hydra-crawler 2.0.1 → 2.0.2

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.
@@ -14,7 +14,6 @@ export declare class ImageTagsParser extends HtmlParser<IDataConfig> {
14
14
  private url?;
15
15
  static parseImageTags(dom: cheerio.Root): TImageTag[];
16
16
  constructor(url?: string | undefined, outcome?: IRequestOutcome, config?: TKeyObject<IParserConfig>);
17
- init(database: DatabaseService): Promise<void>;
18
17
  parse(database: DatabaseService): Promise<void>;
19
18
  }
20
19
  export {};
@@ -34,14 +34,6 @@ export class ImageTagsParser extends HtmlParser {
34
34
  });
35
35
  return images;
36
36
  }
37
- init(database) {
38
- return __awaiter(this, void 0, void 0, function* () {
39
- const urls = database.getUrls();
40
- if (!urls)
41
- return;
42
- yield urls.createIndex({ 'images.alt': 1 }, { unique: false });
43
- });
44
- }
45
37
  parse(database) {
46
38
  return __awaiter(this, void 0, void 0, function* () {
47
39
  if (!this.dom || !this.url)
@@ -1 +1 @@
1
- {"version":3,"file":"image-tags.parser.js","sourceRoot":"","sources":["../../src/parsers/image-tags.parser.ts"],"names":[],"mappings":";;;;;;;;;AAMA,OAAO,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AAOxD,OAAO,EAAE,UAAU,EAAE,MAAM,eAAe,CAAC;AAS3C,MAAM,OAAO,eAAgB,SAAQ,UAAuB;IA2B3D,YACU,GAAY,EACpB,OAAyB,EACzB,MAAkC;QAEnC,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAC;QAJ3B,QAAG,GAAH,GAAG,CAAS;IAKtB,CAAC;IAhCM,MAAM,CAAC,cAAc,CAC1B,GAAiB;QAElB,MAAM,MAAM,GAAgB,EAAE,CAAC;QAE/B,GAAG,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,MAAc,EAAE,OAAwB,EAAW,EAAE;YACrE,MAAM,GAAG,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC5D,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,IAAI,CAAC;YAEnD,MAAM,GAAG,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC5D,MAAM,KAAK,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAEhE,MAAM,KAAK,GAAc;gBACvB,GAAG,EAAE,GAAG;aACT,CAAC;YACF,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,IAAI;gBAAE,KAAK,CAAC,GAAG,GAAG,GAAG,CAAC;YACvD,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;gBAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC;YAE/D,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAEnB,OAAO,IAAI,CAAC;QACb,CAAC,CAAC,CAAC;QAEH,OAAO,MAAM,CAAC;IACf,CAAC;IAUY,IAAI,CAAC,QAAyB;;YAC1C,MAAM,IAAI,GAA+B,QAAQ,CAAC,OAAO,EAAE,CAAC;YAC5D,IAAI,CAAC,IAAI;gBAAE,OAAO;YAElB,MAAM,IAAI,CAAC,WAAW,CAAC,EAAE,YAAY,EAAE,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;QAChE,CAAC;KAAA;IAEY,KAAK,CAAC,QAAyB;;YAC3C,IAAI,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG;gBAAE,OAAO,CAAC,gBAAgB;YAEpD,MAAM,MAAM,GAAgB,eAAe,CAAC,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACrE,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE;gBACtB,kBAAkB,CAAC,uBAAuB,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;gBACtD,MAAM,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;aACnD;iBAAM;gBACN,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;aAC7C;QACF,CAAC;KAAA;CACD"}
1
