@rivi-software/agentic-seo-check 0.0.0-stage → 0.1.0
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/LICENSE +21 -0
- package/README.md +160 -2
- package/dist/cli.d.ts +2 -0
- package/dist/cli.js +149 -0
- package/dist/cli.js.map +1 -0
- package/dist/core/audit.d.ts +72 -0
- package/dist/core/audit.js +337 -0
- package/dist/core/audit.js.map +1 -0
- package/dist/core/concurrency.d.ts +4 -0
- package/dist/core/concurrency.js +29 -0
- package/dist/core/concurrency.js.map +1 -0
- package/dist/core/crawl.d.ts +40 -0
- package/dist/core/crawl.js +195 -0
- package/dist/core/crawl.js.map +1 -0
- package/dist/core/fetch.d.ts +27 -0
- package/dist/core/fetch.js +135 -0
- package/dist/core/fetch.js.map +1 -0
- package/dist/core/html/extract.d.ts +10 -0
- package/dist/core/html/extract.js +150 -0
- package/dist/core/html/extract.js.map +1 -0
- package/dist/core/html/structured-data.d.ts +11 -0
- package/dist/core/html/structured-data.js +185 -0
- package/dist/core/html/structured-data.js.map +1 -0
- package/dist/core/html/text.d.ts +23 -0
- package/dist/core/html/text.js +149 -0
- package/dist/core/html/text.js.map +1 -0
- package/dist/core/lighthouse.d.ts +49 -0
- package/dist/core/lighthouse.js +140 -0
- package/dist/core/lighthouse.js.map +1 -0
- package/dist/core/performance-rules.d.ts +3 -0
- package/dist/core/performance-rules.js +53 -0
- package/dist/core/performance-rules.js.map +1 -0
- package/dist/core/render-rules.d.ts +7 -0
- package/dist/core/render-rules.js +77 -0
- package/dist/core/render-rules.js.map +1 -0
- package/dist/core/render.d.ts +22 -0
- package/dist/core/render.js +73 -0
- package/dist/core/render.js.map +1 -0
- package/dist/core/report/markdown.d.ts +2 -0
- package/dist/core/report/markdown.js +152 -0
- package/dist/core/report/markdown.js.map +1 -0
- package/dist/core/robots.d.ts +11 -0
- package/dist/core/robots.js +33 -0
- package/dist/core/robots.js.map +1 -0
- package/dist/core/rules/canonical.d.ts +2 -0
- package/dist/core/rules/canonical.js +105 -0
- package/dist/core/rules/canonical.js.map +1 -0
- package/dist/core/rules/content.d.ts +1 -0
- package/dist/core/rules/content.js +41 -0
- package/dist/core/rules/content.js.map +1 -0
- package/dist/core/rules/document.d.ts +1 -0
- package/dist/core/rules/document.js +60 -0
- package/dist/core/rules/document.js.map +1 -0
- package/dist/core/rules/headings.d.ts +1 -0
- package/dist/core/rules/headings.js +61 -0
- package/dist/core/rules/headings.js.map +1 -0
- package/dist/core/rules/hreflang.d.ts +1 -0
- package/dist/core/rules/hreflang.js +68 -0
- package/dist/core/rules/hreflang.js.map +1 -0
- package/dist/core/rules/images.d.ts +1 -0
- package/dist/core/rules/images.js +59 -0
- package/dist/core/rules/images.js.map +1 -0
- package/dist/core/rules/index.d.ts +14 -0
- package/dist/core/rules/index.js +76 -0
- package/dist/core/rules/index.js.map +1 -0
- package/dist/core/rules/indexing.d.ts +10 -0
- package/dist/core/rules/indexing.js +53 -0
- package/dist/core/rules/indexing.js.map +1 -0
- package/dist/core/rules/keyword.d.ts +1 -0
- package/dist/core/rules/keyword.js +74 -0
- package/dist/core/rules/keyword.js.map +1 -0
- package/dist/core/rules/links.d.ts +1 -0
- package/dist/core/rules/links.js +72 -0
- package/dist/core/rules/links.js.map +1 -0
- package/dist/core/rules/meta.d.ts +1 -0
- package/dist/core/rules/meta.js +60 -0
- package/dist/core/rules/meta.js.map +1 -0
- package/dist/core/rules/response.d.ts +1 -0
- package/dist/core/rules/response.js +59 -0
- package/dist/core/rules/response.js.map +1 -0
- package/dist/core/rules/rule.d.ts +42 -0
- package/dist/core/rules/rule.js +51 -0
- package/dist/core/rules/rule.js.map +1 -0
- package/dist/core/rules/social.d.ts +1 -0
- package/dist/core/rules/social.js +74 -0
- package/dist/core/rules/social.js.map +1 -0
- package/dist/core/rules/structured-data.d.ts +23 -0
- package/dist/core/rules/structured-data.js +169 -0
- package/dist/core/rules/structured-data.js.map +1 -0
- package/dist/core/rules/title.d.ts +1 -0
- package/dist/core/rules/title.js +59 -0
- package/dist/core/rules/title.js.map +1 -0
- package/dist/core/score.d.ts +3 -0
- package/dist/core/score.js +15 -0
- package/dist/core/score.js.map +1 -0
- package/dist/core/site-rules.d.ts +64 -0
- package/dist/core/site-rules.js +251 -0
- package/dist/core/site-rules.js.map +1 -0
- package/dist/core/sitemap.d.ts +13 -0
- package/dist/core/sitemap.js +95 -0
- package/dist/core/sitemap.js.map +1 -0
- package/dist/core/static.d.ts +9 -0
- package/dist/core/static.js +184 -0
- package/dist/core/static.js.map +1 -0
- package/dist/core/summary.d.ts +2 -0
- package/dist/core/summary.js +28 -0
- package/dist/core/summary.js.map +1 -0
- package/dist/core/types.d.ts +296 -0
- package/dist/core/types.js +26 -0
- package/dist/core/types.js.map +1 -0
- package/dist/core/url.d.ts +20 -0
- package/dist/core/url.js +139 -0
- package/dist/core/url.js.map +1 -0
- package/dist/index.d.ts +11 -0
- package/dist/index.js +10 -0
- package/dist/index.js.map +1 -0
- package/dist/mcp-bin.d.ts +2 -0
- package/dist/mcp-bin.js +7 -0
- package/dist/mcp-bin.js.map +1 -0
- package/dist/mcp.d.ts +5 -0
- package/dist/mcp.js +190 -0
- package/dist/mcp.js.map +1 -0
- package/dist/version.d.ts +2 -0
- package/dist/version.js +4 -0
- package/dist/version.js.map +1 -0
- package/package.json +89 -3
- package/skills/seo-audit/SKILL.md +53 -0
|
@@ -0,0 +1,337 @@
|
|
|
1
|
+
import { promises as fs } from 'node:fs';
|
|
2
|
+
import path from 'node:path';
|
|
3
|
+
import { crawlSite } from './crawl.js';
|
|
4
|
+
import { DEFAULT_USER_AGENT, fetchPage, toResponseInfo } from './fetch.js';
|
|
5
|
+
import { extractPage } from './html/extract.js';
|
|
6
|
+
import { runLighthouse, LighthouseUnavailableError } from './lighthouse.js';
|
|
7
|
+
import { performanceFindings } from './performance-rules.js';
|
|
8
|
+
import { compareRender } from './render-rules.js';
|
|
9
|
+
import { renderPage, RenderUnavailableError } from './render.js';
|
|
10
|
+
import { runPageRules, sortFindings } from './rules/index.js';
|
|
11
|
+
import { pageScore, siteScore } from './score.js';
|
|
12
|
+
import { runSiteRules } from './site-rules.js';
|
|
13
|
+
import { scanDirectory } from './static.js';
|
|
14
|
+
import { buildSummary } from './summary.js';
|
|
15
|
+
import { countFindings, DEFAULT_THRESHOLDS, } from './types.js';
|
|
16
|
+
import { isHttpUrl } from './url.js';
|
|
17
|
+
export const DEFAULT_CRAWL = { maxPages: 50, maxDepth: 3, concurrency: 5, delayMs: 0, maxLinkChecks: 200 };
|
|
18
|
+
/** Audit an HTML string. `url` is used to resolve links and for display. */
|
|
19
|
+
export async function auditHtml(html, input = {}, options = {}) {
|
|
20
|
+
const url = input.url ?? options.baseUrl ?? 'http://localhost/';
|
|
21
|
+
const page = extractPage(html, { url, includeSubdomains: options.includeSubdomains });
|
|
22
|
+
const source = input.source ?? 'html';
|
|
23
|
+
const thresholds = { ...DEFAULT_THRESHOLDS, ...(options.thresholds ?? {}) };
|
|
24
|
+
let findings = runPageRules({ page, source, response: input.response, keyword: options.keyword, thresholds, depth: input.depth }, { disable: new Set(options.disableRules ?? []) });
|
|
25
|
+
if (options.includeInfo === false)
|
|
26
|
+
findings = findings.filter((f) => f.severity !== 'info');
|
|
27
|
+
const audit = {
|
|
28
|
+
schemaVersion: 1,
|
|
29
|
+
kind: 'page',
|
|
30
|
+
target: input.target ?? url,
|
|
31
|
+
url: input.response?.finalUrl ?? url,
|
|
32
|
+
source,
|
|
33
|
+
auditedAt: new Date().toISOString(),
|
|
34
|
+
summary: buildSummary(page, input.response),
|
|
35
|
+
findings,
|
|
36
|
+
counts: countFindings(findings),
