webforai 1.6.2 → 2.0.0
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.
- package/LICENSE +201 -21
- package/README.md +45 -111
- package/dist/bin.js +59 -110
- package/dist/cjs/constants.js +34 -0
- package/dist/cjs/extractors/index.js +31 -0
- package/dist/cjs/{extract-hast/index.js → extractors/pipeExtractors.js} +9 -19
- package/dist/cjs/{extract-hast/readability.js → extractors/presets/takumi.js} +6 -6
- package/dist/cjs/extractors/types.js +16 -0
- package/dist/cjs/html-to-mdast.js +2 -2
- package/dist/cjs/index.js +7 -7
- package/dist/cjs/loaders/puppeteer.js +1 -1
- package/dist/constants.js +10 -0
- package/dist/extractors/index.js +9 -0
- package/dist/{extract-hast/index.js → extractors/pipeExtractors.js} +6 -14
- package/dist/{extract-hast/readability.js → extractors/presets/takumi.js} +3 -3
- package/dist/{extract-hast → extractors/presets}/utils.js +1 -1
- package/dist/extractors/types.js +0 -0
- package/dist/html-to-mdast.js +2 -2
- package/dist/index.js +7 -4
- package/dist/loaders/puppeteer.js +1 -1
- package/dist/types/cli/commands/webforai/loadHtml.js.map +1 -1
- package/dist/types/constants.d.ts +4 -0
- package/dist/types/constants.js.map +1 -0
- package/dist/types/extractors/index.d.ts +3 -0
- package/dist/types/extractors/index.js.map +1 -0
- package/dist/types/extractors/pipeExtractors.d.ts +7 -0
- package/dist/types/extractors/pipeExtractors.js.map +1 -0
- package/dist/types/extractors/presets/takumi.d.ts +10 -0
- package/dist/types/extractors/presets/takumi.js.map +1 -0
- package/dist/types/extractors/presets/utils.js.map +1 -0
- package/dist/types/extractors/types.d.ts +7 -0
- package/dist/types/extractors/types.js.map +1 -0
- package/dist/types/html-to-mdast.d.ts +2 -2
- package/dist/types/html-to-mdast.js.map +1 -1
- package/dist/types/index.d.ts +1 -1
- package/dist/types/index.js.map +1 -1
- package/dist/types/loaders/puppeteer.js.map +1 -1
- package/package.json +3 -3
- package/dist/cjs/extract-hast/strongly.js +0 -57
- package/dist/extract-hast/strongly.js +0 -34
- package/dist/types/extract-hast/index.d.ts +0 -15
- package/dist/types/extract-hast/index.js.map +0 -1
- package/dist/types/extract-hast/readability.d.ts +0 -3
- package/dist/types/extract-hast/readability.js.map +0 -1
- package/dist/types/extract-hast/strongly.d.ts +0 -3
- package/dist/types/extract-hast/strongly.js.map +0 -1
- package/dist/types/extract-hast/utils.js.map +0 -1
- /package/dist/cjs/{extract-hast → extractors/presets}/utils.js +0 -0
- /package/dist/types/{extract-hast → extractors/presets}/utils.d.ts +0 -0
package/dist/index.js
CHANGED
|
@@ -3,12 +3,15 @@ import { htmlToMarkdown } from "./html-to-markdown.js";
|
|
|
3
3
|
import { mdastSplitter } from "./md-splitter.js";
|
|
4
4
|
import { htmlToMdast } from "./html-to-mdast.js";
|
|
5
5
|
import { mdastToMarkdown } from "./mdast-to-markdown.js";
|
|
6
|
-
import {
|
|
6
|
+
import {
|
|
7
|
+
pipeExtractors,
|
|
8
|
+
takumiExtractor
|
|
9
|
+
} from "./extractors/index.js";
|
|
7
10
|
export {
|
|
8
|
-
DEFAULT_EXTRACT_HAST,
|
|
9
|
-
PRESET_EXTRACT_HAST,
|
|
10
11
|
htmlToMarkdown,
|
|
11
12
|
htmlToMdast,
|
|
12
13
|
mdastSplitter,
|
|
13
|
-
mdastToMarkdown
|
|
14
|
+
mdastToMarkdown,
|
|
15
|
+
pipeExtractors,
|
|
16
|
+
takumiExtractor
|
|
14
17
|
};
