@dancrumb/web-crawler 1.4.0 → 3.0.0

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.
@@ -13,7 +13,9 @@ export declare class WebCrawler extends EventEmitter<{
13
13
  private uncrawled;
14
14
  private documents;
15
15
  private roots;
16
+ private supportedTypes;
16
17
  private maxDepth;
18
+ constructor();
17
19
  private reportProgress;
18
20
  private addUncrawled;
19
21
  private markCrawled;
@@ -28,6 +30,8 @@ export declare class WebCrawler extends EventEmitter<{
28
30
  crawlLink(initialLink: string, maxDepth?: number): Promise<string[][]>;
29
31
  clear(): void;
30
32
  addRoot(root: string): Promise<string>;
33
+ addSupportedType(supportedType: `${string}/${string}`): void;
34
+ setSupportedTypes(...supportedTypes: `${string}/${string}`[]): void;
31
35
  /**
32
36
  * Sets the maximum crawl depth.
33
37
  *
@@ -1 +1 @@
1
- {"version":3,"file":"WebCrawler.d.ts","sourceRoot":"","sources":["../src/WebCrawler.ts"],"names":[],"mappings":"AAIA,OAAO,YAAY,MAAM,aAAa,CAAC;AAMvC,KAAK,QAAQ,GAAG;IACd,GAAG,EAAE,MAAM,CAAC;IACZ,YAAY,EAAE,MAAM,CAAC;IACrB,OAAO,EAAE,MAAM,GAAG,MAAM,CAAC;IACzB,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;CACtB,CAAC;AAEF,qBAAa,UAAW,SAAQ,YAAY,CAAC;IAC3C,QAAQ,EAAE,CAAC,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC,CAAC;CACrC,CAAC;IACA,OAAO,CAAC,OAAO,CAA6B;IAC5C,OAAO,CAAC,SAAS,CAA6B;IAC9C,OAAO,CAAC,SAAS,CAA+B;IAChD,OAAO,CAAC,KAAK,CAAqB;IAClC,OAAO,CAAC,QAAQ,CAAa;IAE7B,OAAO,CAAC,cAAc;IAWtB,OAAO,CAAC,YAAY;IAKpB,OAAO,CAAC,WAAW;YAUL,YAAY;YA0BZ,WAAW;IAOzB,OAAO,CAAC,WAAW;YAIL,WAAW;YAoBX,YAAY;YAkBZ,WAAW;YAUX,SAAS;IAiCvB,OAAO,CAAC,cAAc;IAYhB,SAAS,CAAC,WAAW,EAAE,MAAM,EAAE,QAAQ,GAAE,MAAsB;IAwDrE,KAAK;IAOC,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IAM5C;;;;;;;OAOG;IACH,WAAW,CAAC,KAAK,EAAE,MAAM;IAWzB,cAAc;IAId,YAAY,IAAI,QAAQ,EAAE;CAK3B"}
1
+ {"version":3,"file":"WebCrawler.d.ts","sourceRoot":"","sources":["../src/WebCrawler.ts"],"names":[],"mappings":"AAIA,OAAO,YAAY,MAAM,aAAa,CAAC;AAMvC,KAAK,QAAQ,GAAG;IACd,GAAG,EAAE,MAAM,CAAC;IACZ,YAAY,EAAE,MAAM,CAAC;IACrB,OAAO,EAAE,MAAM,GAAG,MAAM,CAAC;IACzB,WAAW,EAAE,MAAM,CAAC;IACpB,KAAK,EAAE,MAAM,GAAG,IAAI,CAAC;CACtB,CAAC;AAEF,qBAAa,UAAW,SAAQ,YAAY,CAAC;IAC3C,QAAQ,EAAE,CAAC,QAAQ,EAAE,MAAM,GAAG,IAAI,CAAC,CAAC;CACrC,CAAC;IACA,OAAO,CAAC,OAAO,