|
|
37
|
+
score: 0,
|
|
38
|
+
};
|
|
39
|
+
if (input.response)
|
|
40
|
+
audit.response = input.response;
|
|
41
|
+
if (input.depth !== undefined)
|
|
42
|
+
audit.depth = input.depth;
|
|
43
|
+
audit.score = pageScore(audit.counts);
|
|
44
|
+
return audit;
|
|
45
|
+
}
|
|
46
|
+
/** Fetch and audit one URL. Optionally render and/or run Lighthouse. */
|
|
47
|
+
export async function auditUrl(url, options = {}) {
|
|
48
|
+
const res = await fetchPage(url, options.fetch ?? {});
|
|
49
|
+
const response = toResponseInfo(res);
|
|
50
|
+
const audit = await auditHtml(res.body ?? '', { url: res.finalUrl, response, source: 'url', target: url }, options);
|
|
51
|
+
if (res.error) {
|
|
52
|
+
audit.findings.unshift({
|
|
53
|
+
rule: 'response.fetch-failed',
|
|
54
|
+
severity: 'error',
|
|
55
|
+
category: 'response',
|
|
56
|
+
message: `Could not fetch ${url}: ${res.error}`,
|
|
57
|
+
recommendation: 'Check that the URL is reachable from this machine (DNS, TLS, firewall, dev server running) and that it responds within the timeout.',
|
|
58
|
+
url,
|
|
59
|
+
evidence: { error: res.error, redirectChain: res.redirectChain },
|
|
60
|
+
});
|
|
61
|
+
audit.counts = countFindings(audit.findings);
|
|
62
|
+
audit.score = pageScore(audit.counts);
|
|
63
|
+
return audit;
|
|
64
|
+
}
|
|
65
|
+
await enrichPage(audit, options);
|
|
66
|
+
return audit;
|
|
67
|
+
}
|
|
68
|
+
/** Audit a single .html file from disk. */
|
|
69
|
+
export async function auditFile(filePath, options = {}) {
|
|
70
|
+
const html = await fs.readFile(filePath, 'utf8');
|
|
71
|
+
const base = options.baseUrl ?? 'http://localhost/';
|
|
72
|
+
const url = new URL(`/${path.basename(filePath)}`, base.endsWith('/') ? base : `${base}/`).href;
|
|
73
|
+
return auditHtml(html, { url, source: 'file', target: filePath }, options);
|
|
74
|
+
}
|
|
75
|
+
/** Crawl a site and audit every page plus site-level checks. */
|
|
76
|
+
export async function auditSite(startUrl, options = {}) {
|
|
77
|
+
const fetchOpts = { ...(options.fetch ?? {}) };
|
|
78
|
+
fetchOpts.userAgent ??= DEFAULT_USER_AGENT;
|
|
79
|
+
const internal = {
|
|
80
|
+
maxPages: options.maxPages ?? DEFAULT_CRAWL.maxPages,
|
|
81
|
+
maxDepth: options.maxDepth ?? DEFAULT_CRAWL.maxDepth,
|
|
82
|
+
concurrency: options.concurrency ?? DEFAULT_CRAWL.concurrency,
|
|
83
|
+
delayMs: options.delayMs ?? DEFAULT_CRAWL.delayMs,
|
|
84
|
+
respectRobots: options.respectRobots ?? true,
|
|
85
|
+
includeSubdomains: options.includeSubdomains ?? false,
|
|
86
|
+
checkExternalLinks: options.checkExternalLinks ?? false,
|
|
87
|
+
maxLinkChecks: options.maxLinkChecks ?? DEFAULT_CRAWL.maxLinkChecks,
|
|
88
|
+
includePatterns: (options.includePatterns ?? []).map((p) => new RegExp(p)),
|
|
89
|
+
excludePatterns: (options.excludePatterns ?? []).map((p) => new RegExp(p)),
|
|
90
|
+
fetch: fetchOpts,
|
|
91
|
+
onProgress: options.onProgress,
|
|
92
|
+
};
|
|
93
|
+
const result = await crawlSite(startUrl, internal);
|
|
94
|
+
const site = await buildSiteAudit(result, 'crawl', startUrl, options);
|
|
95
|
+
// Render/Lighthouse are expensive: apply them to the start page only.
|
|
96
|
+
const startPage = site.pages.find((p) => p.depth === 0) ?? site.pages[0];
|
|
97
|
+
if (startPage && (options.render || options.lighthouse)) {
|
|
98
|
+
await enrichPage(startPage, options);
|
|
99
|
+
recomputeSite(site);
|
|
100
|
+
}
|
|
101
|
+
return shapeSite(site, options);
|
|
102
|
+
}
|
|
103
|
+
/** Audit a directory of built HTML files (dist/, out/, build/). */
|
|
104
|
+
export async function auditDirectory(dir, options = {}) {
|
|
105
|
+
const result = await scanDirectory(dir, {
|
|
106
|
+
baseUrl: options.baseUrl ?? 'http://localhost/',
|
|
107
|
+
includeSubdomains: options.includeSubdomains ?? false,
|
|
108
|
+
maxPages: options.maxPages ?? 500,
|
|
109
|
+
userAgent: options.fetch?.userAgent,
|
|
110
|
+
});
|
|
111
|
+
const site = await buildSiteAudit(result, 'directory', dir, options);
|
|
112
|
+
return shapeSite(site, options);
|
|
113
|
+
}
|
|
114
|
+
/**
|
|
115
|
+
* Auto-detect the target type: http(s) URL, directory, .html file, or "-" for stdin.
|
|
116
|
+
* `crawl` forces a site crawl for URLs.
|
|
117
|
+
*/
|
|
118
|
+
export async function auditTarget(target, options = {}) {
|
|
119
|
+
if (target === '-' || options.stdinHtml !== undefined) {
|
|
120
|
+
const html = options.stdinHtml ?? (await readStdin());
|
|
121
|
+
return auditHtml(html, { url: options.baseUrl, source: 'html', target: 'stdin' }, options);
|
|
122
|
+
}
|
|
123
|
+
if (isHttpUrl(target))
|
|
124
|
+
return options.crawl ? auditSite(target, options) : auditUrl(target, options);
|
|
125
|
+
if (/^localhost(:\d+)?(\/|$)/i.test(target) || /^[\w.-]+\.[a-z]{2,}(\/|$)/i.test(target)) {
|
|
126
|
+
const url = `http${/^localhost/i.test(target) ? '' : 's'}://${target}`;
|
|
127
|
+
return options.crawl ? auditSite(url, options) : auditUrl(url, options);
|
|
128
|
+
}
|
|
129
|
+
const stat = await fs.stat(target).catch(() => null);
|
|
130
|
+
if (!stat)
|
|
131
|
+
throw new Error(`Target "${target}" is not a URL, file or directory.`);
|
|
132
|
+
if (stat.isDirectory())
|
|
133
|
+
return auditDirectory(target, options);
|
|
134
|
+
return auditFile(target, options);
|
|
135
|
+
}
|
|
136
|
+
async function enrichPage(audit, options) {
|
|
137
|
+
const url = audit.response?.finalUrl ?? audit.url;
|
|
138
|
+
const extra = [];
|
|
139
|
+
if (options.render) {
|
|
140
|
+
options.onProgress?.({ phase: 'render', message: `Rendering ${url} with Playwright` });
|
|
141
|
+
try {
|
|
142
|
+
const rendered = await renderPage(url, { timeoutMs: Math.max(options.fetch?.timeoutMs ?? 0, 30_000), userAgent: options.fetch?.userAgent, headers: options.fetch?.headers });
|
|
143
|
+
const rawPage = extractPage(await rawHtmlFor(audit, options), { url, includeSubdomains: options.includeSubdomains });
|
|
144
|
+
const renderedPage = extractPage(rendered.html, { url: rendered.finalUrl, siteHost: new URL(url).hostname, includeSubdomains: options.includeSubdomains });
|
|
145
|
+
const { comparison, findings } = compareRender(rawPage, renderedPage, rendered.finalUrl, rendered.consoleErrors, audit.response);
|
|
146
|
+
audit.rendering = comparison;
|
|
147
|
+
extra.push(...findings);
|
|
148
|
+
}
|
|
149
|
+
catch (err) {
|
|
150
|
+
extra.push(toolFinding('tool.render-unavailable', err instanceof RenderUnavailableError ? err.message : `Rendering failed: ${err.message}`, url));
|
|
151
|
+
}
|
|
152
|
+
}
|
|
153
|
+
if (options.lighthouse) {
|
|
154
|
+
options.onProgress?.({ phase: 'lighthouse', message: `Running Lighthouse (${options.formFactor ?? 'mobile'}) on ${url}` });
|
|
155
|
+
try {
|
|
156
|
+
const perf = await runLighthouse(url, { formFactor: options.formFactor ?? 'mobile', userAgent: options.fetch?.userAgent, headers: options.fetch?.headers });
|
|
157
|
+
audit.performance = perf;
|
|
158
|
+
extra.push(...performanceFindings(perf));
|
|
159
|
+
}
|
|
160
|
+
catch (err) {
|
|
161
|
+
extra.push(toolFinding('tool.lighthouse-unavailable', err instanceof LighthouseUnavailableError ? err.message : `Lighthouse failed: ${err.message}`, url));
|
|
162
|
+
}
|
|
163
|
+
}
|
|
164
|
+
if (extra.length) {
|
|
165
|
+
const filtered = options.includeInfo === false ? extra.filter((f) => f.severity !== 'info') : extra;
|
|
166
|
+
audit.findings = sortFindings([...audit.findings, ...filtered]);
|
|
167
|
+
audit.counts = countFindings(audit.findings);
|
|
168
|
+
audit.score = pageScore(audit.counts);
|
|
169
|
+
}
|
|
170
|
+
}
|
|
171
|
+
// The raw HTML is not kept on the audit (too large); re-fetch it for the render comparison.
