metascope 0.7.0 → 0.7.2

package/dist/lib/utilities/codemeta-helpers.js

@@ -3,12 +3,14 @@ import is from "@sindresorhus/is";
 /**
 * Helpers for building codemeta JSON-LD objects.
 *
-* Provides type-safe constructors for Person/Organization and SoftwareApplication
-* dependency nodes, plus deduplication and license URL normalization.
+* Provides type-safe constructors for Person/Organization and
+* SoftwareApplication dependency nodes, plus deduplication and license URL
+* normalization.
 */
 /**
-* Build a codemeta JSON-LD Person or Organization from flexible inputs.
-* Returns undefined if no identifying information (name, givenName+familyName, or email) is present.
+* Build a codemeta JSON-LD Person or Organization from flexible inputs. Returns
+* undefined if no identifying information (name, givenName+familyName, or
+* email) is present.
 *
 * Works with person shapes from any metascope source — the caller maps
 * source-specific field names into this common parameter object.
@@ -18,7 +20,7 @@ function toPersonOrOrgLd(options) {
 	const hasGivenName = is.nonEmptyStringAndNotWhitespace(options.givenName);
 	const hasFamilyName = is.nonEmptyStringAndNotWhitespace(options.familyName);
 	const hasEmail = is.nonEmptyStringAndNotWhitespace(options.email);
-	if (!hasName && !hasGivenName && !hasFamilyName && !hasEmail) return void 0;
+	if (!hasName && !hasGivenName && !hasFamilyName && !hasEmail) return;
 	const person = { "@type": options.type ?? "Person" };
 	if (is.nonEmptyStringAndNotWhitespace(options.id)) person["@id"] = options.id;
 	if (hasName) person.name = options.name;
@@ -33,9 +35,9 @@ function toPersonOrOrgLd(options) {
 	return person;
 }
 /**
-* Deduplicate persons by name (case-insensitive, trimmed).
-* Keeps the first occurrence, so callers should place higher-priority sources first.
-* Returns undefined if the result is empty.
+* Deduplicate persons by name (case-insensitive, trimmed). Keeps the first
+* occurrence, so callers should place higher-priority sources first. Returns
+* undefined if the result is empty.
 */
 function deduplicatePersonsOrOrgs(persons) {
 	const seen = /* @__PURE__ */ new Map();
@@ -60,8 +62,8 @@ function toDependencyLd(name, version, identifier, runtimePlatform) {
 	return dep;
 }
 /**
-* Deduplicate dependencies by name (case-insensitive).
-* Keeps the first occurrence. Returns undefined if the result is empty.
+* Deduplicate dependencies by name (case-insensitive). Keeps the first
+* occurrence. Returns undefined if the result is empty.
 */
 function deduplicateDependencies(deps) {
 	const seen = /* @__PURE__ */ new Map();
@@ -73,8 +75,8 @@ function deduplicateDependencies(deps) {
 	return result.length > 0 ? result : void 0;
 }
 /**
-* Normalize a license identifier to an SPDX URL.
-* Handles bare SPDX IDs ("MIT") and existing SPDX URLs.
+* Normalize a license identifier to an SPDX URL. Handles bare SPDX IDs ("MIT")
+* and existing SPDX URLs.
 */
 function toSpdxLicenseUrl(spdxId) {
 	return `https://spdx.org/licenses/${spdxId.replace("https://spdx.org/licenses/", "").replace("http://spdx.org/licenses/", "")}`;

package/dist/lib/utilities/fetch.js

@@ -1,8 +1,8 @@
 import { log } from "../log.js";
 //#region src/lib/utilities/fetch.ts
 /**
-* Fetch with automatic retries and exponential backoff.
-* Retries on network errors and 429/5xx responses.
+* Fetch with automatic retries and exponential backoff. Retries on network
+* errors and 429/5xx responses.
 */
 async function fetchWithRetry(url, options, maxRetries = 5) {
 	let lastError;

