@gscdump/cli 2.1.0 → 2.2.0

package/dist/commands/entities.mjs

@@ -27,250 +27,254 @@ async function readUrlList(opts) {
 	for await (const chunk of process.stdin) chunks.push(chunk);
 	return Buffer.concat(chunks).toString("utf8").split("\n").map((l) => l.trim()).filter(Boolean);
 }
-const entitiesCommand = defineCommand({
+const inspectSubCommand = defineCommand({
 	meta: {
-		name: "entities",
-		description: "Manage local entity snapshots (URL inspections and indexing metadata)"
+		name: "inspect",
+		description: "Run URL Inspection for a list of URLs and persist results to the local entity store"
 	},
-	subCommands: {
-		inspect: defineCommand({
-			meta: {
-				name: "inspect",
-				description: "Run URL Inspection for a list of URLs and persist results to the local entity store"
-			},
-			args: {
-				site: {
-					type: "string",
-					alias: "s",
-					description: "Site URL (e.g., sc-domain:example.com); defaults to config.defaultSite or prompt"
-				},
-				file: {
-					type: "string",
-					alias: "f",
-					description: "Path to a file with one URL per line. If omitted, reads from stdin."
-				},
-				limit: {
-					type: "string",
-					description: `Max URLs to inspect this run (default: ${INSPECTION_QPD_PER_PROPERTY}, the per-property GSC daily quota)`
-				},
-				concurrency: {
-					type: "string",
-					alias: "c",
-					default: "4",
-					description: "Concurrent in-flight inspect calls (default: 4)"
-				},
-				...OUTPUT_ARGS
-			},
-			async run({ args }) {
-				const { json, quiet } = applyOutputMode(args);
-				const ctx = await createCommandContext({
-					needsAuth: true,
-					needsStore: true
+	args: {
+		site: {
+			type: "string",
+			alias: "s",
+			description: "Site URL (e.g., sc-domain:example.com); defaults to config.defaultSite or prompt"
+		},
+		file: {
+			type: "string",
+			alias: "f",
+			description: "Path to a file with one URL per line. If omitted, reads from stdin."
+		},
+		limit: {
+			type: "string",
+			description: `Max URLs to inspect this run (default: ${INSPECTION_QPD_PER_PROPERTY}, the per-property GSC daily quota)`
+		},
+		concurrency: {
+			type: "string",
+			alias: "c",
+			default: "4",
+			description: "Concurrent in-flight inspect calls (default: 4)"
+		},
+		...OUTPUT_ARGS
+	},
+	async run({ args }) {
+		const { json, quiet } = applyOutputMode(args);
+		const ctx = await createCommandContext({
+			needsAuth: true,
+			needsStore: true
+		});
+		const client = ctx.client;
+		const store = ctx.store;
+		const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
+		const limit = args.limit ? Number.parseInt(String(args.limit), 10) : INSPECTION_QPD_PER_PROPERTY;
+		const concurrency = Math.max(1, Number.parseInt(String(args.concurrency), 10) || 4);
+		const urls = (await readUrlList({ file: args.file ? String(args.file) : void 0 })).slice(0, limit);
+		if (urls.length === 0) {
+			logger.warn("No URLs to inspect.");
+			return;
+		}
+		if (urls.length === limit && limit < INSPECTION_QPD_PER_PROPERTY) logger.info(`Capping at --limit ${limit}`);
+		if (urls.length === INSPECTION_QPD_PER_PROPERTY) logger.info(`Hit per-property daily inspection quota (${INSPECTION_QPD_PER_PROPERTY}); remaining URLs will be queued for tomorrow.`);
+		const inspector = createInspectionStore({ dataSource: store.dataSource });
+		let completed = 0;
+		let failed = 0;
+		const records = [];
+		const failures = [];
+		await runWithConcurrency(urls, concurrency, async (url) => {
+			const result = await client.inspect(siteUrl, url).catch((err) => err);
+			if (result instanceof Error) {
+				failed++;
+				failures.push({
+					url,
+					error: result.message
 				});