|
|
172
|
+
async function rawHtmlFor(audit, options) {
|
|
173
|
+
const res = await fetchPage(audit.response?.finalUrl ?? audit.url, options.fetch ?? {});
|
|
174
|
+
return res.body ?? '';
|
|
175
|
+
}
|
|
176
|
+
function toolFinding(rule, message, url) {
|
|
177
|
+
return {
|
|
178
|
+
rule,
|
|
179
|
+
severity: 'info',
|
|
180
|
+
category: 'tool',
|
|
181
|
+
message,
|
|
182
|
+
recommendation: 'Install the optional dependency (see message) or run without this option. The rest of the audit is complete.',
|
|
183
|
+
url,
|
|
184
|
+
};
|
|
185
|
+
}
|
|
186
|
+
async function buildSiteAudit(result, source, target, options) {
|
|
187
|
+
options.onProgress?.({ phase: 'rules', message: `Evaluating rules for ${result.records.length} pages` });
|
|
188
|
+
const thresholds = { ...DEFAULT_THRESHOLDS, ...(options.thresholds ?? {}) };
|
|
189
|
+
const pages = [];
|
|
190
|
+
for (const record of result.records) {
|
|
191
|
+
const audit = await auditRecord(record, source === 'directory' ? 'file' : 'url', options);
|
|
192
|
+
pages.push(audit);
|
|
193
|
+
}
|
|
194
|
+
let siteFindings = runSiteRules({
|
|
195
|
+
startUrl: result.startUrl,
|
|
196
|
+
source,
|
|
197
|
+
pages: result.records,
|
|
198
|
+
linkChecks: result.linkChecks,
|
|
199
|
+
linkSources: result.linkSources,
|
|
200
|
+
externalChecks: result.externalChecks,
|
|
201
|
+
externalSources: result.externalSources,
|
|
202
|
+
robots: result.robots,
|
|
203
|
+
robotsAllows: result.robotsAllows,
|
|
204
|
+
blockedByRobots: result.blockedByRobots,
|
|
205
|
+
sitemaps: result.sitemaps,
|
|
206
|
+
probes: result.probes,
|
|
207
|
+
crawl: result.stats,
|
|
208
|
+
thresholds,
|
|
209
|
+
pathOnly: source === 'directory',
|
|
210
|
+
unlinkedPages: result.unlinkedPages,
|
|
211
|
+
});
|
|
212
|
+
const disabled = new Set(options.disableRules ?? []);
|
|
213
|
+
siteFindings = sortFindings(siteFindings.filter((f) => !disabled.has(f.rule)));
|
|
214
|
+
if (options.includeInfo === false)
|
|
215
|
+
siteFindings = siteFindings.filter((f) => f.severity !== 'info');
|
|
216
|
+
const site = {
|
|
217
|
+
schemaVersion: 1,
|
|
218
|
+
kind: 'site',
|
|
219
|
+
target,
|
|
220
|
+
source,
|
|
221
|
+
auditedAt: new Date().toISOString(),
|
|
222
|
+
crawl: result.stats,
|
|
223
|
+
robotsTxt: result.robots,
|
|
224
|
+
sitemaps: { ...result.sitemaps, urls: result.sitemaps.urls.slice(0, 200) },
|
|
225
|
+
pages: pages.sort((a, b) => (a.depth ?? 0) - (b.depth ?? 0) || a.score - b.score),
|
|
226
|
+
siteFindings,
|
|
227
|
+
issues: [],
|
|
228
|
+
counts: { error: 0, warning: 0, info: 0 },
|
|
229
|
+
score: 0,
|
|
230
|
+
};
|
|
231
|
+
recomputeSite(site);
|
|
232
|
+
return site;
|
|
233
|
+
}
|
|
234
|
+
async function auditRecord(record, source, options) {
|
|
235
|
+
const thresholds = { ...DEFAULT_THRESHOLDS, ...(options.thresholds ?? {}) };
|
|
236
|
+
let findings = runPageRules({ page: record.page, source, response: record.response, keyword: options.keyword, thresholds, depth: record.depth }, { disable: new Set(options.disableRules ?? []) });
|
|
237
|
+
if (options.includeInfo === false)
|
|
238
|
+
findings = findings.filter((f) => f.severity !== 'info');
|
|
239
|
+
const audit = {
|
|
240
|
+
schemaVersion: 1,
|
|
241
|
+
kind: 'page',
|
|
242
|
+
target: record.file ?? record.requestedUrl,
|
|
243
|
+
url: record.file ?? record.url,
|
|
244
|
+
source,
|
|
245
|
+
auditedAt: new Date().toISOString(),
|
|
246
|
+
summary: buildSummary(record.page, record.response),
|
|
247
|
+
findings,
|
|
248
|
+
counts: countFindings(findings),
|
|
249
|
+
score: 0,
|
|
250
|
+
depth: record.depth,
|
|
251
|
+
};
|
|
252
|
+
if (record.response)
|
|
253
|
+
audit.response = record.response;
|
|
254
|
+
audit.score = pageScore(audit.counts);
|
|
255
|
+
return audit;
|
|
256
|
+
}
|
|
257
|
+
export function recomputeSite(site) {
|
|
258
|
+
const counts = countFindings(site.siteFindings);
|
|
259
|
+
for (const p of site.pages) {
|
|
260
|
+
counts.error += p.counts.error;
|
|
261
|
+
counts.warning += p.counts.warning;
|
|
262
|
+
counts.info += p.counts.info;
|
|
263
|
+
}
|
|
264
|
+
site.counts = counts;
|
|
265
|
+
site.issues = rollupIssues(site.pages);
|
|
266
|
+
site.score = siteScore(site.pages, site.siteFindings);
|
|
267
|
+
}
|
|
268
|
+
export function rollupIssues(pages) {
|
|
269
|
+
const map = new Map();
|
|
270
|
+
for (const page of pages) {
|
|
271
|
+
const seenHere = new Set();
|
|
272
|
+
for (const f of page.findings) {
|
|
273
|
+
if (seenHere.has(f.rule))
|
|
274
|
+
continue;
|
|
275
|
+
seenHere.add(f.rule);
|
|
276
|
+
let entry = map.get(f.rule);
|
|
277
|
+
if (!entry) {
|
|
278
|
+
entry = { rule: f.rule, severity: f.severity, category: f.category, pages: 0, message: f.message, recommendation: f.recommendation, examples: [], urls: new Set() };
|
|
279
|
+
map.set(f.rule, entry);
|
|
280
|
+
}
|
|
281
|
+
entry.pages++;
|
|
282
|
+
entry.urls.add(page.url);
|
|
283
|
+
}
|
|
284
|
+
}
|
|
285
|
+
const rank = { error: 0, warning: 1, info: 2 };
|
|
286
|
+
return [...map.values()]
|
|
287
|
+
.sort((a, b) => rank[a.severity] - rank[b.severity] || b.pages - a.pages)
|
|
288
|
+
.map(({ urls, ...rest }) => ({ ...rest, examples: [...urls].slice(0, 10) }));
|
|
289
|
+
}
|
|
290
|
+
function shapeSite(site, options) {
|
|
291
|
+
if (options.pageDetail === 'summary') {
|
|
292
|
+
site.pages = site.pages.map((p) => ({ ...p, findings: [], summary: { ...p.summary, headings: [], openGraph: {} } }));
|
|
293
|
+
}
|
|
294
|
+
return site;
|
|
295
|
+
}
|
|
296
|
+
async function readStdin() {
|
|
297
|
+
const chunks = [];
|
|
298
|
+
for await (const chunk of process.stdin)
|
|
299
|
+
chunks.push(Buffer.isBuffer(chunk) ? chunk : Buffer.from(chunk));
|
|
300
|
+
return Buffer.concat(chunks).toString('utf8');
|
|
301
|
+
}
|
|
302
|
+
/** Drop bulky, low-value fields to save tokens when the result is consumed by an LLM. */
|
|
303
|
+
export function compactAudit(audit) {
|
|
304
|
+
const trimFinding = (f) => {
|
|
305
|
+
if (!f.evidence)
|
|
306
|
+
return f;
|
|
307
|
+
const evidence = {};
|
|
308
|
+
for (const [k, v] of Object.entries(f.evidence))
|
|
309
|
+
evidence[k] = Array.isArray(v) ? v.slice(0, 3) : v;
|
|
310
|
+
return { ...f, evidence };
|
|
311
|
+
};
|
|
312
|
+
const trimPage = (p) => {
|
|
313
|
+
const { performance, rendering, ...rest } = p;
|
|
314
|
+
const out = {
|
|
315
|
+
...rest,
|
|
316
|
+
summary: { ...p.summary, headings: p.summary.headings.slice(0, 10), openGraph: {} },
|
|
317
|
+
findings: p.findings.map(trimFinding),
|
|
318
|
+
};
|
|
319
|
+
if (p.response)
|
|
320
|
+
out.response = { ...p.response, headers: {} };
|
|
321
|
+
