@xapp/arachne-web-service 1.18.5 → 1.18.6
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/lib/utils/getSiteMap.js
CHANGED
|
@@ -88,8 +88,30 @@ const fetchSitemap = (url) => __awaiter(void 0, void 0, void 0, function* () {
|
|
|
88
88
|
* @returns
|
|
89
89
|
*/
|
|
90
90
|
function getSitemap(sitemapUrl) {
|
|
91
|
+
return __awaiter(this, void 0, void 0, function* () {
|
|
92
|
+
return crawlSitemap(sitemapUrl, new Set());
|
|
93
|
+
});
|
|
94
|
+
}
|
|
95
|
+
/**
|
|
96
|
+
* Recursive worker behind getSitemap().
|
|
97
|
+
*
|
|
98
|
+
* `visited` carries every sitemap URL already fetched anywhere in the descent. Comparing a
|
|
99
|
+
* child only against its immediate parent is not enough: a site behind a catch-all can serve
|
|
100
|
+
* the same <sitemapindex> for every *sitemap.xml path, so its entries reference each other and
|
|
101
|
+
* the walk ping-pongs between siblings until the process is killed.
|
|
102
|
+
*
|
|
103
|
+
* @param sitemapUrl
|
|
104
|
+
* @param visited Normalized URLs already crawled in this descent.
|
|
105
|
+
* @returns
|
|
106
|
+
*/
|
|
107
|
+
function crawlSitemap(sitemapUrl, visited) {
|
|
91
108
|
return __awaiter(this, void 0, void 0, function* () {
|
|
92
109
|
const normalizedSitemapUrl = (0, arachne_1.normalizeURL)(sitemapUrl);
|
|
110
|
+
if (visited.has(normalizedSitemapUrl)) {
|
|
111
|
+
(0, stentor_logger_1.log)().debug(`Already crawled sitemap ${normalizedSitemapUrl}, skipping.`);
|
|
112
|
+
return [];
|
|
113
|
+
}
|
|
114
|
+
visited.add(normalizedSitemapUrl);
|
|
93
115
|
const xmlData = yield fetchSitemap(normalizedSitemapUrl);
|
|
94
116
|
if (!xmlData) {
|
|
95
117
|
console.error("No sitemap data retrieved.");
|
|
@@ -108,12 +130,11 @@ function getSitemap(sitemapUrl) {
|
|
|
108
130
|
const sitemapNodes = $("sitemap").toArray();
|
|
109
131
|
for (const sitemapNode of sitemapNodes) {
|
|
110
132
|
const loc = $(sitemapNode).find("loc").text();
|
|
111
|
-
|
|
112
|
-
|
|
113
|
-
if (normalizedLoc !== normalizedSitemapUrl) {
|
|
114
|
-
const urls = yield getSitemap(loc);
|
|
115
|
-
allUrls = allUrls.concat(urls);
|
|
133
|
+
if (!loc) {
|
|
134
|
+
continue;
|
|
116
135
|
}
|
|
136
|
+
const urls = yield crawlSitemap(loc, visited);
|
|
137
|
+
allUrls = allUrls.concat(urls);
|
|
117
138
|
}
|
|
118
139
|
return allUrls;
|
|
119
140
|
});
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"getSiteMap.js","sourceRoot":"","sources":["../../src/utils/getSiteMap.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA0DA,
|
|
1
|
+
{"version":3,"file":"getSiteMap.js","sourceRoot":"","sources":["../../src/utils/getSiteMap.ts"],"names":[],"mappings":";;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;AA0DA,gCAEC;AAqGD,wCAqHC;AAtRD,kCAAkC;AAClC,iDAAmC;AAEnC,2CAA6C;AAC7C,mDAAqC;AAErC,iDAOwB;AAExB,MAAM,YAAY,GAAG,CAAO,GAAW,EAAmB,EAAE;IACxD,MAAM,UAAU,GAAG,IAAI,eAAe,EAAE,CAAC;IACzC,MAAM,EAAE,MAAM,EAAE,GAAG,UAAU,CAAC;IAE9B,MAAM,OAAO,GAAG,IAAI,CAAC;IAErB,wFAAwF;IACxF,MAAM,SAAS,GAAG,UAAU,CAAC,GAAG,EAAE;QAC9B,