CAA6B;IAC5C,OAAO,CAAC,SAAS,CAA6B;IAC9C,OAAO,CAAC,SAAS,CAA+B;IAChD,OAAO,CAAC,KAAK,CAAqB;IAClC,OAAO,CAAC,cAAc,CAAoC;IAC1D,OAAO,CAAC,QAAQ,CAAa;;IAO7B,OAAO,CAAC,cAAc;IAWtB,OAAO,CAAC,YAAY;IAKpB,OAAO,CAAC,WAAW;YAUL,YAAY;YA0BZ,WAAW;IAOzB,OAAO,CAAC,WAAW;YAIL,WAAW;YAoBX,YAAY;YAkBZ,WAAW;YAUX,SAAS;IAiCvB,OAAO,CAAC,cAAc;IAYhB,SAAS,CAAC,WAAW,EAAE,MAAM,EAAE,QAAQ,GAAE,MAAsB;IA6DrE,KAAK;IASC,OAAO,CAAC,IAAI,EAAE,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;IAM5C,gBAAgB,CAAC,aAAa,EAAE,GAAG,MAAM,IAAI,MAAM,EAAE;IAIrD,iBAAiB,CAAC,GAAG,cAAc,EAAE,GAAG,MAAM,IAAI,MAAM,EAAE,EAAE;IAK5D;;;;;;;OAOG;IACH,WAAW,CAAC,KAAK,EAAE,MAAM;IAWzB,cAAc;IAId,YAAY,IAAI,QAAQ,EAAE;CAK3B"}
@@ -10,7 +10,12 @@ export class WebCrawler extends EventEmitter {
10
10
  uncrawled = new Map();
11
11
  documents = new Map();
12
12
  roots = new Set();
13
+ supportedTypes = new Set();
13
14
  maxDepth = 0;
15
+ constructor() {
16
+ super();
17
+ this.clear();
18
+ }
14
19
  reportProgress() {
15
20
  if (this.uncrawled.size === 0) {
16
21
  this.emit("progress", null);
@@ -174,7 +179,8 @@ export class WebCrawler extends EventEmitter {
174
179
  continue;
175
180
  }
176
181
  const contentType = this.getContentType(head);
177
- if (contentType === null || !contentType.startsWith("text/html")) {
182
+ if (contentType === null ||
183
+ ![...this.supportedTypes.values()].some((supportedType) => contentType.startsWith(supportedType))) {
178
184
  logDebug(`Page at ${resolved} has MIME type ${contentType}, so skipping`);
179
185
  continue;
180
186
  }
@@ -192,12 +198,21 @@ export class WebCrawler extends EventEmitter {
192
198
  this.uncrawled.clear();
193
199
  this.documents.clear();
194
200
  this.roots.clear();
201
+ this.supportedTypes.clear();
202
+ this.supportedTypes.add("text/html");
195
203
  }
196
204
  async addRoot(root) {
197
205
  const canonical = await this.getCanonical(root);
198
206
  this.roots.add(canonical);
199
207
  return canonical;
200
208
  }
209
+ addSupportedType(supportedType) {
210
+ this.supportedTypes.add(supportedType);
211
+ }
212
+ setSupportedTypes(...supportedTypes) {
213
+ this.supportedTypes.clear();
214
+ supportedTypes.forEach((s) => this.supportedTypes.add(s));
215
+ }
201
216
  /**
202
217
  * Sets the maximum crawl depth.
