@mintlify/scraping 4.0.36 → 4.0.37

package/bin/scrapingPipeline/tabs.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"tabs.js","sourceRoot":"","sources":["../../src/scrapingPipeline/tabs.ts"],"names":[],"mappings":"AAGA,OAAO,EAAE,gBAAgB,EAAE,MAAM,2BAA2B,CAAC;AAE7D,OAAO,EAAE,eAAe,EAAE,SAAS,EAAE,MAAM,6BAA6B,CAAC;AACzE,OAAO,EAAE,eAAe,EAAE,MAAM,oBAAoB,CAAC;AACrD,OAAO,EAAE,GAAG,EAAE,MAAM,iBAAiB,CAAC;AACtC,OAAO,EAAE,aAAa,EAAE,cAAc,EAAE,MAAM,qBAAqB,CAAC;AACpE,OAAO,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAC;AACrD,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAC3C,OAAO,EAAE,aAAa,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,UAAU,EAAE,MAAM,WAAW,CAAC;AACvC,OAAO,EAAE,aAAa,EAAE,MAAM,YAAY,CAAC;AAE3C,MAAM,CAAC,KAAK,UAAU,iBAAiB,CACrC,IAAY,EACZ,GAAiB;IAEjB,MAAM,IAAI,GAAG,UAAU,CAAC,IAAI,CAAC,CAAC;IAC9B,GAAG,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;IAEnB,eAAe,CAAC,IAAI,CAAC,CAAC;IAEtB,IAAI,SAAS,CAAC,MAAM,KAAK,QAAQ,IAAI,SAAS,CAAC,MAAM,KAAK,YAAY,EAAE,CAAC;QACvE,MAAM,KAAK,GAAG,gBAAgB,CAAC,IAAI,CAAC,CAAC;QACrC,IACE,CAAC,KAAK;YACN,CAAC,KAAK,CAAC,MAAM;YACb,CAAC,KAAK,CAAC,MAAM,KAAK,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,CAAC,CAAC,CAAC,GAAG,KAAK,GAAG,CAAC,QAAQ,CAAC;YAEjE,OAAO,UAAU,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;QAEzC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,CAAC,QAAQ,CAAC,UAAU,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;YAC7D,KAAK,CAAC,IAAI,CAAC;gBACT,IAAI,EAAE,gBAAgB,CAAC,GAAG,CAAC,QAAQ,CAAC;gBACpC,GAAG,EAAE,GAAG,CAAC,QAAQ;aAClB,CAAC,CAAC;QACL,CAAC;QAED,MAAM,OAAO,GAAG,MAAM,OAAO,CAAC,GAAG,CAC/B,KAAK,CAAC,GAAG,CAAC,KAAK,EAAE,QAAQ,EAAE,EAAE;YAC3B,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;YAC5B,MAAM,CAAC,QAAQ,GAAG,QAAQ,CAAC,GAAG,CAAC;YAC/B,IAAI,CAAC;gBACH,MAAM,OAAO,GAAG,MAAM,aAAa,CAAC,MAAM,EAAE,SAAS,CAAC,CAAC;gBACvD,OAAO,MAAM,UAAU,CAAC,OAAO,EAAE,MAAM,EAAE,EAAE,IAAI,EAAE,CAAC,QAAQ,CAAC,EAAE,CAAC,CAAC;YACjE,CAAC;YAAC,OAAO,KAAK,EAAE,CAAC;gBACf,OAAO,EAAE,OAAO,EAAE,KAAc,EAAE,OAAO,EAAE,eAAe,CAAC,KAAK,CAAC,EAAE,CAAC;YACtE,CAAC;QACH,CAAC,CAAC,CACH,CAAC;QAEF,MAAM,WAAW,GAA2B,EAAE,CAAC;QAC/C,MAAM,IAAI,GAAe,EAAE,CAAC;QAC5B,IAAI,OAAO,GAAG,cAAc,CAAC;QAC7B,IAAI,MAAM,GAAW,aAAa,CAAC;QAEnC,MAAM,SAAS,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC7D,SAAS,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE;YAC3B,IAAI,CAAC,MAAM,CAAC,IAAI;gBAAE,OAAO;YACzB,WAAW,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,UAAU,CAAC,CAAC;YAC5C,IAAI,MAAM,CAAC,IAAI,CAAC,IAAI;gBAAE,IAAI,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YACrD,IAAI,MAAM,CAAC,IAAI,CAAC,OAAO,KAAK,cAAc;gBAAE,OAAO,GAAG,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC;YAC1E,IAAI,MAAM,CAAC,IAAI,CAAC,MAAM,KAAK,aAAa;gBAAE,MAAM,GAAG,MAAM,CAAC,IAAI,CAAC,MAAM,CAAC;QACxE,CAAC,CAAC,CAAC;QAEH,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,MAAM,EAAE,EAAE,CAAC,CAAC,MAAM,CAAC,OAAO,CAAC,CAAC;QAC7D,QAAQ,CAAC,OAAO,CAAC,CAAC,MAAM,EAAE,EAAE;YAC1B,GAAG,CAAC,sBAAsB,GAAG,MAAM,CAAC,OAAO,CAAC,CAAC;QAC/C,CAAC,CAAC,CAAC;QAEH,MAAM,YAAY,GAAG,SAAS,CAAC,MAAM,KAAK,QAAQ,CAAC;QACnD,MAAM,OAAO,GAAG,YAAY,CAAC,CAAC,CAAC,MAAM,cAAc,EAAE,CAAC,CAAC,CAAC,SAAS,CAAC;QAClE,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QAC/C,MAAM,IAAI,GAAG,MAAM,aAAa,CAAC,IAAI,CAAC,CAAC;QAEvC,OAAO;YACL,OAAO,EAAE,IAAI;YACb,IAAI,EAAE;gBACJ,OAAO,EAAE,kCAAkC;gBAC3C,IAAI;gBACJ,IAAI;gBACJ,MAAM;gBACN,OAAO;gBACP,UAAU,EAAE,WAAyB;gBACrC,IAAI;aACL;SACF,CAAC;IACJ,CAAC;IAED,OAAO,UAAU,CAAC,IAAI,EAAE,GAAG,EAAE,EAAE,IAAI,EAAE,CAAC,CAAC;AACzC,CAAC"}

package/bin/scrapingPipeline/title.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"title.js","sourceRoot":"","sources":["../../src/scrapingPipeline/title.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,IAAI,EAAE,KAAK,EAAE,MAAM,kBAAkB,CAAC;AAEzD,MAAM,YAAY,GAAG,iBAAiB,CAAC;AAEvC,MAAM,CAAC,KAAK,UAAU,aAAa,CAAC,IAAc;IAChD,IAAI,IAAI,GAAuB,SAA+B,CAAC;IAE/D,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,UAAU,IAAI;QACnC,IAAI,IAAI,CAAC,OAAO,KAAK,OAAO;YAAE,OAAO,QAAQ,CAAC;QAE9C,KAAK,CAAC,IAAI,EAAE,MAAM,EAAE,UAAU,OAAO;YACnC,IAAI,GAAG,OAAO,CAAC,KAAK,CAAC;YACrB,OAAO,IAAI,CAAC;QACd,CAAC,CAAC,CAAC;QAEH,IAAI,IAAI,EAAE,CAAC;YACT,OAAO,IAAI,CAAC;QACd,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,IAAI;QAAE,OAAO,YAAY,CAAC;IAE/B,MAAM,KAAK,GAAG,IAAc,CAAC;IAC7B,IAAI,cAAc,GAAG,EAAE,CAAC;IAExB,IAAI,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QACxB,cAAc,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAY,CAAC;IACpE,CAAC;SAAM,IAAI,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/B,cAAc,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAY,CAAC;IACpE,CAAC;SAAM,IAAI,KAAK,CAAC,QAAQ,CAAC,GAAG,CAAC,EAAE,CAAC;QAC/B,cAAc,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,CAAC,IAAI,EAAE,CAAC,CAAC,IAAI,EAAY,CAAC;IACpE,CAAC;SAAM,CAAC;QACN,cAAc,GAAG,KAAK,CAAC,IAAI,EAAE,CAAC;IAChC,CAAC;IAED,OAAO,cAAc,CAAC,CAAC,CAAC,cAAc,CAAC,CAAC,CAAC,YAAY,CAAC;AACxD,CAAC"}

package/bin/tabs/retrieveReadme.js.map

@@ -1 +0,0 @@