-				const client = ctx.client;
-				const store = ctx.store;
-				const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
-				const limit = args.limit ? Number.parseInt(String(args.limit), 10) : INSPECTION_QPD_PER_PROPERTY;
-				const concurrency = Math.max(1, Number.parseInt(String(args.concurrency), 10) || 4);
-				const urls = (await readUrlList({ file: args.file ? String(args.file) : void 0 })).slice(0, limit);
-				if (urls.length === 0) {
-					logger.warn("No URLs to inspect.");
-					return;
-				}
-				if (urls.length === limit && limit < INSPECTION_QPD_PER_PROPERTY) logger.info(`Capping at --limit ${limit}`);
-				if (urls.length === INSPECTION_QPD_PER_PROPERTY) logger.info(`Hit per-property daily inspection quota (${INSPECTION_QPD_PER_PROPERTY}); remaining URLs will be queued for tomorrow.`);
-				const inspector = createInspectionStore({ dataSource: store.dataSource });
-				let completed = 0;
-				let failed = 0;
-				const records = [];
-				const failures = [];
-				await runWithConcurrency(urls, concurrency, async (url) => {
-					const result = await client.inspect(siteUrl, url).catch((err) => err);
-					if (result instanceof Error) {
-						failed++;
-						failures.push({
-							url,
-							error: result.message
-						});
-					} else {
-						const ix = result.inspectionResult;
-						const indexStatus = ix?.indexStatusResult;
-						records.push({
-							url,
-							inspectedAt: (/* @__PURE__ */ new Date()).toISOString(),
-							indexStatus: indexStatus?.verdict ?? void 0,
-							lastCrawlTime: indexStatus?.lastCrawlTime ?? void 0,
-							googleCanonical: indexStatus?.googleCanonical ?? void 0,
-							userCanonical: indexStatus?.userCanonical ?? void 0,
-							coverageState: indexStatus?.coverageState ?? void 0,
-							robotsTxtState: indexStatus?.robotsTxtState ?? void 0,
-							indexingState: indexStatus?.indexingState ?? void 0,
-							pageFetchState: indexStatus?.pageFetchState ?? void 0,
-							mobileUsabilityVerdict: ix?.mobileUsabilityResult?.verdict ?? void 0,
-							richResultsVerdict: ix?.richResultsResult?.verdict ?? void 0,
-							raw: ix
-						});
-					}
-					completed++;
-					if (!quiet) process.stdout.write(`\r${progressBar(completed, urls.length, `${url.slice(0, 60)}`)}`);
+			} else {
+				const ix = result.inspectionResult;
+				const indexStatus = ix?.indexStatusResult;
+				records.push({
+					url,
+					inspectedAt: (/* @__PURE__ */ new Date()).toISOString(),
+					indexStatus: indexStatus?.verdict ?? void 0,
+					lastCrawlTime: indexStatus?.lastCrawlTime ?? void 0,
+					googleCanonical: indexStatus?.googleCanonical ?? void 0,
+					userCanonical: indexStatus?.userCanonical ?? void 0,
+					coverageState: indexStatus?.coverageState ?? void 0,
+					robotsTxtState: indexStatus?.robotsTxtState ?? void 0,
+					indexingState: indexStatus?.indexingState ?? void 0,
+					pageFetchState: indexStatus?.pageFetchState ?? void 0,
+					mobileUsabilityVerdict: ix?.mobileUsabilityResult?.verdict ?? void 0,
+					richResultsVerdict: ix?.richResultsResult?.verdict ?? void 0,
+					raw: ix
 				});
-				if (!quiet) process.stdout.write("\n");
-				await inspector.appendHistory({
-					userId: store.userId,
-					siteId: store.siteIdFor(siteUrl)
-				}, records);
-				if (json) console.log(JSON.stringify({
-					site: siteUrl,
-					inspected: records.length,
-					failed,
-					failures,
-					records
-				}, null, 2));
-				else if (!quiet) {
-					logger.success(`Inspected ${records.length}/${urls.length} URL(s)`);
-					if (failed > 0) {
-						logger.warn(`${failed} failed:`);
-						for (const f of failures.slice(0, 5)) console.log(`  ${f.url}: ${f.error}`);