203
218
  *
@@ -1 +1 @@
1
- {"version":3,"file":"WebCrawler.js","sourceRoot":"","sources":["../src/WebCrawler.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AACnC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,MAAM,MAAM,aAAa,CAAC;AACjC,OAAO,YAAY,MAAM,aAAa,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AACzD,OAAO,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAEtC,MAAM,QAAQ,GAAG,KAAK,CAAC,cAAc,CAAC,CAAC;AAUvC,MAAM,OAAO,UAAW,SAAQ,YAE9B;IACQ,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IACpC,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,SAAS,GAAG,IAAI,GAAG,EAAoB,CAAC;IACxC,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;IAC1B,QAAQ,GAAW,CAAC,CAAC;IAErB,cAAc;QACpB,IAAI,IAAI,CAAC,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YAC9B,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC9B,CAAC;QACD,MAAM,QAAQ,GAAG,CACf,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC/D,GAAG,CACJ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QACb,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAClC,CAAC;IAEO,YAAY,CAAC,GAAW,EAAE,KAAa;QAC7C,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC/B,IAAI,CAAC,cAAc,EAAE,CAAC;IACxB,CAAC;IAEO,WAAW,CAAC,GAAW,EAAE,QAAiB;QAChD,IAAI,KAAK,GAAG,QAAQ,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAChD,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC7B,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC3B,IAAI,CAAC,cAAc,EAAE,CAAC;IACxB,CAAC;IAEO,KAAK,CAAC,YAAY,CAAC,GAAW;QACpC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC;QAC7C,MAAM,WAAW,GACf,IAAI,CAAC,cAAc,CAAC,OAAO,CAAC,IAAI,0BAA0B,CAAC;QAC7D,MAAM,OAAO,GAAG,IAAI,CAAC;QACrB,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE;YACtB,GAAG;YACH,YAAY,EAAE,EAAE;YAChB,OAAO;YACP,KAAK,EAAE,IAAI;YACX,WAAW;SACZ,CAAC,CAAC;QACH,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;QAClD,IAAI,YAAY,KAAK,GAAG,EAAE,CAAC;YACzB,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAC3B,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,YAAY,EAAE;gBAC/B,GAAG;gBACH,YAAY;gBACZ,OAAO;gBACP,KAAK,EAAE,IAAI;gBACX,WAAW;aACZ,CAAC,CAAC;QACL,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,GAAW;QACnC,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAC5B,OAAO,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC,OAAO,CAAC;QAC1C,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;IAChC,CAAC;IAEO,WAAW,CAAC,GAAW;QAC7B,OAAO,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;IACvE,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,GAAW;QACnC,IAAI,CAAC;YACH,QAAQ,CAAC,sBAAsB,GAAG,EAAE,CAAC,CAAC;YAEtC,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC;YACpC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAE7B,MAAM,WAAW,GAAa,EAAE,CAAC;YAEjC,KAAK,MAAM,MAAM,IAAI,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5B,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;YACjD,CAAC;YAED,OAAO,WAAW,CAAC;QACrB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,QAAS,CAAW,CAAC,OAAO,6BAA6B,CAAC,CAAC;YACpE,