;AACH,SAAS,SAAS,CAAC,CAAS;IACxB,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAChB,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAChB,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,GAAG,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,CAAC,EAAE,GAAG,CAAC,GAAG,EAAE,CAAC,CAAC,CAAC;IAC1E,MAAM,EAAE,GAAG,GAAG,GAAG,CAAC,CAAC;IACnB,MAAM,CAAC,GAAG,GAAG,GAAG,EAAE,CAAC;IACnB,OAAO,CAAC,GAAG,CAAC,GAAG,GAAG,CAAC,GAAG,EAAE,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC;AACtC,CAAC;AAED;;;;GAIG;AACH,MAAM,UAAU,aAAa,CAAC,CAAS,EAAE,GAAiB;IACtD,MAAM,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;IAChB,MAAM,MAAM,GAAG,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAC7B,MAAM,CAAC,GAAG,CAAC,CAAC,GAAG,MAAM,CAAC,GAAG,IAAI,CAAC;IAC9B,8FAA8F;IAC9F,MAAM,GAAG,GAAG,CAAC,MAAM,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC;IAC/B,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC7B,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;IAC5C,MAAM,CAAC,GAAG,GAAG,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC;IAC5C,QAAQ,CAAC,MAAM,IAAI,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC;QACxB,KAAK,CAAC;YACF,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACX,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACX,MAAM;QACV,KAAK,CAAC;YACF,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACZ,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACX,MAAM;QACV,KAAK,CAAC;YACF,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACZ,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;YACZ,MAAM;QACV;YACI,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC;YACX,GAAG,CAAC,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC;IACpB,CAAC;AACL,CAAC"}

package/dist/src/types.d.ts

@@ -0,0 +1,28 @@
+import { type ColumnInput, type F32, type TypedArrayData, type U32 } from "@graphty/graph-format";
+/**
+ * A graph as typed arrays: the output of every generator and every dataset loader.
+ *
+ * It is structurally an `EdgeArraysInput` of @graphty/graph-format, so `fromEdgeArrays(graph)`
+ * builds a snapshot from it in one call, and every array is a plain typed array that a Web Worker
+ * can transfer. Edge `e` joins `src[e]` and `dst[e]`, both node indices in `[0, nodeCount)`.
+ */
+export interface SampleGraph {
+    /** Whether the edges are directed (src to dst). */
+    readonly directed: boolean;
+    /** The number of nodes; isolated nodes count. */
+    readonly nodeCount: number;
+    /** Source node index of every edge. */
+    readonly src: U32;
+    /** Target node index of every edge. */
+    readonly dst: U32;
+    /** Per-edge weights, when the graph is weighted. */
+    readonly weights?: F32 | undefined;
+    /** External node ids in index order, when the graph has meaningful names for its nodes. */
+    readonly ids?: readonly string[] | undefined;
+    /**
+     * Node attribute columns in index order, keyed by name: ground truth such as `community`
+     * (u32), `side` (u8, bipartite) or `layer` (u32, layered DAG), and a dataset's own attributes.
+     */
+    readonly nodeColumns?: Readonly<Record<string, TypedArrayData | ColumnInput>> | undefined;
+}
+//# sourceMappingURL=types.d.ts.map

package/dist/src/types.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.d.ts","sourceRoot":"","sources":["../../src/types.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,KAAK,WAAW,EAAE,KAAK,GAAG,EAAE,KAAK,cAAc,EAAE,KAAK,GAAG,EAAE,MAAM,uBAAuB,CAAC;AAElG;;;;;;GAMG;AACH,MAAM,WAAW,WAAW;IACxB,mDAAmD;IACnD,QAAQ,CAAC,QAAQ,EAAE,OAAO,CAAC;IAC3B,iDAAiD;IACjD,QAAQ,CAAC,SAAS,EAAE,MAAM,CAAC;IAC3B,uCAAuC;IACvC,QAAQ,CAAC,GAAG,EAAE,GAAG,CAAC;IAClB,uCAAuC;IACvC,QAAQ,CAAC,GAAG,EAAE,GAAG,CAAC;IAClB,oDAAoD;IACpD,QAAQ,CAAC,OAAO,CAAC,EAAE,GAAG,GAAG,SAAS,CAAC;IACnC,2FAA2F;IAC3F,QAAQ,CAAC,GAAG,CAAC,EAAE,SAAS,MAAM,EAAE,GAAG,SAAS,CAAC;IAC7C;;;OAGG;IACH,QAAQ,CAAC,WAAW,CAAC,EAAE,QAAQ,CAAC,MAAM,CAAC,MAAM,EAAE,cAAc,GAAG,WAAW,CAAC,CAAC,GAAG,SAAS,CAAC;CAC7F"}