package/dist/lib/utilities/formatting.js

@@ -3,9 +3,10 @@ import { relative } from "node:path";
 //#region src/lib/utilities/formatting.ts
 const URL_PROTOCOL_REGEX = /^[a-z]+:\/\//i;
 /**
-* Format an absolute path as either absolute or relative, based on the `absolute` option.
-* When relative, paths identical to `basePath` are returned as `'.'`.
-* Correctly handles Windows paths (normalizes to POSIX) and ignores URLs.
+* Format an absolute path as either absolute or relative, based on the
+* `absolute` option. When relative, paths identical to `basePath` are returned
+* as `'.'`. Correctly handles Windows paths (normalizes to POSIX) and ignores
+* URLs.
 */
 function formatPath(absolutePath, basePath, absolute = DEFAULT_GET_METADATA_OPTIONS.absolute) {
 	if (URL_PROTOCOL_REGEX.test(absolutePath)) return absolutePath;

package/dist/lib/utilities/github.js

@@ -4,11 +4,11 @@ import gitUrlParse from "git-url-parse";
 * Shared GitHub utilities used across multiple sources.
 */
 /**
-* Extract a GitHub owner/repo from git config remote URLs.
-* Prefers the "origin" remote, falls back to the first GitHub remote found.
+* Extract a GitHub owner/repo from git config remote URLs. Prefers the "origin"
+* remote, falls back to the first GitHub remote found.
 */
 function getGitHubRemoteFromConfig(remotes) {
-	if (!remotes) return void 0;
+	if (!remotes) return;
 	const sorted = Object.entries(remotes).toSorted(([a], [b]) => {
 		if (a === "origin") return -1;
 		if (b === "origin") return 1;

package/dist/lib/utilities/license-identification.d.ts

@@ -15,7 +15,9 @@
  * - YAML front matter stripping
  */
 type LicenseMatch = {
-  /** Dice coefficient confidence score (0–1). */confidence: number; /** SPDX license identifier (e.g. "MIT", "Apache-2.0"). */
+  /** Dice coefficient confidence score (0–1). */confidence: number; /** Full license name (e.g. "MIT License", "Apache License 2.0"). */
+  name: string; /** Whether the license is OSI approved. */
+  osiApproved: boolean; /** SPDX license identifier (e.g. "MIT", "Apache-2.0"). */
   spdxId: string; /** SPDX license URL. */
   spdxUrl: string;
 };