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,YAAY,CAAC,GAAW;QACpC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACzC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,KAAK,MAAM,IAAI,IAAI,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC7B,IAAI,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,WAAW,EAAE,CAAC;oBACxC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,CAAE,CAAC;gBAC/B,CAAC;YACH,CAAC;YACD,OAAO,GAAG,CAAC;QACb,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CACN,QAAS,CAAW,CAAC,OAAO,qCAAqC,CAClE,CAAC;YACF,OAAO,GAAG,CAAC;QACb,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,IAAY,EAAE,IAAY;QAClD,IAAI,QAAQ,GAAG,IAAI,CAAC;QACpB,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YAC5B,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;aAAM,CAAC;YACN,QAAQ,GAAG,IAAI,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACzD,CAAC;QACD,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IAC3C,CAAC;IAEO,KAAK,CAAC,SAAS,CACrB,IAAY;QAEZ,IAAI,OAAY,CAAC;QACjB,IAAI,CAAC;YACH,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;QAC1B,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,GAAG,IAAI,0BAA0B,CAAC,CAAC;YAC5C,OAAO,IAAI,CAAC;QACd,CAAC;QAED,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YACzC,QAAQ,CAAC,cAAc,OAAO,CAAC,QAAQ,8BAA8B,CAAC,CAAC;YACvE,OAAO,IAAI,CAAC;QACd,CAAC;QAED,QAAQ,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC;QACzB,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,CAAC;YAEtC,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;gBAC5B,OAAO,QAAQ,CAAC,OAAO,CAAC;YAC1B,CAAC;YACD,QAAQ,CACN,OAAO,QAAQ,CAAC,MAAM,KAAK,QAAQ,CAAC,UAAU,oBAAoB,IAAI,EAAE,CACzE,CAAC;QACJ,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,OAAQ,CAAW,CAAC,OAAO,oBAAoB,IAAI,EAAE,CAAC,CAAC;QAClE,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;IAEO,cAAc,CAAC,OAAiC;QACtD,MAAM,WAAW,GACf,OAAO,CAAC,cAAc,CAAC,IAAI,OAAO,CAAC,cAAc,CAAC,IAAI,IAAI,CAAC;QAC7D,IAAI,WAAW,KAAK,IAAI,EAAE,CAAC;YACzB,OAAO,WAAW,CAAC;QACrB,CAAC;QACD,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;YACpC,OAAO,IAAI,CAAC;QACd,CAAC;QACD,OAAO,WAAW,CAAC;IACrB,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,WAAmB,EAAE,WAAmB,IAAI,CAAC,QAAQ;QACnE,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC;QAClD,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACnB,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;QAEzB,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,GAA8C,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAExE,OAAO,GAAG,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;YAC5B,QAAQ,CAAC,eAAe,CAAC,CAAC;YAC1B,QAAQ,CACN,YAAY,IAAI,CAAC,OAAO,CAAC,IAAI,gBAAgB,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CACnE,CAAC;YACF,QAAQ,CAAC,QAAQ,GAAG,WAAW,KAAK,IAAI,CAAC,CAAC;YAC1C,IAAI,KAAK,IAAI,QAAQ,EAAE,CAAC;gBACtB,IAAI,CAAC,WAAW,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;gBAC7B,QAAQ,CAAC,YAAY,GAAG,KAAK,CAAC,CAAC;gBAC/B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;gBAC1C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBACzB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;oBACnD,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,QAAQ,CAAC,EAAE,CAAC;wBAChC,QAAQ,CAAC,iBAAiB,QAAQ,KAAK,IAAI,GAAG,CAAC,CAAC;wBAChD,SAAS;oBACX,CAAC;oBACD,QAAQ,CACN,GAAG,QAAQ,oBAAoB,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CACrE,CAAC;oBACF,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,QAAQ,CAAC,GAAG,KAAK,EAAE,CAAC;wBACrD,QAAQ,CACN,mBAAmB,QAAQ,aAAa,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CACrE,CAAC;wBACF,SAAS;oBACX,CAAC;oBACD,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;wBAClC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;wBAC5C,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;4BAClB,SAAS;wBACX,CAAC;wBACD,MAAM,WAAW,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;wBAC9C,IAAI,WAAW,KAAK,IAAI,IAAI,CAAC,WAAW,CAAC,UAAU,CAAC,WAAW,CAAC,EAAE,CAAC;4BACjE,QAAQ,CACN,WAAW,QAAQ,kBAAkB,WAAW,eAAe,CAChE,CAAC;4BACF,SAAS;wBACX,CAAC;wBACD,QAAQ,CAAC,UAAU,QAAQ,uBAAuB,CAAC,CAAC;wBACpD,IAAI,CAAC,YAAY,CAAC,QAAQ,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;oBACzC,CAAC;gBACH,CAAC;YACH,CAAC;YACD,CAAC,GAAG,EAAE,KAAK,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACxC,CAAC;QAED,OAAO,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAChE,CAAC;IAED,KAAK;QACH,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;QACrB,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,CAAC;QACvB,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,CAAC;QACvB,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;IACrB,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,IAAY;QACxB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC1B,OAAO,SAAS,CAAC;IACnB,CAAC;IAED;;;;;;;OAOG;IACH,WAAW,CAAC,KAAa;QACvB,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC;QACpB,CAAC;aAAM,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAClC,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC;QACpB,CAAC;QACD,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED,cAAc;QACZ,OAAO,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;IAClC,CAAC;IAED,YAAY;QACV,OAAO,IAAI,CAAC,cAAc,EAAE;aACzB,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;aACrC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC,CAAC;IACpC,CAAC;CACF"}
1
+ {"version":3,"file":"WebCrawler.js","sourceRoot":"","sources":["../src/WebCrawler.ts"],"names":[],"mappings":"AACA,OAAO,KAAK,OAAO,MAAM,SAAS,CAAC;AACnC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,MAAM,MAAM,aAAa,CAAC;AACjC,OAAO,YAAY,MAAM,aAAa,CAAC;AACvC,OAAO,EAAE,OAAO,EAAE,QAAQ,EAAE,MAAM,sBAAsB,CAAC;AACzD,OAAO,EAAE,MAAM,EAAE,MAAM,cAAc,CAAC;AAEtC,MAAM,QAAQ,GAAG,KAAK,CAAC,cAAc,CAAC,CAAC;AAUvC,MAAM,OAAO,UAAW,SAAQ,YAE9B;IACQ,OAAO,GAAG,IAAI,GAAG,EAAkB,CAAC;IACpC,SAAS,GAAG,IAAI,GAAG,EAAkB,CAAC;IACtC,SAAS,GAAG,IAAI,GAAG,EAAoB,CAAC;IACxC,KAAK,GAAG,IAAI,GAAG,EAAU,CAAC;IAC1B,cAAc,GAAG,IAAI,GAAG,EAAyB,CAAC;IAClD,QAAQ,GAAW,CAAC,CAAC;IAE7B;QACE,KAAK,EAAE,CAAC;QACR,IAAI,CAAC,KAAK,EAAE,CAAC;IACf,CAAC;IAEO,cAAc;QACpB,IAAI,IAAI,CAAC,SAAS,CAAC,IAAI,KAAK,CAAC,EAAE,CAAC;YAC9B,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,IAAI,CAAC,CAAC;QAC9B,CAAC;QACD,MAAM,QAAQ,GAAG,CACf,CAAC,IAAI,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,IAAI,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;YAC/D,GAAG,CACJ,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;QACb,IAAI,CAAC,IAAI,CAAC,UAAU,EAAE,QAAQ,CAAC,CAAC;IAClC,CAAC;IAEO,YAAY,CAAC,GAAW,EAAE,KAAa;QAC7C,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC/B,IAAI,CAAC,cAAc,EAAE,CAAC;IACxB,CAAC;IAEO,WAAW,CAAC,GAAW,EAAE,QAAiB;QAChD,IAAI,KAAK,GAAG,QAAQ,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;QAChD,IAAI,KAAK,KAAK,SAAS,EAAE,CAAC;YACxB,OAAO;QACT,CAAC;QACD,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;QAC7B,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;QAC3B,IAAI,CAAC,cAAc,EAAE,CAAC;IACxB,CAAC;IAEO,KAAK,CAAC,YAAY,CAAC,GAAW;QACpC,MAAM,EAAE,IAAI,EAAE,OAAO,EAAE,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC;QAC7C,MAAM,WAAW,GACf,IAAI,CAAC,cAAc,CAAC,OAAO,CAAC,IAAI,0BAA0B,CAAC;QAC7D,MAAM,OAAO,GAAG,IAAI,CAAC;QACrB,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,EAAE;YACtB,GAAG;YACH,YAAY,EAAE,EAAE;YAChB,OAAO;YACP,KAAK,EAAE,IAAI;YACX,WAAW;SACZ,CAAC,CAAC;QACH,MAAM,YAAY,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;QAClD,IAAI,YAAY,KAAK,GAAG,EAAE,CAAC;YACzB,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAC3B,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,YAAY,EAAE;gBAC/B,GAAG;gBACH,YAAY;gBACZ,OAAO;gBACP,KAAK,EAAE,IAAI;gBACX,WAAW;aACZ,CAAC,CAAC;QACL,CAAC;QACD,OAAO,IAAI,CAAC;IACd,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,GAAW;QACnC,IAAI,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,EAAE,CAAC;YAC5B,OAAO,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAE,CAAC,OAAO,CAAC;QAC1C,CAAC;QACD,OAAO,IAAI,CAAC,YAAY,CAAC,GAAG,CAAC,CAAC;IAChC,CAAC;IAEO,WAAW,CAAC,GAAW;QAC7B,OAAO,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,GAAG,CAAC,UAAU,CAAC,IAAI,CAAC,CAAC,CAAC;IACvE,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,GAAW;QACnC,IAAI,CAAC;YACH,QAAQ,CAAC,sBAAsB,GAAG,EAAE,CAAC,CAAC;YAEtC,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,OAAO,CAAC,GAAG,CAAC,CAAC;YACpC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAE7B,MAAM,WAAW,GAAa,EAAE,CAAC;YAEjC,KAAK,MAAM,MAAM,IAAI,CAAC,CAAC,GAAG,CAAC,EAAE,CAAC;gBAC5B,WAAW,CAAC,IAAI,CAAC,MAAM,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,EAAE,CAAC,CAAC;YACjD,CAAC;YAED,OAAO,WAAW,CAAC;QACrB,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,QAAS,CAAW,CAAC,OAAO,6BAA6B,CAAC,CAAC;YACpE,OAAO,EAAE,CAAC;QACZ,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,YAAY,CAAC,GAAW;QACpC,IAAI,CAAC;YACH,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;YACzC,MAAM,CAAC,GAAG,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;YAC7B,KAAK,MAAM,IAAI,IAAI,CAAC,CAAC,MAAM,CAAC,EAAE,CAAC;gBAC7B,IAAI,IAAI,CAAC,OAAO,CAAC,KAAK,CAAC,KAAK,WAAW,EAAE,CAAC;oBACxC,OAAO,IAAI,CAAC,OAAO,CAAC,MAAM,CAAE,CAAC;gBAC/B,CAAC;YACH,CAAC;YACD,OAAO,GAAG,CAAC;QACb,