if (performance)
|
|
322
|
+
out.performance = { ...performance, opportunities: performance.opportunities.slice(0, 5), agenticBrowsingAudits: [], failedSeoAudits: performance.failedSeoAudits.slice(0, 5) };
|
|
323
|
+
if (rendering)
|
|
324
|
+
out.rendering = { ...rendering, consoleErrors: rendering.consoleErrors.slice(0, 3) };
|
|
325
|
+
return out;
|
|
326
|
+
};
|
|
327
|
+
if (audit.kind === 'page')
|
|
328
|
+
return trimPage(audit);
|
|
329
|
+
return {
|
|
330
|
+
...audit,
|
|
331
|
+
robotsTxt: audit.robotsTxt ? { ...audit.robotsTxt, text: null } : null,
|
|
332
|
+
sitemaps: { ...audit.sitemaps, urls: audit.sitemaps.urls.slice(0, 20) },
|
|
333
|
+
pages: audit.pages.map(trimPage),
|
|
334
|
+
siteFindings: audit.siteFindings.map(trimFinding),
|
|
335
|
+
};
|
|
336
|
+
}
|
|
337
|
+
//# sourceMappingURL=audit.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"audit.js","sourceRoot":"","sources":["../../src/core/audit.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,QAAQ,IAAI,EAAE,EAAE,MAAM,SAAS,CAAC;AACzC,OAAO,IAAI,MAAM,WAAW,CAAC;AAC7B,OAAO,EAAE,SAAS,EAAmF,MAAM,YAAY,CAAC;AACxH,OAAO,EAAE,kBAAkB,EAAE,SAAS,EAAE,cAAc,EAAqB,MAAM,YAAY,CAAC;AAC9F,OAAO,EAAE,WAAW,EAAE,MAAM,mBAAmB,CAAC;AAChD,OAAO,EAAE,aAAa,EAAE,0BAA0B,EAAE,MAAM,iBAAiB,CAAC;AAC5E,OAAO,EAAE,mBAAmB,EAAE,MAAM,wBAAwB,CAAC;AAC7D,OAAO,EAAE,aAAa,EAAE,MAAM,mBAAmB,CAAC;AAClD,OAAO,EAAE,UAAU,EAAE,sBAAsB,EAAE,MAAM,aAAa,CAAC;AACjE,OAAO,EAAE,YAAY,EAAE,YAAY,EAAE,MAAM,kBAAkB,CAAC;AAC9D,OAAO,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,YAAY,CAAC;AAClD,OAAO,EAAE,YAAY,EAAmB,MAAM,iBAAiB,CAAC;AAChE,OAAO,EAAE,aAAa,EAAE,MAAM,aAAa,CAAC;AAC5C,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAC5C,OAAO,EACL,aAAa,EACb,kBAAkB,GASnB,MAAM,YAAY,CAAC;AACpB,OAAO,EAAE,SAAS,EAAE,MAAM,UAAU,CAAC;AA8CrC,MAAM,CAAC,MAAM,aAAa,GAAG,EAAE,QAAQ,EAAE,EAAE,EAAE,QAAQ,EAAE,CAAC,EAAE,WAAW,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,aAAa,EAAE,GAAG,EAAW,CAAC;AAEpH,4EAA4E;AAC5E,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,IAAY,EAAE,QAAwB,EAAE,EAAE,UAAwB,EAAE;IAClG,MAAM,GAAG,GAAG,KAAK,CAAC,GAAG,IAAI,OAAO,CAAC,OAAO,IAAI,mBAAmB,CAAC;IAChE,MAAM,IAAI,GAAG,WAAW,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;IACtF,MAAM,MAAM,GAAG,KAAK,CAAC,MAAM,IAAI,MAAM,CAAC;IACtC,MAAM,UAAU,GAAG,EAAE,GAAG,kBAAkB,EAAE,GAAG,CAAC,OAAO,CAAC,UAAU,IAAI,EAAE,CAAC,EAAE,CAAC;IAC5E,IAAI,QAAQ,GAAG,YAAY,CACzB,EAAE,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,KAAK,CAAC,QAAQ,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,UAAU,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,EACpG,EAAE,OAAO,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,YAAY,IAAI,EAAE,CAAC,EAAE,CACjD,CAAC;IACF,IAAI,OAAO,CAAC,WAAW,KAAK,KAAK;QAAE,QAAQ,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC;IAC5F,MAAM,KAAK,GAAc;QACvB,aAAa,EAAE,CAAC;QAChB,IAAI,EAAE,MAAM;QACZ,MAAM,EAAE,KAAK,CAAC,MAAM,IAAI,GAAG;QAC3B,GAAG,EAAE,KAAK,CAAC,QAAQ,EAAE,QAAQ,IAAI,GAAG;QACpC,MAAM;QACN,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACnC,OAAO,EAAE,YAAY,CAAC,IAAI,EAAE,KAAK,CAAC,QAAQ,CAAC;QAC3C,QAAQ;QACR,MAAM,EAAE,aAAa,CAAC,QAAQ,CAAC;QAC/B,KAAK,EAAE,CAAC;KACT,CAAC;IACF,IAAI,KAAK,CAAC,QAAQ;QAAE,KAAK,CAAC,QAAQ,GAAG,KAAK,CAAC,QAAQ,CAAC;IACpD,IAAI,KAAK,CAAC,KAAK,KAAK,SAAS;QAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;IACzD,KAAK,CAAC,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACtC,OAAO,KAAK,CAAC;AACf,CAAC;AAED,wEAAwE;AACxE,MAAM,CAAC,KAAK,UAAU,QAAQ,CAAC,GAAW,EAAE,UAAwB,EAAE;IACpE,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IACtD,MAAM,QAAQ,GAAG,cAAc,CAAC,GAAG,CAAC,CAAC;IACrC,MAAM,KAAK,GAAG,MAAM,SAAS,CAAC,GAAG,CAAC,IAAI,IAAI,EAAE,EAAE,EAAE,GAAG,EAAE,GAAG,CAAC,QAAQ,EAAE,QAAQ,EAAE,MAAM,EAAE,KAAK,EAAE,MAAM,EAAE,GAAG,EAAE,EAAE,OAAO,CAAC,CAAC;IACpH,IAAI,GAAG,CAAC,KAAK,EAAE,CAAC;QACd,KAAK,CAAC,QAAQ,CAAC,OAAO,CAAC;YACrB,IAAI,EAAE,uBAAuB;YAC7B,QAAQ,EAAE,OAAO;YACjB,QAAQ,EAAE,UAAU;YACpB,OAAO,EAAE,mBAAmB,GAAG,KAAK,GAAG,CAAC,KAAK,EAAE;YAC/C,cAAc,EAAE,qIAAqI;YACrJ,GAAG;YACH,QAAQ,EAAE,EAAE,KAAK,EAAE,GAAG,CAAC,KAAK,EAAE,aAAa,EAAE,GAAG,CAAC,aAAa,EAAE;SACjE,CAAC,CAAC;QACH,KAAK,CAAC,MAAM,GAAG,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QAC7C,KAAK,CAAC,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;QACtC,OAAO,KAAK,CAAC;IACf,CAAC;IACD,MAAM,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;IACjC,OAAO,KAAK,CAAC;AACf,CAAC;AAED,2CAA2C;AAC3C,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,QAAgB,EAAE,UAAwB,EAAE;IAC1E,MAAM,IAAI,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,QAAQ,EAAE,MAAM,CAAC,CAAC;IACjD,MAAM,IAAI,GAAG,OAAO,CAAC,OAAO,IAAI,mBAAmB,CAAC;IACpD,MAAM,GAAG,GAAG,IAAI,GAAG,CAAC,IAAI,IAAI,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAAE,EAAE,IAAI,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,GAAG,IAAI,GAAG,CAAC,CAAC,IAAI,CAAC;IAChG,OAAO,SAAS,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,EAAE,OAAO,CAAC,CAAC;AAC7E,CAAC;AAED,gEAAgE;AAChE,MAAM,CAAC,KAAK,UAAU,SAAS,CAAC,QAAgB,EAAE,UAAwB,EAAE;IAC1E,MAAM,SAAS,GAAiB,EAAE,GAAG,CAAC,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,EAAE,CAAC;IAC7D,SAAS,CAAC,SAAS,KAAK,kBAAkB,CAAC;IAC3C,MAAM,QAAQ,GAAyB;QACrC,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,aAAa,CAAC,QAAQ;QACpD,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,aAAa,CAAC,QAAQ;QACpD,WAAW,EAAE,OAAO,CAAC,WAAW,IAAI,aAAa,CAAC,WAAW;QAC7D,OAAO,EAAE,OAAO,CAAC,OAAO,IAAI,aAAa,CAAC,OAAO;QACjD,aAAa,EAAE,OAAO,CAAC,aAAa,IAAI,IAAI;QAC5C,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,IAAI,KAAK;QACrD,kBAAkB,EAAE,OAAO,CAAC,kBAAkB,IAAI,KAAK;QACvD,aAAa,EAAE,OAAO,CAAC,aAAa,IAAI,aAAa,CAAC,aAAa;QACnE,eAAe,EAAE,CAAC,OAAO,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC;QAC1E,eAAe,EAAE,CAAC,OAAO,CAAC,eAAe,IAAI,EAAE,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,MAAM,CAAC,CAAC,CAAC,CAAC;QAC1E,KAAK,EAAE,SAAS;QAChB,UAAU,EAAE,OAAO,CAAC,UAAU;KAC/B,CAAC;IACF,MAAM,MAAM,GAAG,MAAM,SAAS,CAAC,QAAQ,EAAE,QAAQ,CAAC,CAAC;IACnD,MAAM,IAAI,GAAG,MAAM,cAAc,CAAC,MAAM,EAAE,OAAO,EAAE,QAAQ,EAAE,OAAO,CAAC,CAAC;IACtE,sEAAsE;IACtE,MAAM,SAAS,GAAG,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,KAAK,KAAK,CAAC,CAAC,IAAI,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IACzE,IAAI,SAAS,IAAI,CAAC,OAAO,CAAC,MAAM,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,CAAC;QACxD,MAAM,UAAU,CAAC,SAAS,EAAE,OAAO,CAAC,CAAC;QACrC,aAAa,CAAC,IAAI,CAAC,CAAC;IACtB,CAAC;IACD,OAAO,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;AAClC,CAAC;AAED,mEAAmE;AACnE,MAAM,CAAC,KAAK,UAAU,cAAc,CAAC,GAAW,EAAE,UAAwB,EAAE;IAC1E,MAAM,MAAM,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE;QACtC,OAAO,EAAE,OAAO,CAAC,OAAO,IAAI,mBAAmB;QAC/C,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,IAAI,KAAK;QACrD,QAAQ,EAAE,OAAO,CAAC,QAAQ,IAAI,GAAG