package/dist/lib/utilities/license-identification.js

@@ -29,31 +29,36 @@ function identifyLicense(text) {
 	const headerMatch = identifyByHeader(text);
 	if (headerMatch) return headerMatch;
 	const normalizedInput = normalizeInput(text);
-	if (normalizedInput.length < 2) return void 0;
+	if (normalizedInput.length < 2) return;
 	const inputBigramsMap = computeBigrams(normalizedInput);
 	const inputTotal = normalizedInput.length - 1;
 	let bestMatch;
 	let bestScore = 0;
-	for (const { bigramsMap, normalized, spdxId, spdxUrl, totalBigrams } of getNormalizedLicenses()) {
-		if (normalizedInput === normalized) return {
-			confidence: 1,
-			spdxId,
-			spdxUrl
-		};
+	for (const { bigramsMap, normalized, spdxId, totalBigrams } of getNormalizedLicenses()) {
+		if (normalizedInput === normalized) return buildMatch(spdxId, 1);
 		const score = diceCoefficientCached(inputBigramsMap, inputTotal, bigramsMap, totalBigrams);
 		if (score > bestScore) {
 			bestScore = score;
-			bestMatch = {
-				confidence: score,
-				spdxId,
-				spdxUrl
-			};
+			bestMatch = buildMatch(spdxId, score);
 			if (bestScore > .98) break;
 		}
 	}
 	if (bestMatch && bestMatch.confidence >= CONFIDENCE_THRESHOLD) return bestMatch;
 }
 /**
+* Build a LicenseMatch for the given SPDX ID with the supplied confidence.
+*/
+function buildMatch(spdxId, confidence) {
+	const entry = spdxLicenseList[spdxId];
+	return {
+		confidence,
+		name: entry.name,
+		osiApproved: entry.osiApproved,
+		spdxId,
+		spdxUrl: getLicenseUrl(spdxId)
+	};
+}
+/**
 * Convert an SPDX license identifier to its canonical SPDX URL.
 */
 function spdxIdToUrl(spdxId) {
@@ -111,7 +116,6 @@ function getNormalizedLicenses() {
 			bigramsMap: computeBigrams(normalized),
 			normalized,
 			spdxId,
-			spdxUrl: getLicenseUrl(spdxId),
 			totalBigrams: normalized.length - 1
 		};
 	});
@@ -155,11 +159,7 @@ const HEADER_PATTERNS = [
 ];
 function identifyByHeader(text) {
 	const header = text.slice(0, 500);
-	for (const { pattern, spdxId } of HEADER_PATTERNS) if (pattern.test(header)) return {
-		confidence: 1,
-		spdxId,
-		spdxUrl: getLicenseUrl(spdxId)
-	};
+	for (const { pattern, spdxId } of HEADER_PATTERNS) if (pattern.test(header)) return buildMatch(spdxId, 1);
 }
 /**
 * Extracts `http(s)://...` URLs from the raw text, stripping trailing
@@ -187,7 +187,7 @@ function normalizeUrl(url) {
 	} catch {
 		return;
 	}
-	if (parsed.protocol !== "http:" && parsed.protocol !== "https:") return void 0;
+	if (parsed.protocol !== "http:" && parsed.protocol !== "https:") return;
 	const host = parsed.hostname.toLowerCase().replace(WWW_PREFIX_REGEX, "");
 	let path = parsed.pathname.toLowerCase().replace(LEGALCODE_SUFFIX_REGEX, "");
 	path = path.replace(TRAILING_SLASH_REGEX, "");
@@ -232,17 +232,13 @@ function getUrlIndex() {
 }
 function identifyByUrl(text) {
 	const matches = text.match(URL_REGEX);
-	if (!matches) return void 0;
+	if (!matches) return;
 	const index = getUrlIndex();
 	for (const raw of matches) {
 		const normalized = normalizeUrl(raw.replace(TRAILING_PUNCTUATION_REGEX, ""));
 		if (!normalized) continue;
 		const spdxId = index.get(normalized);
-		if (spdxId) return {
-			confidence: 1,
-			spdxId,
-			spdxUrl: getLicenseUrl(spdxId)
-		};
+		if (spdxId) return buildMatch(spdxId, 1);
 	}
 }
 //#endregion

package/dist/lib/utilities/schema-primitives.js

@@ -5,35 +5,38 @@ import { z } from "zod";
 * Shared Zod schema primitives for parser schemas.
 *
 * These handle common coercions:
-*   - Empty/whitespace-only strings → undefined
-*   - URL validation with empty → undefined fallback
-*   - String arrays with empty-element filtering
+*
+* - Empty/whitespace-only strings → undefined
+* - URL validation with empty → undefined fallback
+* - String arrays with empty-element filtering
 */
 /**
-* A string that treats empty or whitespace-only values as undefined.
-* Use `.optional()` on the result for optional fields.
+* A string that treats empty or whitespace-only values as undefined. The outer
+* `.optional()` lets the field be absent in parent objects (zod 4.4+ no longer
+* bubbles inner-schema optionality through `z.preprocess`).
 */
 const nonEmptyString = z.preprocess((value) => {
 	if (typeof value !== "string") return;
 	const trimmed = value.trim();
 	return trimmed.length > 0 ? trimmed : void 0;
-}, z.string().optional());
+}, z.string().optional()).optional();
 /**
-* A URL string that treats empty/whitespace as undefined.
-* Does not reject invalid URLs — just ensures non-empty.
-* Use z.url() in place of this if strict URL validation is desired.
+* A URL string that treats empty/whitespace as undefined. Does not reject
+* invalid URLs — just ensures non-empty. Use z.url() in place of this if strict
+* URL validation is desired.
 */
 const optionalUrl = nonEmptyString.describe("A URL string");
 /**
-* An array of strings, filtering out empty/whitespace-only elements.
+* An array of strings, filtering out empty/whitespace-only elements. Defaults
+* to `[]` so absent/non-array fields parse cleanly under zod 4.4+.
 */
 const stringArray = z.preprocess((value) => {
 	if (!Array.isArray(value)) return [];
 	return value.filter((item) => typeof item === "string" && item.trim().length > 0);
-}, z.array(z.string()));
+}, z.array(z.string())).default([]);
 /**
-* Parse a JSON string into a plain object.
-* Returns `undefined` for malformed JSON or non-object values (arrays, primitives).
+* Parse a JSON string into a plain object. Returns `undefined` for malformed
+* JSON or non-object values (arrays, primitives).
 */
 function parseJsonRecord(content) {
 	try {