package/dist/src/types.js

@@ -0,0 +1,2 @@
+export {};
+//# sourceMappingURL=types.js.map

package/dist/src/types.js.map

@@ -0,0 +1 @@
+{"version":3,"file":"types.js","sourceRoot":"","sources":["../../src/types.ts"],"names":[],"mappings":""}

package/package.json

@@ -1,8 +1,105 @@
 {
   "name": "@graphty/graph-samples",
-  "version": "0.0.1",
-  "description": "Placeholder so npm trusted publishing can be configured. The real package is published from https://github.com/graphty-org/graphty-monorepo",
+  "version": "0.1.2",
+  "private": false,
+  "description": "Seeded, reproducible graph generators and classic sample datasets as typed arrays for the @graphty/graph-format snapshot",
+  "author": "Adam Powers <apowers@ato.ms>",
+  "type": "module",
+  "main": "dist/graph-samples.js",
+  "types": "dist/graph-samples.d.ts",
+  "typesVersions": {
+    "*": {
+      "generators": [
+        "dist/generators.d.ts"
+      ],
+      "datasets/*": [
+        "dist/datasets/*.d.ts"
+      ]
+    }
+  },
+  "exports": {
+    ".": {
+      "types": "./dist/graph-samples.d.ts",
+      "import": "./dist/graph-samples.js",
+      "default": "./dist/graph-samples.js"
+    },
+    "./generators": {
+      "types": "./dist/generators.d.ts",
+      "import": "./dist/generators.js",
+      "default": "./dist/generators.js"
+    },
+    "./datasets/*": {
+      "types": "./dist/datasets/*.d.ts",
+      "import": "./dist/datasets/*.js",
+      "default": "./dist/datasets/*.js"
+    }
+  },
+  "sideEffects": false,
+  "files": [
+    "dist/",
+    "!dist/*.tsbuildinfo",
+    "src/",
+    "README.md",
+    "LICENSE",
+    "NOTICE"
+  ],
+  "publishConfig": {
+    "access": "public",
+    "provenance": true
+  },
+  "engines": {
+    "node": ">=18.19.0"
+  },
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/graphty-org/graphty-monorepo.git",
+    "directory": "graph-samples"
+  },
+  "keywords": [
+    "graph",
+    "network",
+    "generator",
+    "random-graph",
+    "erdos-renyi",
+    "barabasi-albert",
+    "watts-strogatz",
+    "stochastic-block-model",
+    "karate-club",
+    "dataset",
+    "graph-format"
+  ],
   "license": "MIT",
-  "repository": { "type": "git", "url": "git+https://github.com/graphty-org/graphty-monorepo.git", "directory": "graph-samples" },
-  "publishConfig": { "access": "public" }
-}
+  "bugs": {
+    "url": "https://github.com/graphty-org/graphty-monorepo/issues"
+  },
+  "homepage": "https://github.com/graphty-org/graphty-monorepo/tree/master/graph-samples#readme",
+  "dependencies": {
+    "@graphty/graph-format": "^1.0.5"
+  },
+  "peerDependencies": {
+    "@graphty/graph-format": "^1.0.0"
+  },
+  "devDependencies": {
+    "@vitest/coverage-v8": "^3.2.4",
+    "@vitest/ui": "^3.2.4",
+    "typescript": "^5.9.3",
+    "vite": "^7.0.5",
+    "vitest": "^3.2.4"
+  },
+  "scripts": {
+    "build": "tsc -p tsconfig.build.json",
+    "build:bundle": "node scripts/build-bundle.js",
+    "build:all": "npm run build && npm run build:bundle",
+    "lint": "eslint && tsc --noEmit -p tsconfig.json",
+    "lint:fix": "eslint --fix",
+    "typecheck": "tsc --noEmit -p tsconfig.json",
+    "test": "vitest",
+    "test:ui": "vitest --ui",
+    "test:run": "vitest run",
+    "coverage": "vitest run --coverage",
+    "coverage:preview": "npx serve coverage -p ${PORT:?start it through servherd, which sets PORT}",
+    "datasets:convert": "node scripts/convert-datasets.mjs",
+    "datasets:hosted": "node scripts/build-hosted.mjs",
+    "ready:commit": "npm run build:all && npm run lint && npm run test:run"
+  }
+}