-{"version":3,"file":"retrieveReadme.js","sourceRoot":"","sources":["../../src/tabs/retrieveReadme.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,KAAK,EAAE,IAAI,EAAE,QAAQ,EAAE,MAAM,kBAAkB,CAAC;AAEzD,OAAO,EAAE,SAAS,EAAE,MAAM,6BAA6B,CAAC;AACxD,OAAO,EAAE,SAAS,EAAE,gBAAgB,EAAE,MAAM,mBAAmB,CAAC;AAEhE,MAAM,UAAU,gBAAgB,CAAC,QAAkB;IACjD,IAAI,SAAS,CAAC,MAAM,KAAK,QAAQ,IAAI,SAAS,CAAC,MAAM,KAAK,YAAY;QAAE,OAAO,SAAS,CAAC;IAEzF,IAAI,OAAO,GAAwB,SAAgC,CAAC;IACpE,KAAK,CAAC,QAAQ,EAAE,SAAS,EAAE,UAAU,IAAI;QACvC,IAAI,SAAS,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YAClC,IACE,IAAI,CAAC,OAAO,KAAK,QAAQ;gBACzB,IAAI,CAAC,UAAU,CAAC,SAAS;gBACzB,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC;gBACxC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,QAAQ,CAAC,WAAW,CAAC,EAC/C,CAAC;gBACD,OAAO,GAAG,IAAI,CAAC;gBACf,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;QAED,IAAI,SAAS,CAAC,MAAM,KAAK,YAAY,EAAE,CAAC;YACtC,IACE,IAAI,CAAC,OAAO,KAAK,KAAK;gBACtB,IAAI,CAAC,UAAU,CAAC,SAAS;gBACzB,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC;gBACxC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,QAAQ,CAAC,QAAQ,CAAC,EAC5C,CAAC;gBACD,OAAO,GAAG,IAAI,CAAC;gBACf,OAAO,IAAI,CAAC;YACd,CAAC;QACH,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,IAAI,CAAC,OAAO;QAAE,OAAO,SAAS,CAAC;IAE/B,MAAM,KAAK,GAAe,EAAE,CAAC;IAC7B,KAAK,CAAC,OAAkB,EAAE,SAAS,EAAE,UAAU,IAAI;QACjD,IAAI,SAAS,CAAC,MAAM,KAAK,QAAQ,EAAE,CAAC;YAClC,IACE,IAAI,CAAC,OAAO,KAAK,KAAK;gBACtB,CAAC,CACC,IAAI,CAAC,OAAO,KAAK,KAAK;oBACtB,IAAI,CAAC,UAAU,CAAC,SAAS;oBACzB,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC;oBACxC,IAAI,CAAC,UAAU,CAAC,SAAS,CAAC,QAAQ,CAAC,iBAAiB,CAAC,CACtD;gBAED,OAAO,QAAQ,CAAC;YAElB,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,UAAU,OAAO;gBACtC,IACE,OAAO,CAAC,OAAO,KAAK,GAAG;oBACvB,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI;oBACxB,OAAO,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ;oBAC3C,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;oBAE1C,OAAO,QAAQ,CAAC;gBAClB,MAAM,KAAK,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;gBACjC,KAAK,CAAC,IAAI,CAAC;oBACT,IAAI,EAAE,KAAK,IAAI,gBAAgB,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC;oBACxD,GAAG,EAAE,OAAO,CAAC,UAAU,CAAC,IAAI;iBAC7B,CAAC,CAAC;YACL,CAAC,CAAC,CAAC;QACL,CAAC;QAED,IAAI,SAAS,CAAC,MAAM,KAAK,YAAY,EAAE,CAAC;YACtC,IAAI,IAAI,CAAC,OAAO,KAAK,KAAK;gBAAE,OAAO,QAAQ,CAAC;YAE5C,KAAK,CAAC,IAAI,EAAE,SAAS,EAAE,UAAU,OAAO,EAAE,CAAC,EAAE,MAAM;gBACjD,IACE,OAAO,CAAC,OAAO,KAAK,GAAG;oBACvB,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI;oBACxB,OAAO,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ;oBAC3C,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC;oBAC1C,CAAC,MAAM;oBACP,MAAM,CAAC,IAAI,KAAK,SAAS;oBACzB,CAAC,KAAK,CAAC,OAAO,CAAC,