;QACjC,SAAS,EAAE,OAAO,CAAC,KAAK,EAAE,SAAS;KACpC,CAAC,CAAC;IACH,MAAM,IAAI,GAAG,MAAM,cAAc,CAAC,MAAM,EAAE,WAAW,EAAE,GAAG,EAAE,OAAO,CAAC,CAAC;IACrE,OAAO,SAAS,CAAC,IAAI,EAAE,OAAO,CAAC,CAAC;AAClC,CAAC;AAED;;;GAGG;AACH,MAAM,CAAC,KAAK,UAAU,WAAW,CAAC,MAAc,EAAE,UAAkE,EAAE;IACpH,IAAI,MAAM,KAAK,GAAG,IAAI,OAAO,CAAC,SAAS,KAAK,SAAS,EAAE,CAAC;QACtD,MAAM,IAAI,GAAG,OAAO,CAAC,SAAS,IAAI,CAAC,MAAM,SAAS,EAAE,CAAC,CAAC;QACtD,OAAO,SAAS,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,OAAO,CAAC,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,OAAO,EAAE,EAAE,OAAO,CAAC,CAAC;IAC7F,CAAC;IACD,IAAI,SAAS,CAAC,MAAM,CAAC;QAAE,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IACrG,IAAI,0BAA0B,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,4BAA4B,CAAC,IAAI,CAAC,MAAM,CAAC,EAAE,CAAC;QACzF,MAAM,GAAG,GAAG,OAAO,aAAa,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,MAAM,MAAM,EAAE,CAAC;QACvE,OAAO,OAAO,CAAC,KAAK,CAAC,CAAC,CAAC,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;IAC1E,CAAC;IACD,MAAM,IAAI,GAAG,MAAM,EAAE,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC;IACrD,IAAI,CAAC,IAAI;QAAE,MAAM,IAAI,KAAK,CAAC,WAAW,MAAM,oCAAoC,CAAC,CAAC;IAClF,IAAI,IAAI,CAAC,WAAW,EAAE;QAAE,OAAO,cAAc,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;IAC/D,OAAO,SAAS,CAAC,MAAM,EAAE,OAAO,CAAC,CAAC;AACpC,CAAC;AAED,KAAK,UAAU,UAAU,CAAC,KAAgB,EAAE,OAAqB;IAC/D,MAAM,GAAG,GAAG,KAAK,CAAC,QAAQ,EAAE,QAAQ,IAAI,KAAK,CAAC,GAAG,CAAC;IAClD,MAAM,KAAK,GAAc,EAAE,CAAC;IAC5B,IAAI,OAAO,CAAC,MAAM,EAAE,CAAC;QACnB,OAAO,CAAC,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,aAAa,GAAG,kBAAkB,EAAE,CAAC,CAAC;QACvF,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,UAAU,CAAC,GAAG,EAAE,EAAE,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,OAAO,CAAC,KAAK,EAAE,SAAS,IAAI,CAAC,EAAE,MAAM,CAAC,EAAE,SAAS,EAAE,OAAO,CAAC,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC;YAC7K,MAAM,OAAO,GAAG,WAAW,CAAC,MAAM,UAAU,CAAC,KAAK,EAAE,OAAO,CAAC,EAAE,EAAE,GAAG,EAAE,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;YACrH,MAAM,YAAY,GAAG,WAAW,CAAC,QAAQ,CAAC,IAAI,EAAE,EAAE,GAAG,EAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,EAAE,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC,QAAQ,EAAE,iBAAiB,EAAE,OAAO,CAAC,iBAAiB,EAAE,CAAC,CAAC;YAC3J,MAAM,EAAE,UAAU,EAAE,QAAQ,EAAE,GAAG,aAAa,CAAC,OAAO,EAAE,YAAY,EAAE,QAAQ,CAAC,QAAQ,EAAE,QAAQ,CAAC,aAAa,EAAE,KAAK,CAAC,QAAQ,CAAC,CAAC;YACjI,KAAK,CAAC,SAAS,GAAG,UAAU,CAAC;YAC7B,KAAK,CAAC,IAAI,CAAC,GAAG,QAAQ,CAAC,CAAC;QAC1B,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,yBAAyB,EAAE,GAAG,YAAY,sBAAsB,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,qBAAsB,GAAa,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;QAC/J,CAAC;IACH,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,EAAE,CAAC;QACvB,OAAO,CAAC,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,YAAY,EAAE,OAAO,EAAE,uBAAuB,OAAO,CAAC,UAAU,IAAI,QAAQ,QAAQ,GAAG,EAAE,EAAE,CAAC,CAAC;QAC3H,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,EAAE,UAAU,EAAE,OAAO,CAAC,UAAU,IAAI,QAAQ,EAAE,SAAS,EAAE,OAAO,CAAC,KAAK,EAAE,SAAS,EAAE,OAAO,EAAE,OAAO,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,CAAC;YAC5J,KAAK,CAAC,WAAW,GAAG,IAAI,CAAC;YACzB,KAAK,CAAC,IAAI,CAAC,GAAG,mBAAmB,CAAC,IAAI,CAAC,CAAC,CAAC;QAC3C,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,KAAK,CAAC,IAAI,CAAC,WAAW,CAAC,6BAA6B,EAAE,GAAG,YAAY,0BAA0B,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,sBAAuB,GAAa,CAAC,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,CAAC;QACxK,CAAC;IACH,CAAC;IACD,IAAI,KAAK,CAAC,MAAM,EAAE,CAAC;QACjB,MAAM,QAAQ,GAAG,OAAO,CAAC,WAAW,KAAK,KAAK,CAAC,CAAC,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC;QACpG,KAAK,CAAC,QAAQ,GAAG,YAAY,CAAC,CAAC,GAAG,KAAK,CAAC,QAAQ,EAAE,GAAG,QAAQ,CAAC,CAAC,CAAC;QAChE,KAAK,CAAC,MAAM,GAAG,aAAa,CAAC,KAAK,CAAC,QAAQ,CAAC,CAAC;QAC7C,KAAK,CAAC,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACxC,CAAC;AACH,CAAC;AAED,4FAA4F;AAC5F,KAAK,UAAU,UAAU,CAAC,KAAgB,EAAE,OAAqB;IAC/D,MAAM,GAAG,GAAG,MAAM,SAAS,CAAC,KAAK,CAAC,QAAQ,EAAE,QAAQ,IAAI,KAAK,CAAC,GAAG,EAAE,OAAO,CAAC,KAAK,IAAI,EAAE,CAAC,CAAC;IACxF,OAAO,GAAG,CAAC,IAAI,IAAI,EAAE,CAAC;AACxB,CAAC;AAED,SAAS,WAAW,CAAC,IAAY,EAAE,OAAe,EAAE,GAAW;IAC7D,OAAO;QACL,IAAI;QACJ,QAAQ,EAAE,MAAM;QAChB,QAAQ,EAAE,MAAM;QAChB,OAAO;QACP,cAAc,EAAE,8GAA8G;QAC9H,GAAG;KACJ,CAAC;AACJ,CAAC;AAED,KAAK,UAAU,cAAc,CAAC,MAAmB,EAAE,MAA6B,EAAE,MAAc,EAAE,OAAqB;IACrH,OAAO,CAAC,UAAU,EAAE,CAAC,EAAE,KAAK,EAAE,OAAO,EAAE,OAAO,EAAE,wBAAwB,MAAM,CAAC,OAAO,CAAC,MAAM,QAAQ,EAAE,CAAC,CAAC;IACzG,MAAM,UAAU,GAAG,EAAE,GAAG,kBAAkB,EAAE,GAAG,CAAC,OAAO,CAAC,UAAU,IAAI,EAAE,CAAC,EAAE,CAAC;IAC5E,MAAM,KAAK,GAAgB,EAAE,CAAC;IAC9B,KAAK,MAAM,MAAM,IAAI,MAAM,CAAC,OAAO,EAAE,CAAC;QACpC,MAAM,KAAK,GAAG,MAAM,WAAW,CAAC,MAAM,EAAE,MAAM,KAAK,WAAW,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,KAAK,EAAE,OAAO,CAAC,CAAC;QAC1F,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACpB,CAAC;IACD,IAAI,YAAY,GAAG,YAAY,CAAC;QAC9B,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,MAAM;QACN,KAAK,EAAE,MAAM,CAAC,OAAO;QACrB,UAAU,EAAE,MAAM,CAAC,UAAU;QAC7B,WAAW,EAAE,MAAM,CAAC,WAAW;QAC/B,cAAc,EAAE,MAAM,CAAC,cAAc;QACrC,eAAe,EAAE,MAAM,CAAC,eAAe;QACvC,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,YAAY,EAAE,MAAM,CAAC,YAAY;QACjC,eAAe,EAAE,MAAM,CAAC,eAAe;QACvC,QAAQ,EAAE,MAAM,CAAC,QAAQ;QACzB,MAAM,EAAE,MAAM,CAAC,MAAM;QACrB,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,UAAU;QACV,QAAQ,EAAE,MAAM,KAAK,WAAW;QAChC,aAAa,EAAE,MAAM,CAAC,aAAa;KACpC,CAAC,CAAC;IACH,MAAM,QAAQ,GAAG,IAAI,GAAG,CAAC,OAAO,CAAC,YAAY,IAAI,EAAE,CAAC,CAAC;IACrD,YAAY,GAAG,YAAY,CAAC,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;IAC/E,IAAI,OAAO,CAAC,WAAW,KAAK,KAAK;QAAE,YAAY,GAAG,YAAY,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC;IAEpG,MAAM,IAAI,GAAc;QACtB,aAAa,EAAE,CAAC;QAChB,IAAI,EAAE,MAAM;QACZ,MAAM;QACN,MAAM;QACN,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACnC,KAAK,EAAE,MAAM,CAAC,KAAK;QACnB,SAAS,EAAE,MAAM,CAAC,MAAM;QACxB,QAAQ,EAAE,EAAE,GAAG,MAAM,CAAC,QAAQ,EAAE,IAAI,EAAE,MAAM,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE;QAC1E,KAAK,EAAE,KAAK,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,GAAG,CAAC,CAAC,CAAC,KAAK,IAAI,CAAC,CAAC,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;QACjF,YAAY;QACZ,MAAM,EAAE,EAAE;QACV,MAAM,EAAE,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAE;QACzC,KAAK,EAAE,CAAC;KACT,CAAC;IACF,aAAa,CAAC,IAAI,CAAC,CAAC;IACpB,OAAO,IAAI,CAAC;AACd,CAAC;AAED,KAAK,UAAU,WAAW,CAAC,MAAkB,EAAE,MAAkB,EAAE,OAAqB;IACtF,MAAM,UAAU,GAAG,EAAE,GAAG,kBAAkB,EAAE,GAAG,CAAC,OAAO,CAAC,UAAU,IAAI,EAAE,CAAC,EAAE,CAAC;IAC5E,IAAI,QAAQ,GAAG,YAAY,CACzB,EAAE,IAAI,EAAE,MAAM,CAAC,IAAI,EAAE,MAAM,EAAE,QAAQ,EAAE,MAAM,CAAC,QAAQ,EAAE,OAAO,EAAE,OAAO,CAAC,OAAO,EAAE,UAAU,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,EAAE,EACnH,EAAE,OAAO,EAAE,IAAI,GAAG,CAAC,OAAO,CAAC,YAAY,IAAI,EAAE,CAAC,EAAE,CACjD,CAAC;IACF,IAAI,OAAO,CAAC,WAAW,KAAK,KAAK;QAAE,QAAQ,GAAG,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC,QAAQ,KAAK,MAAM,CAAC,CAAC;IAC5F,MAAM,KAAK,GAAc;QACvB,aAAa,EAAE,CAAC;QAChB,IAAI,EAAE,MAAM;QACZ,MAAM,EAAE,MAAM,CAAC,IAAI,IAAI,MAAM,CAAC,YAAY;QAC1C,GAAG,EAAE,MAAM,CAAC,IAAI,IAAI,MAAM,CAAC,GAAG;QAC9B,MAAM;QACN,SAAS,EAAE,IAAI,IAAI,EAAE,CAAC,WAAW,EAAE;QACnC,OAAO,EAAE,YAAY,CAAC,MAAM,CAAC,IAAI,EAAE,MAAM,CAAC,QAAQ,CAAC;QACnD,QAAQ;QACR,MAAM,EAAE,aAAa,CAAC,QAAQ,CAAC;QAC/B,KAAK,EAAE,CAAC;QACR,KAAK,EAAE,MAAM,CAAC,KAAK;KACpB,CAAC;IACF,IAAI,MAAM,CAAC,QAAQ;QAAE,KAAK,CAAC,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC;IACtD,KAAK,CAAC,KAAK,GAAG,SAAS,CAAC,KAAK,CAAC,MAAM,CAAC,CAAC;IACtC,OAAO,KAAK,CAAC;AACf,CAAC;AAED,MAAM,UAAU,aAAa,CAAC,IAAe;IAC3C,MAAM,MAAM,GAAG,aAAa,CAAC,IAAI,CAAC,YAAY,CAAC,CAAC;IAChD,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,KAAK,EAAE,CAAC;QAC3B,MAAM,CAAC,KAAK,IAAI,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC;QAC/B,MAAM,CAAC,OAAO,IAAI,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC;QACnC,MAAM,CAAC,IAAI,IAAI,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC;IAC/B,CAAC;IACD,IAAI,CAAC,MAAM,GAAG,MAAM,CAAC;IACrB,IAAI,CAAC,MAAM,GAAG,YAAY,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;IACvC,IAAI,CAAC,KAAK,GAAG,SAAS,CAAC,IAAI,CAAC,KAAK,EAAE,IAAI,CAAC,YAAY,CAAC,CAAC;AACxD,CAAC;AAED,MAAM,UAAU,YAAY,CAAC,KAAkB;IAC7C,MAAM,GAAG,GAAG,IAAI,GAAG,EAA+C,CAAC;IACnE,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;QACzB,MAAM,QAAQ,GAAG,IAAI,GAAG,EAAU,CAAC;QACnC,KAAK,MAAM,CAAC,IAAI,IAAI,CAAC,QAAQ,EAAE,CAAC;YAC9B,IAAI,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC;gBAAE,SAAS;YACnC,QAAQ,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YACrB,IAAI,KAAK,GAAG,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC;YAC5B,IAAI,CAAC,KAAK,EAAE,CAAC;gBACX,KAAK,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,QAAQ,EAAE,CAAC,CAAC,QAAQ,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,OAAO,EAAE,cAAc,EAAE,CAAC,CAAC,cAAc,EAAE,QAAQ,EAAE,EAAE,EAAE,IAAI,EAAE,IAAI,GAAG,EAAE,EAAE,CAAC;gBACpK,GAAG,CAAC,GAAG,CAAC,CAAC,CAAC,IAAI,EAAE,KAAK,CAAC,CAAC;YACzB,CAAC;YACD,KAAK,CAAC,KAAK,EAAE,CAAC;YACd,KAAK,CAAC,IAAI,CAAC,GAAG,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QAC3B,CAAC;IACH,CAAC;IACD,MAAM,IAAI,GAAG,EAAE,KAAK,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,CAAC,EAAW,CAAC;IACxD,OAAO,CAAC,GAAG,GAAG,CAAC,MAAM,EAAE,CAAC;SACrB,IAAI,CAAC,CAAC,CAAC,EAAE,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,GAAG,IAAI,CAAC,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,GAAG,CAAC,CAAC,KAAK,CAAC;SACxE,GAAG,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,IAAI,EAAE,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,IAAI,EAAE,QAAQ,EAAE,CAAC,GAAG,IAAI,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;AACjF,CAAC;AAED,SAAS,SAAS,CAAC,IAAe,EAAE,OAAqB;IACvD,IAAI,OAAO,CAAC,UAAU,KAAK,SAAS,EAAE,CAAC;QACrC,IAAI,CAAC,KAAK,GAAG,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,QAAQ,EAAE,EAAE,EAAE,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,OAAO,EAAE,QAAQ,EAAE,EAAE,EAAE,SAAS,EAAE,EAAE,EAAE,EAAE,CAAC,CAAC,CAAC;IACvH,CAAC;IACD,OAAO,IAAI,CAAC;AACd,CAAC;AAED,KAAK,UAAU,SAAS;IACtB,MAAM,MAAM,GAAa,EAAE,CAAC;IAC5B,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,OAAO,CAAC,KAAK;QAAE,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,CAAC;IAC1G,OAAO,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC;AAChD,CAAC;AAED,yFAAyF;AACzF,MAAM,UAAU,YAAY,CAAkB,KAAQ;IACpD,MAAM,WAAW,GAAG,CAAC,CAAU,EAAW,EAAE;QAC1C,IAAI,CAAC,CAAC,CAAC,QAAQ;YAAE,OAAO,CAAC,CAAC;QAC1B,MAAM,QAAQ,GAA4B,EAAE,CAAC;QAC7C,KAAK,MAAM,CAAC,CAAC,EAAE,CAAC,CAAC,IAAI,MAAM,CAAC,OAAO,CAAC,CAAC,CAAC,QAAQ,CAAC;YAAE,QAAQ,CAAC,CAAC,CAAC,GAAG,KAAK,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;QACpG,OAAO,EAAE,GAAG,CAAC,EAAE,QAAQ,EAAE,CAAC;IAC5B,CAAC,CAAC;IACF,MAAM,QAAQ,GAAG,CAAC,CAAY,EAAa,EAAE;QAC3C,MAAM,EAAE,WAAW,EAAE,SAAS,EAAE,GAAG,IAAI,EAAE,GAAG,CAAC,CAAC;QAC9C,MAAM,GAAG,GAAc;YACrB,GAAG,IAAI;YACP,OAAO,EAAE,EAAE,GAAG,CAAC,CAAC,OAAO,EAAE,QAAQ,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE,SAAS,EAAE,EAAE,EAAE;YACnF,QAAQ,EAAE,CAAC,CAAC,QAAQ,CAAC,GAAG,CAAC,WAAW,CAAC;SACtC,CAAC;QACF,IAAI,CAAC,CAAC,QAAQ;YAAE,GAAG,CAAC,QAAQ,GAAG,EAAE,GAAG,CAAC,CAAC,QAAQ,EAAE,OAAO,EAAE,EAAE,EAAE,CAAC;QAC9D,IAAI,WAAW;YAAE,GAAG,CAAC,WAAW,GAAG,EAAE,GAAG,WAAW,EAAE,aAAa,EAAE,WAAW,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,qBAAqB,EAAE,EAAE,EAAE,eAAe,EAAE,WAAW,CAAC,eAAe,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACjM,IAAI,SAAS;YAAE,GAAG,CAAC,SAAS,GAAG,EAAE,GAAG,SAAS,EAAE,aAAa,EAAE,SAAS,CAAC,aAAa,CAAC,KAAK,CAAC,CAAC,EAAE,CAAC,CAAC,EAAE,CAAC;QACpG,OAAO,GAAG,CAAC;IACb,CAAC,CAAC;IACF,IAAI,KAAK,CAAC,IAAI,KAAK,MAAM;QAAE,OAAO,QAAQ,CAAC,KAAK,CAAM,CAAC;IACvD,OAAO;QACL,GAAG,KAAK;QACR,SAAS,EAAE,KAAK,CAAC,SAAS,CAAC,CAAC,CAAC,EAAE,GAAG,KAAK,CAAC,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,IAAI;QACtE,QAAQ,EAAE,EAAE,GAAG,KAAK,CAAC,QAAQ,EAAE,IAAI,EAAE,KAAK,CAAC,QAAQ,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,EAAE,EAAE,CAAC,EAAE;QACvE,KAAK,EAAE,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,CAAC;QAChC,YAAY,EAAE,KAAK,CAAC,YAAY,CAAC,GAAG,CAAC,WAAW,CAAC;KAC7C,CAAC;AACT,CAAC"}
|
|
@@ -0,0 +1,4 @@
|
|
|
1
|
+
/** Minimal promise pool. */
|
|
2
|
+
export declare function createPool(concurrency: number): <T>(task: () => Promise<T>) => Promise<T>;
|
|
3
|
+
export declare function mapConcurrent<T, R>(items: T[], concurrency: number, fn: (item: T, index: number) => Promise<R>): Promise<R[]>;
|
|
4
|
+
export declare function sleep(ms: number): Promise<void>;
|
|
@@ -0,0 +1,29 @@
|
|
|
1
|
+
/** Minimal promise pool. */
|
|
2
|
+
export function createPool(concurrency) {
|
|
3
|
+
let active = 0;
|
|
4
|
+
const queue = [];
|
|
5
|
+
const next = () => {
|
|
6
|
+
active--;
|
|
7
|
+
queue.shift()?.();
|
|
8
|
+
};
|
|
9
|
+
return function run(task) {
|
|
10
|
+
return new Promise((resolve, reject) => {
|
|
11
|
+
const start = () => {
|
|
12
|
+
active++;
|
|
13
|
+
task().then(resolve, reject).finally(next);
|
|
14
|
+
};
|
|
15
|
+
if (active < Math.max(1, concurrency))
|
|
16
|
+
start();
|
|
17
|
+
else
|
|
18
|
+
queue.push(start);
|
|
19
|
+
});
|
|
20
|
+
};
|
|
21
|
+
}
|
|
22
|
+
export async function mapConcurrent(items, concurrency, fn) {
|
|
23
|
+
const run = createPool(concurrency);
|
|
24
|
+
return Promise.all(items.map((item, i) => run(() => fn(item, i))));
|
|
25
|
+
}
|
|
26
|
+
export function sleep(ms) {
|
|
27
|
+
return new Promise((r) => setTimeout(r, ms));
|
|
28
|
+
}
|
|
29
|
+
//# sourceMappingURL=concurrency.js.map
