@crawlee/utils 3.17.1-beta.63 → 3.17.1-beta.65
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/internals/sitemap.js +1 -1
- package/package.json +3 -3
package/internals/sitemap.js
CHANGED
|
@@ -448,7 +448,7 @@ async function* discoverValidSitemaps(urls, options = {}) {
|
|
|
448
448
|
catch (err) {
|
|
449
449
|
log_1.default.warning(`Failed to fetch robots.txt file for ${hostname}`, { error: err });
|
|
450
450
|
}
|
|
451
|
-
const sitemapUrl = domainUrls.find((url) => /sitemap
|
|
451
|
+
const sitemapUrl = domainUrls.find((url) => /sitemap(?:_index)?\.(?:xml|txt)(?:\.gz)?$/i.test(url));
|
|
452
452
|
if (sitemapUrl !== undefined) {
|
|
453
453
|
if (addSitemapUrl(sitemapUrl)) {
|
|
454
454
|
yield sitemapUrl;
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@crawlee/utils",
|
|
3
|
-
"version": "3.17.1-beta.
|
|
3
|
+
"version": "3.17.1-beta.65",
|
|
4
4
|
"description": "A set of shared utilities that can be used by crawlers",
|
|
5
5
|
"engines": {
|
|
6
6
|
"node": ">=16.0.0"
|
|
@@ -49,7 +49,7 @@
|
|
|
49
49
|
"dependencies": {
|
|
50
50
|
"@apify/log": "^2.4.0",
|
|
51
51
|
"@apify/ps-tree": "^1.2.0",
|
|
52
|
-
"@crawlee/types": "3.17.1-beta.
|
|
52
|
+
"@crawlee/types": "3.17.1-beta.65",
|
|
53
53
|
"@types/sax": "^1.2.7",
|
|
54
54
|
"cheerio": "1.0.0-rc.12",
|
|
55
55
|
"file-type": "^21.3.1",
|
|
@@ -71,5 +71,5 @@
|
|
|
71
71
|
}
|
|
72
72
|
}
|
|
73
73
|
},
|
|
74
|
-
"gitHead": "
|
|
74
|
+
"gitHead": "5f275abc2a1e8234df28bddffd12c162c211234e"
|
|
75
75
|
}
|