package/dist/lib/utilities/template-helpers.js

@@ -53,7 +53,7 @@ const casePoliceDict = {
 * regardless of cardinality.
 */
 function firstOf(value) {
-	if (value === void 0) return void 0;
+	if (value === void 0) return;
 	return Array.isArray(value) ? value[0] : value;
 }
 /**
@@ -104,7 +104,7 @@ function splitCommaSeparated(value) {
 * Join an array of strings with a delimiter, filtering out undefined values.
 */
 function toDelimitedString(source, delimiter = ", ") {
-	if (source === void 0) return void 0;
+	if (source === void 0) return;
 	if (Array.isArray(source)) {
 		const filtered = source.filter((s) => s !== void 0);
 		return filtered.length > 0 ? filtered.join(delimiter) : void 0;
@@ -213,7 +213,7 @@ function stripUndefined(value) {
 				hasKeys = true;
 			}
 		}
-		if (!hasKeys) return void 0;
+		if (!hasKeys) return;
 		return result;
 	}
 	return value;

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "metascope",
-  "version": "0.7.0",
+  "version": "0.7.2",
   "description": "A CLI tool and TypeScript library to easily extract metadata from all kinds of software repositories.",
   "keywords": [
     "metadata",
@@ -49,9 +49,9 @@
     "@types/node": "~22.17.2",
     "@types/plist": "^3.0.5",
     "@types/yargs": "^17.0.35",
-    "case-police": "^2.2.0",
+    "case-police": "^2.2.1",
     "defu": "^6.1.7",
-    "fast-xml-parser": "^5.5.11",
+    "fast-xml-parser": "^5.7.2",
     "find-workspaces": "^0.3.1",
     "git-url-parse": "^16.1.0",
     "gray-matter-es": "^0.2.1",
@@ -60,45 +60,45 @@
     "octokit": "^5.0.5",
     "package-json": "^10.0.1",
     "picomatch": "^4.0.4",
-    "pkg-types": "^2.3.0",
-    "plist": "^3.1.0",
+    "pkg-types": "^2.3.1",
+    "plist": "^3.1.1",
     "pretty-ms": "^9.3.0",
     "read-pkg": "^10.1.0",
     "read-pyproject": "^0.3.2",
     "remark-parse": "^11.0.0",
     "semver": "^7.7.4",
-    "simple-git": "^3.35.2",
+    "simple-git": "^3.36.0",
     "smol-toml": "^1.6.1",
     "spdx-license-list": "^6.11.0",
     "string-ts": "^2.3.1",
-    "tinyexec": "^1.1.1",
+    "tinyexec": "^1.1.2",
     "tinyglobby": "^0.2.16",
     "unified": "^11.0.5",
-    "updates": "^17.14.0",
+    "updates": "^17.16.7",
     "web-tree-sitter": "^0.26.8",
     "yaml": "^2.8.3",
     "yargs": "^18.0.0",
-    "zod": "^4.3.6"
+    "zod": "^4.4.1"
   },
   "devDependencies": {
     "@arethetypeswrong/core": "^0.18.2",
     "@fast-csv/parse": "^5.0.5",
-    "@kitschpatrol/shared-config": "^7.3.0",
+    "@kitschpatrol/shared-config": "^7.5.2",
     "@types/jsonld": "^1.5.15",
     "@types/picomatch": "^4.0.3",
     "@types/semver": "^7.7.1",
     "bumpp": "^11.0.1",
     "jsonld": "^9.0.0",
-    "mdat-plugin-cli-help": "^2.1.2",
+    "mdat-plugin-cli-help": "^3.0.0",
     "msw": "2.12.14",
     "publint": "^0.3.18",
     "shx": "^0.4.0",
     "tree-sitter-python": "^0.25.0",
     "tree-sitter-ruby": "^0.23.1",
-    "tsdown": "^0.21.7",
+    "tsdown": "^0.21.10",
     "tsx": "^4.21.0",
-    "typescript": "~6.0.2",
-    "vitest": "^4.1.4"
+    "typescript": "~6.0.3",
+    "vitest": "^4.1.5"
   },
   "engines": {
     "node": ">=22.17.0"