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"concurrency.js","sourceRoot":"","sources":["../../src/core/concurrency.ts"],"names":[],"mappings":"AAAA,4BAA4B;AAC5B,MAAM,UAAU,UAAU,CAAC,WAAmB;IAC5C,IAAI,MAAM,GAAG,CAAC,CAAC;IACf,MAAM,KAAK,GAAmB,EAAE,CAAC;IACjC,MAAM,IAAI,GAAG,GAAG,EAAE;QAChB,MAAM,EAAE,CAAC;QACT,KAAK,CAAC,KAAK,EAAE,EAAE,EAAE,CAAC;IACpB,CAAC,CAAC;IACF,OAAO,SAAS,GAAG,CAAI,IAAsB;QAC3C,OAAO,IAAI,OAAO,CAAI,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE;YACxC,MAAM,KAAK,GAAG,GAAG,EAAE;gBACjB,MAAM,EAAE,CAAC;gBACT,IAAI,EAAE,CAAC,IAAI,CAAC,OAAO,EAAE,MAAM,CAAC,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC7C,CAAC,CAAC;YACF,IAAI,MAAM,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,WAAW,CAAC;gBAAE,KAAK,EAAE,CAAC;;gBAC1C,KAAK,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;QACzB,CAAC,CAAC,CAAC;IACL,CAAC,CAAC;AACJ,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,aAAa,CAAO,KAAU,EAAE,WAAmB,EAAE,EAA0C;IACnH,MAAM,GAAG,GAAG,UAAU,CAAC,WAAW,CAAC,CAAC;IACpC,OAAO,OAAO,CAAC,GAAG,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,CAAC,EAAE,EAAE,CAAC,GAAG,CAAC,GAAG,EAAE,CAAC,EAAE,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AACrE,CAAC;AAED,MAAM,UAAU,KAAK,CAAC,EAAU;IAC9B,OAAO,IAAI,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,UAAU,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,CAAC;AAC/C,CAAC"}
|
|
@@ -0,0 +1,40 @@
|
|
|
1
|
+
import { type FetchOptions } from './fetch.js';
|
|
2
|
+
import type { LinkCheck, PageRecord, ProbeResults } from './site-rules.js';
|
|
3
|
+
import type { CrawlStats, RobotsTxtInfo, SitemapInfo } from './types.js';
|
|
4
|
+
export interface ProgressEvent {
|
|
5
|
+
phase: 'robots' | 'sitemap' | 'crawl' | 'links' | 'probes' | 'rules' | 'render' | 'lighthouse';
|
|
6
|
+
message: string;
|
|
7
|
+
crawled?: number;
|
|
8
|
+
discovered?: number;
|
|
9
|
+
url?: string;
|
|
10
|
+
}
|
|
11
|
+
export interface CrawlOptions {
|
|
12
|
+
maxPages: number;
|
|
13
|
+
maxDepth: number;
|
|
14
|
+
concurrency: number;
|
|
15
|
+
delayMs: number;
|
|
16
|
+
respectRobots: boolean;
|
|
17
|
+
includeSubdomains: boolean;
|
|
18
|
+
checkExternalLinks: boolean;
|
|
19
|
+
maxLinkChecks: number;
|
|
20
|
+
includePatterns: RegExp[];
|
|
21
|
+
excludePatterns: RegExp[];
|
|
22
|
+
fetch: FetchOptions;
|
|
23
|
+
onProgress?: (event: ProgressEvent) => void;
|
|
24
|
+
}
|
|
25
|
+
export interface CrawlResult {
|
|
26
|
+
startUrl: string;
|
|
27
|
+
records: PageRecord[];
|
|
28
|
+
linkChecks: Map<string, LinkCheck>;
|
|
29
|
+
linkSources: Map<string, Set<string>>;
|
|
30
|
+
externalChecks?: Map<string, LinkCheck>;
|
|
31
|
+
externalSources?: Map<string, Set<string>>;
|
|
32
|
+
robots: RobotsTxtInfo | null;
|
|
33
|
+
robotsAllows: ((url: string) => boolean) | null;
|
|
34
|
+
blockedByRobots: string[];
|
|
35
|
+
sitemaps: SitemapInfo;
|
|
36
|
+
probes: ProbeResults;
|
|
37
|
+
stats: CrawlStats;
|
|
38
|
+
unlinkedPages?: string[];
|
|
39
|
+
}
|
|
40
|
+
export declare function crawlSite(startInput: string, options: CrawlOptions): Promise<CrawlResult>;
|
|
@@ -0,0 +1,195 @@
|
|
|
1
|
+
import { extractPage } from './html/extract.js';
|
|
2
|
+
import { DEFAULT_USER_AGENT, fetchPage, isHtmlResponse, probeUrl, toResponseInfo } from './fetch.js';
|
|
3
|
+
import { loadRobotsTxt } from './robots.js';
|
|
4
|
+
import { loadSitemaps } from './sitemap.js';
|
|
5
|
+
import { isLocalHost } from './rules/canonical.js';
|
|
6
|
+
import { hostOf, isSameSite, looksLikeHtml, normalizeUrl, stripWww } from './url.js';
|
|
7
|
+
import { mapConcurrent, sleep } from './concurrency.js';
|
|
8
|
+
export async function crawlSite(startInput, options) {
|
|
9
|
+
const started = performance.now();
|
|
10
|
+
const progress = options.onProgress ?? (() => undefined);
|
|
11
|
+
const startUrl = normalizeUrl(startInput) ?? startInput;
|
|
12
|
+
const origin = new URL(startUrl).origin;
|
|
13
|
+
const startHost = hostOf(startUrl) ?? '';
|
|
14
|
+
const userAgent = options.fetch.userAgent ?? DEFAULT_USER_AGENT;
|
|
15
|
+
progress({ phase: 'robots', message: `Fetching ${origin}/robots.txt` });
|
|
16
|
+
const robots = await loadRobotsTxt(origin, userAgent, options.fetch);
|
|
17
|
+
const allows = options.respectRobots ? robots.allows : () => true;
|
|
18
|
+
progress({ phase: 'sitemap', message: 'Fetching sitemaps' });
|
|
19
|
+
const sitemapCandidates = [...new Set([...robots.info.sitemaps, new URL('/sitemap.xml', origin).href])];
|
|
20
|
+
const sitemaps = await loadSitemaps(sitemapCandidates, options.fetch);
|
|
21
|
+
const visited = new Set([startUrl]);
|
|
22
|
+
const discovered = new Set([startUrl]);
|
|
23
|
+
const linkChecks = new Map();
|
|
24
|
+
const linkSources = new Map();
|
|
25
|
+
const externalSources = new Map();
|
|
26
|
+
const records = [];
|
|
27
|
+
const blocked = [];
|
|
28
|
+
let limitSkips = 0;
|
|
29
|
+
const accepts = (url) => {
|
|
30
|
+
if (options.includePatterns.length && !options.includePatterns.some((re) => re.test(url)))
|
|
31
|
+
return false;
|
|
32
|
+
if (options.excludePatterns.some((re) => re.test(url)))
|
|
33
|
+
return false;
|
|
34
|
+
return true;
|
|
35
|
+
};
|
|
36
|
+
const queue = [{ url: startUrl, depth: 0 }];
|
|
37
|
+
const processItem = async (item) => {
|
|
38
|
+
if (!allows(item.url)) {
|
|
39
|
+
blocked.push(item.url);
|
|
40
|
+
return;
|
|
41
|
+
}
|
|
42
|
+
const res = await fetchPage(item.url, options.fetch);
|
|
43
|
+
const requestedKey = item.url;
|
|
44
|
+
const finalKey = normalizeUrl(res.finalUrl) ?? res.finalUrl;
|
|
45
|
+
linkChecks.set(requestedKey, { url: item.url, status: res.status, finalUrl: res.finalUrl, redirectChain: res.redirectChain, error: res.error });
|
|
46
|
+
if (finalKey !== requestedKey) {
|
|
47
|
+
if (!linkChecks.has(finalKey))
|
|
48
|
+
linkChecks.set(finalKey, { url: res.finalUrl, status: res.status, finalUrl: res.finalUrl, redirectChain: [], error: res.error });
|
|
49
|
+
if (visited.has(finalKey) && records.some((r) => r.url === finalKey))
|
|
50
|
+
return;
|
|
51
|
+
visited.add(finalKey);
|
|
52
|
+
if (!isSameSite(finalKey, startHost, options.includeSubdomains))
|
|
53
|
+
return; // redirected off-site
|
|
54
|
+
}
|
|
55
|
+
progress({ phase: 'crawl', message: `${res.status ?? 'ERR'} ${res.finalUrl}`, crawled: records.length + 1, discovered: discovered.size, url: res.finalUrl });
|
|
56
|
+
if (res.error || typeof res.body !== 'string' || !isHtmlResponse(res)) {
|
|
57
|
+
if (res.status !== null && res.status < 400 && !res.error && typeof res.body === 'string' && !isHtmlResponse(res))
|
|
58
|
+
return; // non-HTML asset
|
|
59
|
+
// Record error pages so page-level findings (response.error-status) appear for the start URL.