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,WAAW,OAAO,+BAA+B,GAAG,EAAE,CAAC,CAAC;QACpE,UAAU,CAAC,KAAK,EAAE,CAAC;IACvB,CAAC,EAAE,OAAO,CAAC,CAAC;IAEZ,aAAa;IACb,IAAI,CAAC;QACD,MAAM,IAAI,GAAG,MAAM,KAAK,CAAC,GAAG,EAAE,EAAE,MAAM,EAAE,CAAC;aACpC,IAAI,CAAC,CAAC,QAAQ,EAAE,EAAE;YACf,IAAI,CAAC,QAAQ,CAAC,EAAE,EAAE,CAAC;gBACf,MAAM,IAAI,KAAK,CAAC,yCAAyC,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;YAChF,CAAC;YAED,OAAO,QAAQ,CAAC,IAAI,EAAE,CAAC;QAC3B,CAAC,CAAC;aACD,OAAO,CAAC,GAAG,EAAE;YACV,YAAY,CAAC,SAAS,CAAC,CAAC;QAC5B,CAAC,CAAC,CAAC;QAEP,OAAO,IAAI,CAAC;IAChB,CAAC;IAAC,OAAO,KAAK,EAAE,CAAC;QACb,OAAO,CAAC,KAAK,CAAC,4BAA4B,KAAK,EAAE,CAAC,CAAC;QACnD,OAAO,EAAE,CAAC;IACd,CAAC;AACL,CAAC,CAAA,CAAC;AAEF;;;;;;;;;GASG;AACH,SAAsB,UAAU,CAAC,UAAkB;;QAC/C,OAAO,YAAY,CAAC,UAAU,EAAE,IAAI,GAAG,EAAU,CAAC,CAAC;IACvD,CAAC;CAAA;AAED;;;;;;;;;;;GAWG;AACH,SAAe,YAAY,CAAC,UAAkB,EAAE,OAAoB;;QAChE,MAAM,oBAAoB,GAAG,IAAA,sBAAY,EAAC,UAAU,CAAC,CAAC;QAEtD,IAAI,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,EAAE,CAAC;YACpC,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,2BAA2B,oBAAoB,aAAa,CAAC,CAAC;YAC1E,OAAO,EAAE,CAAC;QACd,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,oBAAoB,CAAC,CAAC;QAElC,MAAM,OAAO,GAAG,MAAM,YAAY,CAAC,oBAAoB,CAAC,CAAC;QAEzD,IAAI,CAAC,OAAO,EAAE,CAAC;YACX,OAAO,CAAC,KAAK,CAAC,4BAA4B,CAAC,CAAC;YAC5C,OAAO,EAAE,CAAC;QACd,CAAC;QAED,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,OAAO,EAAE;YAC5B,OAAO,EAAE,IAAI;SAChB,CAAC,CAAC;QAEH,IAAI,OAAO,GAAa,EAAE,CAAC;QAE3B,mBAAmB;QACnB,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,EAAE;YACvB,MAAM,GAAG,GAAG,CAAC,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;YACxC,OAAO,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;QACtB,CAAC,CAAC,CAAC;QAEH,yCAAyC;QACzC,MAAM,YAAY,GAAG,CAAC,CAAC,SAAS,CAAC,CAAC,OAAO,EAAE,CAAC;QAE5C,KAAK,MAAM,WAAW,IAAI,YAAY,EAAE,CAAC;YACrC,MAAM,GAAG,GAAG,CAAC,CAAC,WAAW,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC,IAAI,EAAE,CAAC;YAE9C,IAAI,CAAC,GAAG,EAAE,CAAC;gBACP,SAAS;YACb,CAAC;YAED,MAAM,IAAI,GAAG,MAAM,YAAY,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;YAC9C,OAAO,GAAG,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACnC,CAAC;QAED,OAAO,OAAO,CAAC;IACnB,CAAC;CAAA;AAqCD;;;;;GAKG;AACH,SAAsB,cAAc,CAAC,GAAW;;QAC5C,IAAI,OAAY,CAAC;QACjB,IAAI,YAAoB,CAAC;QACzB,IAAI,UAAe,CAAC;QAEpB,IAAI,CAAC;YACD,OAAO,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;YACvB,YAAY,GAAG,OAAO,CAAC,QAAQ,CAAC;YAEhC,UAAU,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;YAC1B,UAAU,CAAC,QAAQ,GAAG,GAAG,YAAY,aAAa,CAAC;QACvD,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACT,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,oCAAoC,GAAG,GAAG,CAAC,CAAC;YACxD,OAAO,SAAS,CAAC;QACrB,CAAC;QAED,IAAI,CAAC,UAAU,EAAE,CAAC;YACd,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,oCAAoC,GAAG,GAAG,CAAC,CAAC;YACxD,OAAO,SAAS,CAAC;QACrB,CAAC;QAED,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,sBAAsB,UAAU,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;QAE3D,MAAM,OAAO,GAAG,MAAM,UAAU,CAAC,UAAU,CAAC,QAAQ,EAAE,CAAC,CAAC;QAExD,QAAQ;QACR,MAAM,OAAO,GAAG,OAAO,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE;YAClC,IAAI,KAAK,CAAC,UAAU,CAAC,GAAG,CAAC,EAAE,CAAC;gBACxB,MAAM,UAAU,GAAG,IAAI,GAAG,CAAC,GAAG,CAAC,CAAC;gBAChC,UAAU,CAAC,QAAQ,GAAG,KAAK,CAAC;gBAC5B,OAAO,UAAU,CAAC,QAAQ,EAAE,CAAC;YACjC,CAAC;iBAAM,CAAC;gBACJ,OAAO,KAAK,CAAC;YACjB,CAAC;QACL,CAAC,CAAC,CAAC;QAEH,wFAAwF;QACxF,MAAM,QAAQ,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,EAAE;YACtC,IAAI,CAAC;gBACD,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;gBAC9B,MAAM,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC;gBACjC,OAAO,QAAQ,CAAC,UAAU,CAAC,YAAY,CAAC,CAAC;YAC7C,CAAC;YAAC,OAAO,CAAC,EAAE,CAAC;gBACT,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,0BAA0B,KAAK,EAAE,CAAC,CAAC;YACnD,CAAC;YACD,OAAO,KAAK,CAAC;QACjB,CAAC,CAAC,CAAC;QAEH,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,eAAe,OAAO,CAAC,MAAM,uBAAuB,QAAQ,CAAC,MAAM,EAAE,CAAC,CAAC;QAEnF,qBAAqB;QACrB,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YACpC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,kCAAmB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACrD,CAAC,CAAC,CAAC;QAEH,IAAI,OAAO,EAAE,CAAC;YACV,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,0BAA0B,OAAO,EAAE,CAAC,CAAC;QACrD,CAAC;QAED,MAAM,SAAS,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YACtC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,oCAAqB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACvD,CAAC,CAAC,CAAC;QAEH,IAAI,SAAS,EAAE,CAAC;YACZ,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,4BAA4B,SAAS,EAAE,CAAC,CAAC;QACzD,CAAC;QAED,MAAM,OAAO,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YACpC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,kCAAmB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACrD,CAAC,CAAC,CAAC;QAEH,IAAI,OAAO,EAAE,CAAC;YACV,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,qBAAqB,OAAO,EAAE,CAAC,CAAC;QAChD,CAAC;QAED,MAAM,WAAW,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YACxC,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,sCAAuB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QACzD,CAAC,CAAC,CAAC;QAEH,IAAI,WAAW,EAAE,CAAC;YACd,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,yBAAyB,WAAW,EAAE,CAAC,CAAC;QACxD,CAAC;QAED,MAAM,aAAa,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YAC1C,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,wCAAyB,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC3D,CAAC,CAAC,CAAC;QAEH,IAAI,aAAa,EAAE,CAAC;YAChB,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,gCAAgC,aAAa,EAAE,CAAC,CAAC;QACjE,CAAC;QAED,MAAM,cAAc,GAAG,QAAQ,CAAC,IAAI,CAAC,CAAC,KAAK,EAAE,EAAE;YAC3C,MAAM,MAAM,GAAG,IAAI,GAAG,CAAC,KAAK,CAAC,CAAC;YAC9B,OAAO,0CAA2B,CAAC,IAAI,CAAC,MAAM,CAAC,QAAQ,CAAC,CAAC;QAC7D,CAAC,CAAC,CAAC;QAEH,IAAI,cAAc,EAAE,CAAC;YACjB,IAAA,oBAAG,GAAE,CAAC,KAAK,CAAC,kCAAkC,cAAc,EAAE,CAAC,CAAC;QACpE,CAAC;QAED,OAAO;YACH,UAAU,EAAE,UAAU,CAAC,QAAQ,EAAE;YAEjC,UAAU,EAAE,QAAQ,CAAC,MAAM;YAE3B,OAAO;YACP,SAAS;YACT,OAAO;YACP,WAAW;YACX,aAAa;YACb,cAAc;SACjB,CAAC;IACN,CAAC;CAAA"}
|