package/src/datasets/build.ts

@@ -0,0 +1,140 @@
+/**
+ * The shapes shared by every dataset module: its metadata, the compact form its generated data
+ * file is written in, and the one function that turns that form into a {@link SampleGraph}.
+ */
+
+import { type ColumnInput, type TypedArrayData } from "@graphty/graph-format";
+
+import { type SampleGraph } from "../types.js";
+
+/** What a dataset is, where it came from and under which terms. */
+export interface DatasetMeta {
+    /** The subpath name: `@graphty/graph-samples/datasets/<name>`. */
+    readonly name: string;
+    /** The human-readable title. */
+    readonly title: string;
+    /** One or two sentences on what the nodes and edges are. */
+    readonly description: string;
+    /** The reference to cite. */
+    readonly citation: string;
+    /** The URL the data was converted from. */
+    readonly source: string;
+    /** The license as known, or "unclear" with what is known. */
+    readonly license: string;
+    /** The node count. */
+    readonly nodes: number;
+    /** The edge count. */
+    readonly edges: number;
+    /** Whether the edges are directed. */
+    readonly directed: boolean;
+    /** Whether the edges carry weights. */
+    readonly weighted: boolean;
+    /** The node columns, name to description. */
+    readonly attributes: Readonly<Record<string, string>>;
+    /** The node column holding the ground-truth grouping, or null when there is none. */
+    readonly groundTruth: string | null;
+    /** The algorithms and layouts it shows off. */
+    readonly showcases: readonly string[];
+    /**
+     * Where the graph data lives: "bundled" (the subpath `@graphty/graph-samples/datasets/<name>`;
+     * the meaning when absent) or "remote" (on graphty.app, loaded with `fetchDataset(name)`).
+     */
+    readonly hosting?: "bundled" | "remote";
+    /** A remote dataset's download size: the bytes of its `.gsnp.gz` file. */
+    readonly bytes?: number;
+    /** A remote dataset's checksum: the SHA-256 of its `.gsnp.gz` file, lowercase hex. */
+    readonly sha256?: string;
+}
+
+/** A categorical column: codes into a list of category names. */
+interface DictColumnData {
+    readonly dtype: "dict";
+    readonly categories: readonly string[];
+    readonly codes: readonly number[];
+}
+
+/** A text column, e.g. the display label. */
+interface StringColumnData {
+    readonly dtype: "string";
+    readonly role?: "label";
+    readonly values: readonly string[];
+}
+
+/** A small unsigned integer column, e.g. a bipartite side. */
+interface U8ColumnData {
+    readonly dtype: "u8";
+    readonly values: readonly number[];
+}
+
+/** A number column, e.g. a latitude. */
+interface F64ColumnData {
+    readonly dtype: "f64";
+    readonly values: readonly number[];
+}
+
+/** An unsigned 32-bit integer column, e.g. a population. */
+interface U32ColumnData {
+    readonly dtype: "u32";
+    readonly values: readonly number[];
+}
+
+/** The compact form of a dataset, as scripts/convert-datasets.mjs writes it into data.ts. */
+export interface DatasetData {
+    readonly directed: boolean;
+    readonly nodeCount: number;
+    /** External node ids in index order, or null when the index is the id. */
+    readonly ids: readonly string[] | null;
+    /** Flat edge endpoints: src0, dst0, src1, dst1, ... */
+    readonly edges: readonly number[];
+    readonly weights: readonly number[] | null;
+    readonly columns: Readonly<Record<string, DictColumnData | StringColumnData | U8ColumnData | F64ColumnData | U32ColumnData>>;
+}
+
+/**
+ * Expand a dataset's compact data into typed arrays and graph-format column inputs.
+ * @param data - the generated data
+ * @returns the graph
+ */
+export function buildDataset(data: DatasetData): SampleGraph {
+    const m = data.edges.length / 2;
+    const src = new Uint32Array(m);
+    const dst = new Uint32Array(m);
+    for (let e = 0; e < m; e++) {
+        src[e] = data.edges[2 * e];
+        dst[e] = data.edges[2 * e + 1];
+    }
+    const nodeColumns: Record<string, TypedArrayData | ColumnInput> = {};
+    for (const [name, column] of Object.entries(data.columns)) {
+        switch (column.dtype) {
+            case "dict":
+                nodeColumns[name] = {
+                    data: column.codes.map((code) => column.categories[code]),
+                    decl: { dtype: "dict", options: column.categories },
+                };
+                break;
+            case "string":
+                nodeColumns[name] = { data: column.values, decl: { dtype: "string", role: column.role } };
+                break;
+            case "u8":
+                nodeColumns[name] = Uint8Array.from(column.values);
+                break;
+            case "u32":
+                nodeColumns[name] = Uint32Array.from(column.values);
+                break;
+            case "f64":
+                nodeColumns[name] = Float64Array.from(column.values);
+                break;
+            default:
+                throw new TypeError(`unknown column dtype in dataset column "${name}"`);
+        }
+    }
+    return {
+        directed: data.directed,
+        nodeCount: data.nodeCount,
+        src,
+        dst,
+        ...(data.weights === null ? {} : { weights: Float32Array.from(data.weights) }),
+        ...(data.ids === null ? {} : { ids: data.ids }),
+        ...(Object.keys(nodeColumns).length === 0 ? {} : { nodeColumns }),
+    };
+}