|
|
60
|
+
if (item.depth === 0 || res.status === null) {
|
|
61
|
+
const page = extractPage(res.body ?? '', { url: res.finalUrl, siteHost: startHost, includeSubdomains: options.includeSubdomains });
|
|
62
|
+
records.push({ url: finalKey, requestedUrl: item.url, page, response: toResponseInfo(res), depth: item.depth });
|
|
63
|
+
}
|
|
64
|
+
return;
|
|
65
|
+
}
|
|
66
|
+
const page = extractPage(res.body, { url: res.finalUrl, siteHost: startHost, includeSubdomains: options.includeSubdomains });
|
|
67
|
+
records.push({ url: finalKey, requestedUrl: item.url, page, response: toResponseInfo(res), depth: item.depth });
|
|
68
|
+
for (const link of page.links) {
|
|
69
|
+
if (!link.url)
|
|
70
|
+
continue;
|
|
71
|
+
const target = normalizeUrl(link.url);
|
|
72
|
+
if (!target)
|
|
73
|
+
continue;
|
|
74
|
+
if (!link.internal) {
|
|
75
|
+
if (!externalSources.has(target))
|
|
76
|
+
externalSources.set(target, new Set());
|
|
77
|
+
externalSources.get(target).add(finalKey);
|
|
78
|
+
continue;
|
|
79
|
+
}
|
|
80
|
+
if (!linkSources.has(target))
|
|
81
|
+
linkSources.set(target, new Set());
|
|
82
|
+
linkSources.get(target).add(finalKey);
|
|
83
|
+
if (!looksLikeHtml(target) || !accepts(target))
|
|
84
|
+
continue;
|
|
85
|
+
if (discovered.has(target))
|
|
86
|
+
continue;
|
|
87
|
+
discovered.add(target);
|
|
88
|
+
if (item.depth + 1 > options.maxDepth)
|
|
89
|
+
continue;
|
|
90
|
+
if (visited.size >= options.maxPages) {
|
|
91
|
+
limitSkips++;
|
|
92
|
+
continue;
|
|
93
|
+
}
|
|
94
|
+
visited.add(target);
|
|
95
|
+
queue.push({ url: target, depth: item.depth + 1 });
|
|
96
|
+
}
|
|
97
|
+
};
|
|
98
|
+
await runQueue(queue, options.concurrency, options.delayMs, processItem);
|
|
99
|
+
/* Check internal link targets that were discovered but not crawled (limits, depth, non-HTML). */
|
|
100
|
+
progress({ phase: 'links', message: 'Checking internal link targets' });
|
|
101
|
+
const toCheck = [...linkSources.keys()].filter((u) => !linkChecks.has(u) && allows(u)).slice(0, options.maxLinkChecks);
|
|
102
|
+
const sitemapExtra = sitemaps.urls
|
|
103
|
+
.map((u) => normalizeUrl(u) ?? u)
|
|
104
|
+
.filter((u) => isSameSite(u, startHost, options.includeSubdomains) && !linkChecks.has(u) && !toCheck.includes(u) && allows(u))
|
|
105
|
+
.slice(0, Math.max(0, options.maxLinkChecks - toCheck.length));
|
|
106
|
+
for (const u of [...linkSources.keys()].filter((u) => !allows(u)))
|
|
107
|
+
if (!blocked.includes(u))
|
|
108
|
+
blocked.push(u);
|
|
109
|
+
await mapConcurrent([...toCheck, ...sitemapExtra], options.concurrency, async (u) => {
|
|
110
|
+
const res = await probeUrl(u, options.fetch);
|
|
111
|
+
linkChecks.set(u, { url: u, status: res.status, finalUrl: res.finalUrl, redirectChain: res.redirectChain, error: res.error });
|
|
112
|
+
if (options.delayMs)
|
|
113
|
+
await sleep(options.delayMs);
|
|
114
|
+
});
|
|
115
|
+
let externalChecks;
|
|
116
|
+
if (options.checkExternalLinks) {
|
|
117
|
+
progress({ phase: 'links', message: `Checking ${Math.min(externalSources.size, options.maxLinkChecks)} external links` });
|
|
118
|
+
externalChecks = new Map();
|
|
119
|
+
const targets = [...externalSources.keys()].slice(0, options.maxLinkChecks);
|
|
120
|
+
await mapConcurrent(targets, Math.min(options.concurrency, 4), async (u) => {
|
|
121
|
+
const res = await probeUrl(u, { ...options.fetch, timeoutMs: Math.min(options.fetch.timeoutMs ?? 15_000, 10_000) });
|
|
122
|
+
externalChecks.set(u, { url: u, status: res.status, finalUrl: res.finalUrl, redirectChain: res.redirectChain, error: res.error });
|
|
123
|
+
});
|
|
124
|
+
}
|
|
125
|
+
progress({ phase: 'probes', message: 'Probing soft-404 and host redirects' });
|
|
126
|
+
const probes = await runProbes(startUrl, options.fetch);
|
|
127
|
+
const stats = {
|
|
128
|
+
pagesCrawled: records.length,
|
|
129
|
+
pagesDiscovered: discovered.size,
|
|
130
|
+
linksChecked: linkChecks.size + (externalChecks?.size ?? 0),
|
|
131
|
+
maxPages: options.maxPages,
|
|
132
|
+
maxDepth: options.maxDepth,
|
|
133
|
+
hitPageLimit: limitSkips > 0,
|
|
134
|
+
durationMs: Math.round(performance.now() - started),
|
|
135
|
+
respectedRobots: options.respectRobots,
|
|
136
|
+
};
|
|
137
|
+
return {
|
|
138
|
+
startUrl,
|
|
139
|
+
records,
|
|
140
|
+
linkChecks,
|
|
141
|
+
linkSources,
|
|
142
|
+
externalChecks,
|
|
143
|
+
externalSources: options.checkExternalLinks ? externalSources : undefined,
|
|
144
|
+
robots: robots.info,
|
|
145
|
+
robotsAllows: robots.info.found ? robots.allows : null,
|
|
146
|
+
blockedByRobots: blocked,
|
|
147
|
+
sitemaps,
|
|
148
|
+
probes,
|
|
149
|
+
stats,
|
|
150
|
+
};
|
|
151
|
+
}
|
|
152
|
+
async function runQueue(queue, concurrency, delayMs, work) {
|
|
153
|
+
let active = 0;
|
|
154
|
+
await new Promise((resolve) => {
|
|
155
|
+
const pump = () => {
|
|
156
|
+
while (active < Math.max(1, concurrency) && queue.length) {
|
|
157
|
+
const item = queue.shift();
|
|
158
|
+
active++;
|
|
159
|
+
work(item)
|
|
160
|
+
.catch(() => undefined)
|
|
161
|
+
.finally(() => {
|
|
162
|
+
active--;
|
|
163
|
+
if (delayMs > 0)
|
|
164
|
+
setTimeout(pump, delayMs);
|
|
165
|
+
else
|
|
166
|
+
pump();
|
|
167
|
+
});
|
|
168
|
+
}
|
|
169
|
+
if (active === 0 && queue.length === 0)
|
|
170
|
+
resolve();
|
|
171
|
+
};
|
|
172
|
+
pump();
|
|
173
|
+
});
|
|
174
|
+
}
|
|
175
|
+
async function runProbes(startUrl, fetchOpts) {
|
|
176
|
+
const probes = {};
|
|
177
|
+
const url = new URL(startUrl);
|
|
178
|
+
const soft404Url = new URL(`/agentic-seo-probe-${Date.now().toString(36)}-${Math.random().toString(36).slice(2, 8)}`, url.origin).href;
|
|
179
|
+
const soft = await probeUrl(soft404Url, fetchOpts);
|
|
180
|
+
probes.soft404 = soft.error ? null : { url: soft404Url, status: soft.status };
|
|
181
|
+
if (isLocalHost(startUrl))
|
|
182
|
+
return probes;
|
|
183
|
+
if (url.protocol === 'https:') {
|
|
184
|
+
const httpUrl = `http://${url.host}/`;
|
|
185
|
+
const res = await probeUrl(httpUrl, { ...fetchOpts, timeoutMs: 10_000 });
|
|
186
|
+
probes.httpRedirect = res.error ? null : { url: httpUrl, status: res.status, finalUrl: res.finalUrl, redirectsToHttps: res.finalUrl.startsWith('https://') };
|
|
187
|
+
}
|
|
188
|
+
const host = url.hostname;
|
|
189
|
+
const variantHost = host.startsWith('www.') ? stripWww(host) : `www.${host}`;
|
|
190
|
+
const variantUrl = `${url.protocol}//${variantHost}/`;
|
|
191
|
+
const res = await probeUrl(variantUrl, { ...fetchOpts, timeoutMs: 10_000 });
|
|
192
|
+
probes.hostVariant = res.error ? null : { url: variantUrl, status: res.status, finalUrl: res.finalUrl, redirectsToMain: (hostOf(res.finalUrl) ?? '') === host };
|
|
193
|
+
return probes;
|
|
194
|
+
}
|
|
195
|
+
//# sourceMappingURL=crawl.js.map
|