-						if (failures.length > 5) console.log(`  ... and ${failures.length - 5} more`);
-					}
-				}
-				if (failed > 0) process.exit(1);
 			}
-		}),
-		show: defineCommand({
-			meta: {
-				name: "show",
-				description: "Print the latest inspection record for a URL from the local entity store"
-			},
-			args: {
-				...OUTPUT_ARGS,
-				site: {
-					type: "string",
-					alias: "s",
-					description: "Site URL (defaults to config.defaultSite or prompt)"
-				},
-				url: {
-					type: "positional",
-					required: true,
-					description: "URL to look up"
-				}
-			},
-			async run({ args }) {
-				const { json } = applyOutputMode(args);
-				const ctx = await createCommandContext({
-					needsAuth: true,
-					needsStore: true
+			completed++;
+			if (!quiet) process.stdout.write(`\r${progressBar(completed, urls.length, `${url.slice(0, 60)}`)}`);
+		});
+		if (!quiet) process.stdout.write("\n");
+		await inspector.appendHistory({
+			userId: store.userId,
+			siteId: store.siteIdFor(siteUrl)
+		}, records);
+		if (json) console.log(JSON.stringify({
+			site: siteUrl,
+			inspected: records.length,
+			failed,
+			failures,
+			records
+		}, null, 2));
+		else if (!quiet) {
+			logger.success(`Inspected ${records.length}/${urls.length} URL(s)`);
+			if (failed > 0) {
+				logger.warn(`${failed} failed:`);
+				for (const f of failures.slice(0, 5)) console.log(`  ${f.url}: ${f.error}`);
+				if (failures.length > 5) console.log(`  ... and ${failures.length - 5} more`);
+			}
+		}
+		if (failed > 0) process.exit(1);
+	}
+});
+const showSubCommand = defineCommand({
+	meta: {
+		name: "show",
+		description: "Print the latest inspection record for a URL from the local entity store"
+	},
+	args: {
+		...OUTPUT_ARGS,
+		site: {
+			type: "string",
+			alias: "s",
+			description: "Site URL (defaults to config.defaultSite or prompt)"
+		},
+		url: {
+			type: "positional",
+			required: true,
+			description: "URL to look up"
+		}
+	},
+	async run({ args }) {
+		const { json } = applyOutputMode(args);
+		const ctx = await createCommandContext({
+			needsAuth: true,
+			needsStore: true
+		});
+		const store = ctx.store;
+		const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
+		const record = await findLatestInspection(createInspectionStore({ dataSource: store.dataSource }), {
+			userId: store.userId,
+			siteId: store.siteIdFor(siteUrl)
+		}, String(args.url));
+		if (!record) {
+			logger.warn(`No inspection record for ${args.url}`);
+			process.exit(1);
+		}
+		if (json) {
+			console.log(JSON.stringify(record, null, 2));
+			return;
+		}
+		console.log();
+		console.log(`  \x1B[1m${record.url}\x1B[0m`);
+		console.log(`  Inspected:    ${record.inspectedAt}`);
+		if (record.indexStatus) console.log(`  Index:        ${record.indexStatus}`);
+		if (record.lastCrawlTime) console.log(`  Last crawl:   ${record.lastCrawlTime}`);
+		if (record.googleCanonical) console.log(`  Canonical:    ${record.googleCanonical}`);
+		if (record.coverageState) console.log(`  Coverage:     ${record.coverageState}`);
+		if (record.richResultsVerdict) console.log(`  Rich results: ${record.richResultsVerdict}`);
+		console.log();
+	}
+});
+const indexingSnapshotSubCommand = defineCommand({
+	meta: {
+		name: "snapshot",
+		description: "Fetch Indexing API metadata (latest update/remove per URL) and persist to the local entity store"
+	},
+	args: {
+		site: {
+			type: "string",
+			alias: "s",
+			description: "Site URL (e.g., sc-domain:example.com); defaults to config.defaultSite or prompt"
+		},
+		file: {
+			type: "string",
+			alias: "f",
+			description: "Path to a file with one URL per line. If omitted, reads from stdin."