package/src/datasets/catalog.ts

@@ -0,0 +1,35 @@
+import { type DatasetMeta } from "./build.js";
+import { celegansNeuralMeta } from "./celegans-neural/meta.js";
+import { contiguousUsaMeta } from "./contiguous-usa/meta.js";
+import { davisSouthernWomenMeta } from "./davis-southern-women/meta.js";
+import { dolphinsMeta } from "./dolphins/meta.js";
+import { florentineFamiliesMeta } from "./florentine-families/meta.js";
+import { footballMeta } from "./football/meta.js";
+import { HOSTED_DATASETS } from "./hosted.js";
+import { karateMeta } from "./karate/meta.js";
+import { knuthMilesMeta } from "./knuth-miles/meta.js";
+import { lesMiserablesMeta } from "./les-miserables/meta.js";
+import { openflightsMeta } from "./openflights/meta.js";
+import { politicalBlogsMeta } from "./political-blogs/meta.js";
+import { politicalBooksMeta } from "./political-books/meta.js";
+
+/**
+ * The metadata of every dataset (no graph data). A bundled dataset's edges load only from its own
+ * subpath, `@graphty/graph-samples/datasets/<name>`; a dataset with `hosting: "remote"` loads with
+ * `fetchDataset(name)`.
+ */
+export const DATASETS: readonly DatasetMeta[] = [
+    karateMeta,
+    florentineFamiliesMeta,
+    davisSouthernWomenMeta,
+    lesMiserablesMeta,
+    footballMeta,
+    politicalBooksMeta,
+    dolphinsMeta,
+    contiguousUsaMeta,
+    knuthMilesMeta,
+    celegansNeuralMeta,
+    politicalBlogsMeta,
+    openflightsMeta,
+    ...HOSTED_DATASETS,
+];