package/readme.md

@@ -583,7 +583,7 @@ The architecture and non-boilerplate parts of the documentation were human-drive
 
 ## Maintainers
 
-[@kitschpatrol](https://github.com/kitschpatrol)
+[kitschpatrol](https://github.com/kitschpatrol)
 
 ## Acknowledgments
 
@@ -595,7 +595,11 @@ Jacob Peddicord's [askalono](https://github.com/jpeddicord/askalono) project ins
 
 ## Contributing
 
-[Issues](https://github.com/kitschpatrol/metascope/issues) and pull requests are welcome.
+[Issues](https://github.com/kitschpatrol/metascope/issues) are welcome and appreciated.
+
+Please open an issue to discuss changes before submitting a pull request. Unsolicited PRs (especially AI-generated ones) are unlikely to be merged.
+
+This repository uses [@kitschpatrol/shared-config](https://github.com/kitschpatrol/shared-config) (via its `ksc` CLI) for linting and formatting, plus [MDAT](https://github.com/kitschpatrol/mdat) for readme placeholder expansion.
 
 <!-- /contributing -->
 
@@ -603,6 +607,6 @@ Jacob Peddicord's [askalono](https://github.com/jpeddicord/askalono) project ins
 
 ## License
 
-[MIT](license.txt) © Eric Mika
+[MIT](license.txt) © [Eric Mika](https://ericmika.com)
 
 <!-- /license -->

package/dist/bin/chunk-BjEoQXZ0.js

@@ -1 +0,0 @@
-import{createRequire as e}from"node:module";var t=Object.create,n=Object.defineProperty,r=Object.getOwnPropertyDescriptor,i=Object.getOwnPropertyNames,a=Object.getPrototypeOf,o=Object.prototype.hasOwnProperty,s=(e,t)=>()=>(e&&(t=e(e=0)),t),c=(e,t)=>()=>(t||e((t={exports:{}}).exports,t),t.exports),l=(e,t)=>{let r={};for(var i in e)n(r,i,{get:e[i],enumerable:!0});return t||n(r,Symbol.toStringTag,{value:`Module`}),r},u=(e,t,a,s)=>{if(t&&typeof t==`object`||typeof t==`function`)for(var c=i(t),l=0,u=c.length,d;l<u;l++)d=c[l],!o.call(e,d)&&d!==a&&n(e,d,{get:(e=>t[e]).bind(null,d),enumerable:!(s=r(t,d))||s.enumerable});return e},d=(e,r,i)=>(i=e==null?{}:t(a(e)),u(r||!e||!e.__esModule?n(i,`default`,{value:e,enumerable:!0}):i,e)),f=e=>o.call(e,`module.exports`)?e[`module.exports`]:u(n({},`__esModule`,{value:!0}),e),p=e(import.meta.url);export{f as a,p as i,s as n,d as o,l as r,c as t};