+		},
+		concurrency: {
+			type: "string",
+			alias: "c",
+			default: "4",
+			description: "Concurrent in-flight getMetadata calls (default: 4)"
+		},
+		...OUTPUT_ARGS
+	},
+	async run({ args }) {
+		const { quiet } = applyOutputMode(args);
+		const ctx = await createCommandContext({
+			needsAuth: true,
+			needsStore: true
+		});
+		const client = ctx.client;
+		const store = ctx.store;
+		const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
+		const concurrency = Math.max(1, Number.parseInt(String(args.concurrency), 10) || 4);
+		const urls = await readUrlList({ file: args.file ? String(args.file) : void 0 });
+		if (urls.length === 0) {
+			logger.warn("No URLs to fetch metadata for.");
+			return;
+		}
+		const records = [];
+		const failures = [];
+		let completed = 0;
+		await runWithConcurrency(urls, concurrency, async (url) => {
+			const result = await client.indexing.getMetadata(url).catch((err) => err);
+			if (result instanceof Error) {
+				if (INDEXING_NOT_FOUND_RE.test(result.message)) records.push({
+					url,
+					capturedAt: (/* @__PURE__ */ new Date()).toISOString()
 				});
-				const store = ctx.store;
-				const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
-				const record = await findLatestInspection(createInspectionStore({ dataSource: store.dataSource }), {
-					userId: store.userId,
-					siteId: store.siteIdFor(siteUrl)
-				}, String(args.url));
-				if (!record) {
-					logger.warn(`No inspection record for ${args.url}`);
-					process.exit(1);
-				}
-				if (json) {
-					console.log(JSON.stringify(record, null, 2));
-					return;
-				}
-				console.log();
-				console.log(`  \x1B[1m${record.url}\x1B[0m`);
-				console.log(`  Inspected:    ${record.inspectedAt}`);
-				if (record.indexStatus) console.log(`  Index:        ${record.indexStatus}`);
-				if (record.lastCrawlTime) console.log(`  Last crawl:   ${record.lastCrawlTime}`);
-				if (record.googleCanonical) console.log(`  Canonical:    ${record.googleCanonical}`);
-				if (record.coverageState) console.log(`  Coverage:     ${record.coverageState}`);
-				if (record.mobileUsabilityVerdict) console.log(`  Mobile:       ${record.mobileUsabilityVerdict}`);
-				if (record.richResultsVerdict) console.log(`  Rich results: ${record.richResultsVerdict}`);
-				console.log();
+				else failures.push({
+					url,
+					error: result.message
+				});
+			} else records.push({
+				url,
+				capturedAt: (/* @__PURE__ */ new Date()).toISOString(),
+				latestUpdateAt: result.latestUpdate?.notifyTime ?? void 0,
+				latestRemoveAt: result.latestRemove?.notifyTime ?? void 0,
+				raw: result
+			});
+			completed++;
+			if (!quiet) process.stdout.write(`\r${progressBar(completed, urls.length, url.slice(0, 60))}`);
+		});
+		if (!quiet) process.stdout.write("\n");
+		await createIndexingMetadataStore({ dataSource: store.dataSource }).writeBatch({
+			userId: store.userId,
+			siteId: store.siteIdFor(siteUrl)
+		}, records);
+		if (!quiet) {
+			logger.success(`Captured metadata for ${records.length}/${urls.length} URL(s)`);
+			if (failures.length > 0) {
+				logger.warn(`${failures.length} failed:`);
+				for (const f of failures.slice(0, 5)) console.log(`  ${f.url}: ${f.error}`);
+				if (failures.length > 5) console.log(`  ... and ${failures.length - 5} more`);
 			}
-		}),
-		indexing: defineCommand({
-			meta: {
-				name: "indexing",
-				description: "Snapshot Indexing API metadata per URL"
-			},
-			subCommands: { snapshot: defineCommand({
-				meta: {
-					name: "snapshot",
-					description: "Fetch Indexing API metadata (latest update/remove per URL) and persist to the local entity store"
-				},
-				args: {
-					site: {
-						type: "string",
-						alias: "s",
-						description: "Site URL (e.g., sc-domain:example.com); defaults to config.defaultSite or prompt"
-					},
-					file: {
-						type: "string",
-						alias: "f",
-						description: "Path to a file with one URL per line. If omitted, reads from stdin."