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC;oBAC3C,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC,MAAM,KAAK,CAAC;oBACxC,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,eAAe;oBAClD,MAAM,CAAC,UAAU,CAAC,SAAS,CAAC,QAAQ,CAAC,sBAAsB,CAAC;oBAE5D,OAAO,QAAQ,CAAC;gBAElB,MAAM,KAAK,GAAG,SAAS,CAAC,OAAO,CAAC,CAAC;gBACjC,KAAK,CAAC,IAAI,CAAC;oBACT,IAAI,EAAE,KAAK,IAAI,gBAAgB,CAAC,OAAO,CAAC,UAAU,CAAC,IAAI,CAAC;oBACxD,GAAG,EAAE,OAAO,CAAC,UAAU,CAAC,IAAI;iBAC7B,CAAC,CAAC;YACL,CAAC,CAAC,CAAC;QACL,CAAC;IACH,CAAC,CAAC,CAAC;IAEH,OAAO,KAAK,CAAC;AACf,CAAC"}

package/src/scrapingPipeline/group.ts

@@ -1,59 +0,0 @@
-import type { Browser } from 'puppeteer';
-
-import type { Result } from '../types/result.js';
-import { getErrorMessage } from '../utils/errors.js';
-import { log } from '../utils/log.js';
-import { fetchPageHtml, startPuppeteer } from '../utils/network.js';
-import { scrapePage } from './page.js';
-
-export async function scrapePageGroup(
-  navGroup: Array<URL>,
-  needsBrowser: boolean,
-  opts: {
-    externalLinks: boolean;
-    rootPaths?: Array<string>;
-  } = { externalLinks: false }
-): Promise<Array<Result<[string, string]>>> {
-  const browser: Browser | undefined = needsBrowser ? await startPuppeteer() : undefined;
-
-  try {
-    const res = await Promise.all(
-      navGroup.map(async (url, index) => {
-        try {
-          if (opts.externalLinks) {
-            const res = scrapePage(`external-link-${index}`, url, { externalLink: true });
-            return res;
-          }
-
-          let isOverviewPage = false;
-          if (url.toString().endsWith('/mintie_overview')) {
-            isOverviewPage = true;
-            url = new URL(url.toString().replace('/mintie_overview', ''));
-          }
-
-          const html = await fetchPageHtml(url, browser);
-          const res = scrapePage(html, url, {
-            externalLink: false,
-            isOverviewPage,
-            rootPath: opts.rootPaths ? opts.rootPaths[index] : undefined,
-          });
-          return res;
-        } catch (error) {
-          const errorMessage = getErrorMessage(error);
-          throw new Error(`We encountered an error when scraping ${url}${errorMessage}`);
-        }
-      })
-    );
-    return res;
-  } catch (error) {
-    const errorMessage = getErrorMessage(error);
-    log(
-      `We encountered an error when scraping the page group from ${
-        navGroup[0]?.origin ?? 'the URL provided'
-      }${errorMessage}`
-    );
-    throw error;
-  } finally {
-    if (browser) await browser.close();
-  }
-}