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CACN,QAAS,CAAW,CAAC,OAAO,qCAAqC,CAClE,CAAC;YACF,OAAO,GAAG,CAAC;QACb,CAAC;IACH,CAAC;IAEO,KAAK,CAAC,WAAW,CAAC,IAAY,EAAE,IAAY;QAClD,IAAI,QAAQ,GAAG,IAAI,CAAC;QACpB,IAAI,IAAI,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YAC5B,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;aAAM,CAAC;YACN,QAAQ,GAAG,IAAI,GAAG,CAAC,IAAI,EAAE,IAAI,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QACzD,CAAC;QACD,OAAO,MAAM,IAAI,CAAC,YAAY,CAAC,QAAQ,CAAC,CAAC;IAC3C,CAAC;IAEO,KAAK,CAAC,SAAS,CACrB,IAAY;QAEZ,IAAI,OAAY,CAAC;QACjB,IAAI,CAAC;YACH,OAAO,GAAG,IAAI,GAAG,CAAC,IAAI,CAAC,CAAC;QAC1B,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,GAAG,IAAI,0BAA0B,CAAC,CAAC;YAC5C,OAAO,IAAI,CAAC;QACd,CAAC;QAED,IAAI,CAAC,OAAO,CAAC,QAAQ,CAAC,UAAU,CAAC,MAAM,CAAC,EAAE,CAAC;YACzC,QAAQ,CAAC,cAAc,OAAO,CAAC,QAAQ,8BAA8B,CAAC,CAAC;YACvE,OAAO,IAAI,CAAC;QACd,CAAC;QAED,QAAQ,CAAC,QAAQ,IAAI,EAAE,CAAC,CAAC;QACzB,IAAI,CAAC;YACH,MAAM,QAAQ,GAAG,MAAM,QAAQ,CAAC,IAAI,CAAC,CAAC;YAEtC,IAAI,QAAQ,CAAC,MAAM,KAAK,GAAG,EAAE,CAAC;gBAC5B,OAAO,QAAQ,CAAC,OAAO,CAAC;YAC1B,CAAC;YACD,QAAQ,CACN,OAAO,QAAQ,CAAC,MAAM,KAAK,QAAQ,CAAC,UAAU,oBAAoB,IAAI,EAAE,CACzE,CAAC;QACJ,CAAC;QAAC,OAAO,CAAC,EAAE,CAAC;YACX,QAAQ,CAAC,OAAQ,CAAW,CAAC,OAAO,oBAAoB,IAAI,EAAE,CAAC,CAAC;QAClE,CAAC;QAED,OAAO,IAAI,CAAC;IACd,CAAC;IAEO,cAAc,CAAC,OAAiC;QACtD,MAAM,WAAW,GACf,OAAO,CAAC,cAAc,CAAC,IAAI,OAAO,CAAC,cAAc,CAAC,IAAI,IAAI,CAAC;QAC7D,IAAI,WAAW,KAAK,IAAI,EAAE,CAAC;YACzB,OAAO,WAAW,CAAC;QACrB,CAAC;QACD,IAAI,OAAO,WAAW,KAAK,QAAQ,EAAE,CAAC;YACpC,OAAO,IAAI,CAAC;QACd,CAAC;QACD,OAAO,WAAW,CAAC;IACrB,CAAC;IAED,KAAK,CAAC,SAAS,CAAC,WAAmB,EAAE,WAAmB,IAAI,CAAC,QAAQ;QACnE,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC;QAClD,IAAI,CAAC,OAAO,CAAC,IAAI,CAAC,CAAC;QACnB,IAAI,CAAC,QAAQ,GAAG,QAAQ,CAAC;QAEzB,IAAI,CAAC,GAAG,EAAE,KAAK,CAAC,GAA8C,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;QAExE,OAAO,GAAG,KAAK,SAAS,EAAE,CAAC;YACzB,MAAM,CAAC,KAAK,KAAK,SAAS,CAAC,CAAC;YAC5B,QAAQ,CAAC,eAAe,CAAC,CAAC;YAC1B,QAAQ,CACN,YAAY,IAAI,CAAC,OAAO,CAAC,IAAI,gBAAgB,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CACnE,CAAC;YACF,QAAQ,CAAC,QAAQ,GAAG,WAAW,KAAK,IAAI,CAAC,CAAC;YAC1C,IAAI,KAAK,IAAI,QAAQ,EAAE,CAAC;gBACtB,IAAI,CAAC,WAAW,CAAC,GAAG,EAAE,KAAK,CAAC,CAAC;gBAC7B,QAAQ,CAAC,YAAY,GAAG,KAAK,CAAC,CAAC;gBAC/B,MAAM,KAAK,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,GAAG,CAAC,CAAC;gBAC1C,KAAK,MAAM,IAAI,IAAI,KAAK,EAAE,CAAC;oBACzB,MAAM,QAAQ,GAAG,MAAM,IAAI,CAAC,WAAW,CAAC,IAAI,EAAE,GAAG,CAAC,CAAC;oBACnD,IAAI,CAAC,IAAI,CAAC,WAAW,CAAC,QAAQ,CAAC,EAAE,CAAC;wBAChC,QAAQ,CAAC,iBAAiB,QAAQ,KAAK,IAAI,GAAG,CAAC,CAAC;wBAChD,SAAS;oBACX,CAAC;oBACD,QAAQ,CACN,GAAG,QAAQ,oBAAoB,CAAC,GAAG,IAAI,CAAC,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CACrE,CAAC;oBACF,IAAI,CAAC,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,IAAI,QAAQ,CAAC,GAAG,KAAK,EAAE,CAAC;wBACrD,QAAQ,CACN,mBAAmB,QAAQ,aAAa,IAAI,CAAC,OAAO,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CACrE,CAAC;wBACF,SAAS;oBACX,CAAC;oBACD,IAAI,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,QAAQ,CAAC,EAAE,CAAC;wBAClC,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,SAAS,CAAC,QAAQ,CAAC,CAAC;wBAC5C,IAAI,IAAI,KAAK,IAAI,EAAE,CAAC;4BAClB,SAAS;wBACX,CAAC;wBACD,MAAM,WAAW,GAAG,IAAI,CAAC,cAAc,CAAC,IAAI,CAAC,CAAC;wBAC9C,IACE,WAAW,KAAK,IAAI;4BACpB,CAAC,CAAC,GAAG,IAAI,CAAC,cAAc,CAAC,MAAM,EAAE,CAAC,CAAC,IAAI,CAAC,CAAC,aAAa,EAAE,EAAE,CACxD,WAAW,CAAC,UAAU,CAAC,aAAa,CAAC,CACtC,EACD,CAAC;4BACD,QAAQ,CACN,WAAW,QAAQ,kBAAkB,WAAW,eAAe,CAChE,CAAC;4BACF,SAAS;wBACX,CAAC;wBACD,QAAQ,CAAC,UAAU,QAAQ,uBAAuB,CAAC,CAAC;wBACpD,IAAI,CAAC,YAAY,CAAC,QAAQ,EAAE,KAAK,GAAG,CAAC,CAAC,CAAC;oBACzC,CAAC;gBACH,CAAC;YACH,CAAC;YACD,CAAC,GAAG,EAAE,KAAK,CAAC,GAAG,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC;QACxC,CAAC;QAED,OAAO,CAAC,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,EAAE,CAAC,GAAG,IAAI,CAAC,SAAS,CAAC,IAAI,EAAE,CAAC,CAAC,CAAC;IAChE,CAAC;IAED,KAAK;QACH,IAAI,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC;QACrB,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,CAAC;QACvB,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,CAAC;QACvB,IAAI,CAAC,KAAK,CAAC,KAAK,EAAE,CAAC;QACnB,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,CAAC;QAC5B,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,WAAW,CAAC,CAAC;IACvC,CAAC;IAED,KAAK,CAAC,OAAO,CAAC,IAAY;QACxB,MAAM,SAAS,GAAG,MAAM,IAAI,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC;QAChD,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,SAAS,CAAC,CAAC;QAC1B,OAAO,SAAS,CAAC;IACnB,CAAC;IAED,gBAAgB,CAAC,aAAoC;QACnD,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,aAAa,CAAC,CAAC;IACzC,CAAC;IAED,iBAAiB,CAAC,GAAG,cAAuC;QAC1D,IAAI,CAAC,cAAc,CAAC,KAAK,EAAE,CAAC;QAC5B,cAAc,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,IAAI,CAAC,cAAc,CAAC,GAAG,CAAC,CAAC,CAAC,CAAC,CAAC;IAC5D,CAAC;IAED;;;;;;;OAOG;IACH,WAAW,CAAC,KAAa;QACvB,IAAI,KAAK,GAAG,CAAC,EAAE,CAAC;YACd,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC;QACpB,CAAC;aAAM,IAAI,MAAM,CAAC,QAAQ,CAAC,KAAK,CAAC,EAAE,CAAC;YAClC,IAAI,CAAC,QAAQ,GAAG,IAAI,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACpC,CAAC;aAAM,CAAC;YACN,IAAI,CAAC,QAAQ,GAAG,CAAC,CAAC;QACpB,CAAC;QACD,OAAO,IAAI,CAAC,QAAQ,CAAC;IACvB,CAAC;IAED,cAAc;QACZ,OAAO,CAAC,GAAG,IAAI,CAAC,OAAO,CAAC,IAAI,EAAE,CAAC,CAAC;IAClC,CAAC;IAED,YAAY;QACV,OAAO,IAAI,CAAC,cAAc,EAAE;aACzB,GAAG,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,SAAS,CAAC,GAAG,CAAC,GAAG,CAAC,CAAC;aACrC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,CAAC,KAAK,SAAS,CAAC,CAAC;IACpC,CAAC;CACF"}
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@dancrumb/web-crawler",
3
- "version": "1.4.0",
3
+ "version": "3.0.0",
4
4
  "description": "",
5
5
  "files": [
6
6
  "dist"
@@ -10,7 +10,8 @@
10
10
  },
11
11
  "scripts": {
12
12
  "build": "tsc",
13
- "test": "vitest --run"
13
+ "test": "vitest --run",
14
+ "prepare": "rm -rf dist && npm run build"
14
15
  },
15
16
  "keywords": [],
16
17
  "author": "Dan Rumney<dancrumb@gmail.com>",
@@ -37,4 +38,4 @@
37
38
  "path": "./node_modules/git-cz"
38
39
  }
39
40
  }
40
- }
41
+ }