-					},
-					concurrency: {
-						type: "string",
-						alias: "c",
-						default: "4",
-						description: "Concurrent in-flight getMetadata calls (default: 4)"
-					},
-					...OUTPUT_ARGS
-				},
-				async run({ args }) {
-					const { quiet } = applyOutputMode(args);
-					const ctx = await createCommandContext({
-						needsAuth: true,
-						needsStore: true
-					});
-					const client = ctx.client;
-					const store = ctx.store;
-					const siteUrl = await ctx.resolveSite(args.site ? String(args.site) : void 0);
-					const concurrency = Math.max(1, Number.parseInt(String(args.concurrency), 10) || 4);
-					const urls = await readUrlList({ file: args.file ? String(args.file) : void 0 });
-					if (urls.length === 0) {
-						logger.warn("No URLs to fetch metadata for.");
-						return;
-					}
-					const records = [];
-					const failures = [];
-					let completed = 0;
-					await runWithConcurrency(urls, concurrency, async (url) => {
-						const result = await client.indexing.getMetadata(url).catch((err) => err);
-						if (result instanceof Error) if (INDEXING_NOT_FOUND_RE.test(result.message)) records.push({
-							url,
-							capturedAt: (/* @__PURE__ */ new Date()).toISOString()
-						});
-						else failures.push({
-							url,
-							error: result.message
-						});
-						else records.push({
-							url,
-							capturedAt: (/* @__PURE__ */ new Date()).toISOString(),
-							latestUpdateAt: result.latestUpdate?.notifyTime ?? void 0,
-							latestRemoveAt: result.latestRemove?.notifyTime ?? void 0,
-							raw: result
-						});
-						completed++;
-						if (!quiet) process.stdout.write(`\r${progressBar(completed, urls.length, url.slice(0, 60))}`);
-					});
-					if (!quiet) process.stdout.write("\n");
-					await createIndexingMetadataStore({ dataSource: store.dataSource }).writeBatch({
-						userId: store.userId,
-						siteId: store.siteIdFor(siteUrl)
-					}, records);
-					if (!quiet) {
-						logger.success(`Captured metadata for ${records.length}/${urls.length} URL(s)`);
-						if (failures.length > 0) {
-							logger.warn(`${failures.length} failed:`);
-							for (const f of failures.slice(0, 5)) console.log(`  ${f.url}: ${f.error}`);
-							if (failures.length > 5) console.log(`  ... and ${failures.length - 5} more`);
-						}
-					}
-					if (failures.length > 0) process.exit(1);
-				}
-			}) }
-		})
+		}
+		if (failures.length > 0) process.exit(1);
+	}
+});
+const indexingSubCommand = defineCommand({
+	meta: {
+		name: "indexing",
+		description: "Snapshot Indexing API metadata per URL"
+	},
+	subCommands: { snapshot: indexingSnapshotSubCommand }
+});
+const entitiesCommand = defineCommand({
+	meta: {
+		name: "entities",
+		description: "Manage local entity snapshots (URL inspections and indexing metadata)"
+	},
+	subCommands: {
+		inspect: inspectSubCommand,
+		show: showSubCommand,
+		indexing: indexingSubCommand
 	}
 });
 export { entitiesCommand };