+ {"version":3,"file":"image-tags.parser.js","sourceRoot":"","sources":["../../src/parsers/image-tags.parser.ts"],"names":[],"mappings":";;;;;;;;;AAEA,OAAO,EAAE,kBAAkB,EAAE,MAAM,oBAAoB,CAAC;AAOxD,OAAO,EAAE,UAAU,EAAE,MAAM,eAAe,CAAC;AAS3C,MAAM,OAAO,eAAgB,SAAQ,UAAuB;IA2B3D,YACU,GAAY,EACpB,OAAyB,EACzB,MAAkC;QAEnC,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,WAAW,CAAC,CAAC;QAJ3B,QAAG,GAAH,GAAG,CAAS;IAKtB,CAAC;IAhCM,MAAM,CAAC,cAAc,CAC1B,GAAiB;QAElB,MAAM,MAAM,GAAgB,EAAE,CAAC;QAE/B,GAAG,CAAC,KAAK,CAAC,CAAC,IAAI,CAAC,CAAC,MAAc,EAAE,OAAwB,EAAW,EAAE;YACrE,MAAM,GAAG,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC5D,IAAI,GAAG,KAAK,IAAI,IAAI,GAAG,KAAK,SAAS;gBAAE,OAAO,IAAI,CAAC;YAEnD,MAAM,GAAG,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAC5D,MAAM,KAAK,GAA0B,GAAG,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC;YAEhE,MAAM,KAAK,GAAc;gBACvB,GAAG,EAAE,GAAG;aACT,CAAC;YACF,IAAI,GAAG,KAAK,SAAS,IAAI,GAAG,KAAK,IAAI;gBAAE,KAAK,CAAC,GAAG,GAAG,GAAG,CAAC;YACvD,IAAI,KAAK,KAAK,SAAS,IAAI,KAAK,KAAK,IAAI;gBAAE,KAAK,CAAC,KAAK,GAAG,KAAK,CAAC;YAE/D,MAAM,CAAC,IAAI,CAAC,KAAK,CAAC,CAAC;YAEnB,OAAO,IAAI,CAAC;QACb,CAAC,CAAC,CAAC;QAEH,OAAO,MAAM,CAAC;IACf,CAAC;IAUY,KAAK,CAAC,QAAyB;;YAC3C,IAAI,CAAC,IAAI,CAAC,GAAG,IAAI,CAAC,IAAI,CAAC,GAAG;gBAAE,OAAO,CAAC,gBAAgB;YAEpD,MAAM,MAAM,GAAgB,eAAe,CAAC,cAAc,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;YACrE,IAAI,MAAM,CAAC,MAAM,GAAG,CAAC,EAAE;gBACtB,kBAAkB,CAAC,uBAAuB,IAAI,CAAC,GAAG,EAAE,CAAC,CAAC;gBACtD,MAAM,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,EAAE,MAAM,CAAC,CAAC;aACnD;iBAAM;gBACN,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;aAC7C;QACF,CAAC;KAAA;CACD"}
@@ -6,7 +6,6 @@ import { Parser } from './parser';
6
6
  export declare class ServerParser extends Parser<IParserConfig> {
7
7
  private url?;
8
8
  constructor(url?: string | undefined, outcome?: IRequestOutcome, config?: TKeyObject<IParserConfig>);
9
- init(database: DatabaseService): Promise<void>;
10
9
  supports(_contentType: string): boolean;
11
10
  parse(database: DatabaseService): Promise<void>;
12
11
  }
@@ -15,15 +15,6 @@ export class ServerParser extends Parser {
15
15
  super(outcome, config, 'server');
16
16
  this.url = url;
17
17
  }
18
- init(database) {
19
- return __awaiter(this, void 0, void 0, function* () {
20
- const urls = database.getUrls();
21
- if (!urls)
22
- return;
23
- yield urls.createIndex({ 'server.server': 1 }, { unique: false });
24
- yield urls.createIndex({ 'server.xPoweredBy': 1 }, { unique: false });
25
- });
26
- }
27
18
  supports(_contentType) {
28
19
  return true; // everything
29
20
  }
@@ -1 +1 @@
1
- {"version":3,"file":"server.parser.js","sourceRoot":"","sources":["../../src/parsers/server.parser.ts"],"names":[],"mappings":";;;;;;;;;AAYA,OAAO,EAAE,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAC3C,OAAO,EAAE,WAAW,EAAE,MAAM,wBAAwB,CAAC;AAErD,OAAO,EAAE,MAAM,EAAE,MAAM,UAAU,CAAC;AAElC,MAAM,OAAO,YAAa,SAAQ,MAAqB;IACtD,YACU,GAAY,EACpB,OAAyB,EACzB,MAAkC;QAEnC,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC;QAJxB,QAAG,GAAH,GAAG,CAAS;IAKtB,CAAC;IAEY,IAAI,CAAC,QAAyB;;YAC1C,MAAM,IAAI,GAA+B,QAAQ,CAAC,OAAO,EAAE,CAAC;YAC5D,IAAI,CAAC,IAAI;gBAAE,OAAO;YAElB,MAAM,IAAI,CAAC,WAAW,CAAC,EAAE,eAAe,EAAE,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;YAClE,MAAM,IAAI,CAAC,WAAW,CAAC,EAAE,mBAAmB,EAAE,CAAC,EAAE,EAAE,EAAE,MAAM,EAAE,KAAK,EAAE,CAAC,CAAC;QACvE,CAAC;KAAA;IAEM,QAAQ,CAAC,YAAoB;QACnC,OAAO,IAAI,CAAC,CAAC,aAAa;IAC3B,CAAC;IAEY,KAAK,CAAC,QAAyB;;YAC3C,IAAI,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,CAAC,GAAG;gBAAE,OAAO;YAEvC,MAAM,MAAM,GAAsB,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,QAAQ,CAAY,IAAI,SAAS,CAAC;YACzF,MAAM,UAAU,GAAsB,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,cAAc,CAAY,IAAI,SAAS,CAAC;YAEnG,IAAI,MAAM,KAAK,SAAS,IAAI,UAAU,KAAK,SAAS,EAAE;gBACrD,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;gBAC7C,OAAO;aACP;YAED,MAAM,IAAI,GAAY,EAAE,CAAC;YAEzB,IAAI,MAAM,EAAE;gBACX,IAAI,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;gBAC/D,IAAI,sBAAsB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC;gBACnE,IAAI,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;gBAC7D,IAAI,mBAAmB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC;gBACvE,IAAI,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,QAAQ,CAAC;gBACnE,IAAI,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;aAC7D;YAED,IAAI,UAAU,EAAE;gBACf,IAAI,YAAY,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,GAAG,CAAC;gBACrE,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,MAAM,CAAC;gBAC5E,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,OAAO,CAAC;gBAC7E,IAAI,aAAa,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,IAAI,CAAC;aACvE;YAED,IAAI,IAAI,CAAC,MAAM,KAAK,SAAS,IAAI,IAAI,CAAC,UAAU,KAAK,SAAS,EAAE;gBAC/D,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;gBAC7C,OAAO;aACP;YAED,IAAI;gBACH,MAAM,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,EAAE,IAAI,CAAC,CAAC;aACjD;YAAC,OAAO,EAAE,EAAE,EAAE,gBAAgB,EAAE;QAClC,CAAC;KAAA;CACD"}
1
+ {"version":3,"file":"server.parser.js","sourceRoot":"","sources":["../../src/parsers/server.parser.ts"],"names":[],"mappings":";;;;;;;;;AAQA,OAAO,EAAE,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAC3C,OAAO,EAAE,WAAW,EAAE,MAAM,wBAAwB,CAAC;AAErD,OAAO,EAAE,MAAM,EAAE,MAAM,UAAU,CAAC;AAElC,MAAM,OAAO,YAAa,SAAQ,MAAqB;IACtD,YACU,GAAY,EACpB,OAAyB,EACzB,MAAkC;QAEnC,KAAK,CAAC,OAAO,EAAE,MAAM,EAAE,QAAQ,CAAC,CAAC;QAJxB,QAAG,GAAH,GAAG,CAAS;IAKtB,CAAC;IAEM,QAAQ,CAAC,YAAoB;QACnC,OAAO,IAAI,CAAC,CAAC,aAAa;IAC3B,CAAC;IAEY,KAAK,CAAC,QAAyB;;YAC3C,IAAI,CAAC,IAAI,CAAC,OAAO,IAAI,CAAC,IAAI,CAAC,GAAG;gBAAE,OAAO;YAEvC,MAAM,MAAM,GAAsB,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,QAAQ,CAAY,IAAI,SAAS,CAAC;YACzF,MAAM,UAAU,GAAsB,IAAI,CAAC,OAAO,CAAC,OAAO,CAAC,cAAc,CAAY,IAAI,SAAS,CAAC;YAEnG,IAAI,MAAM,KAAK,SAAS,IAAI,UAAU,KAAK,SAAS,EAAE;gBACrD,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;gBAC7C,OAAO;aACP;YAED,MAAM,IAAI,GAAY,EAAE,CAAC;YAEzB,IAAI,MAAM,EAAE;gBACX,IAAI,eAAe,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC;gBAC/D,IAAI,sBAAsB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,GAAG,CAAC;gBACnE,IAAI,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;gBAC7D,IAAI,mBAAmB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,UAAU,CAAC;gBACvE,IAAI,iBAAiB,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,QAAQ,CAAC;gBACnE,IAAI,cAAc,CAAC,IAAI,CAAC,MAAM,CAAC;oBAAE,IAAI,CAAC,MAAM,GAAG,OAAO,CAAC,KAAK,CAAC;aAC7D;YAED,IAAI,UAAU,EAAE;gBACf,IAAI,YAAY,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,GAAG,CAAC;gBACrE,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,MAAM,CAAC;gBAC5E,IAAI,gBAAgB,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,OAAO,CAAC;gBAC7E,IAAI,aAAa,CAAC,IAAI,CAAC,UAAU,CAAC;oBAAE,IAAI,CAAC,UAAU,GAAG,WAAW,CAAC,IAAI,CAAC;aACvE;YAED,IAAI,IAAI,CAAC,MAAM,KAAK,SAAS,IAAI,IAAI,CAAC,UAAU,KAAK,SAAS,EAAE;gBAC/D,MAAM,QAAQ,CAAC,SAAS,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,CAAC,CAAC;gBAC7C,OAAO;aACP;YAED,IAAI;gBACH,MAAM,QAAQ,CAAC,OAAO,CAAC,IAAI,CAAC,GAAG,EAAE,QAAQ,EAAE,IAAI,CAAC,CAAC;aACjD;YAAC,OAAO,EAAE,EAAE,EAAE,gBAAgB,EAAE;QAClC,CAAC;KAAA;CACD"}
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "hydra-crawler",
3
- "version": "2.0.1",
3
+ "version": "2.0.2",
4
4
  "description": "Node.js Hydra web crawler",
5
5
  "author": "Pete Morris",
6
6
  "license": "ISC",