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"loadHtml.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/loadHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,EAAE,GAAG,EAAE,MAAM,gBAAgB,CAAC;AACrC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,EAAE,QAAQ,EAAE,MAAM,iBAAiB,CAAC;AAC3C,OAAO,EAAE,QAAQ,IAAI,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACrE,OAAO,EAAE,QAAQ,IAAI,oBAAoB,EAAE,MAAM,6BAA6B,CAAC;AAE/E,MAAM,wBAAwB,GAAG,KAAK,IAAI,EAAE;IAC3C,MAAM,IAAI,GAAG,QAAQ,CAAC,cAAc,EAAE,CAAC;IACvC,IAAI,CAAC;QACJ,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACtB,OAAO,IAAI,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAC;IACd,CAAC;AACF,CAAC,CAAC;AAEF,MAAM,oBAAoB,GAAG,KAAK,IAAI,EAAE;IACvC,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,8BAA8B,CAAC,CAAC;IACvE,MAAM,aAAa,GAAG,MAAM,EAAE;SAC5B,QAAQ,CAAC,
|
|
1
|
+
{"version":3,"file":"loadHtml.js","sourceRoot":"","sources":["../../../../../src/cli/commands/webforai/loadHtml.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,MAAM,kBAAkB,CAAC;AAClC,OAAO,EAAE,aAAa,EAAE,MAAM,UAAU,CAAC;AACzC,OAAO,EAAE,GAAG,EAAE,MAAM,gBAAgB,CAAC;AACrC,OAAO,KAAK,MAAM,OAAO,CAAC;AAC1B,OAAO,EAAE,MAAM,YAAY,CAAC;AAC5B,OAAO,EAAE,QAAQ,EAAE,MAAM,iBAAiB,CAAC;AAC3C,OAAO,EAAE,QAAQ,IAAI,eAAe,EAAE,MAAM,wBAAwB,CAAC;AACrE,OAAO,EAAE,QAAQ,IAAI,oBAAoB,EAAE,MAAM,6BAA6B,CAAC;AAE/E,MAAM,wBAAwB,GAAG,KAAK,IAAI,EAAE;IAC3C,MAAM,IAAI,GAAG,QAAQ,CAAC,cAAc,EAAE,CAAC;IACvC,IAAI,CAAC;QACJ,MAAM,EAAE,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;QACtB,OAAO,IAAI,CAAC;IACb,CAAC;IAAC,MAAM,CAAC;QACR,OAAO,KAAK,CAAC;IACd,CAAC;AACF,CAAC,CAAC;AAEF,MAAM,oBAAoB,GAAG,KAAK,IAAI,EAAE;IACvC,MAAM,IAAI,GAAG,MAAM,MAAM,CAAC,IAAI,CAAC,OAAO,CAAC,8BAA8B,CAAC,CAAC;IACvE,MAAM,aAAa,GAAG,MAAM,EAAE;SAC5B,QAAQ,CAAC,aAAa,CAAC,IAAI,CAAC,EAAE,OAAO,CAAC;SACtC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,IAAI,CAAC,KAAK,CAAC,GAAG,CAAC,QAAQ,EAAE,CAAC,CAAC;SACzC,KAAK,CAAC,GAAG,EAAE,CAAC,IAAI,CAAC,CAAC;IACpB,OAAO,aAAa,EAAE,OAAO,CAAC;AAC/B,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,UAAkB,EAAE,MAAc,EAAE,OAA4B,EAAE,EAAE;IAClG,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QACzE,MAAM,OAAO,GAAG,MAAM,EAAE,CAAC,QAAQ,CAAC,UAAU,EAAE,OAAO,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,OAAO,EAAE,CAAC;QACxB,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,0BAA0B,UAAU,EAAE,CAAC,CAAC;QAClE,MAAM,OAAO,GAAG,MAAM,eAAe,CAAC,UAAU,CAAC,CAAC;QAClD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,IAAI,MAAM,KAAK,YAAY,EAAE,CAAC;QAC7B,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,iCAAiC,UAAU,EAAE,CAAC,CAAC;QACzE,MAAM,qBAAqB,GAAG,MAAM,wBAAwB,EAAE,CAAC;QAC/D,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,yBAAyB,qBAAqB,EAAE,CAAC,CAAC;QAE5E,MAAM,SAAS,GAAG,MAAM,oBAAoB,EAAE,CAAC;QAE/C,IAAI,CAAC,qBAAqB,EAAE,CAAC;YAC5B,MAAM,OAAO,GAAG;gBACf,EAAE,CAAC,IAAI,CAAC,oCAAoC,CAAC;gBAC7C,EAAE;gBACF,qEAAqE;gBACrE,EAAE;gBACF,oBAAoB,SAAS,mBAAmB;gBAChD,EAAE;gBACF,6CAA6C;gBAC7C,2HAA2H;gBAC3H,qCAAqC;gBACrC,EAAE;gBACF,iDAAiD;gBACjD,0DAA0D;gBAC1D,6DAA6D;gBAC7D,EAAE;gBACF,yBAAyB,SAAS,eAAe;aACjD,CAAC;YAEF,GAAG,CAAC,KAAK,CAAC,KAAK,CAAC,OAAO,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,WAAW,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC;YAC3E,MAAM,IAAI,KAAK,CAAC,6BAA6B,CAAC,CAAC;QAChD,CAAC;QACD,MAAM,OAAO,GAAG,MAAM,oBAAoB,CAAC,UAAU,CAAC,CAAC;QACvD,OAAO,CAAC,KAAK,IAAI,GAAG,CAAC,IAAI,CAAC,gBAAgB,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,GAAG,CAAC,EAAE,CAAC,CAAC;QACnE,OAAO,OAAO,CAAC;IAChB,CAAC;IAED,MAAM,IAAI,KAAK,CAAC,uBAAuB,MAAM,EAAE,CAAC,CAAC;AAClD,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"constants.js","sourceRoot":"","sources":["../../src/constants.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,eAAe,EAAE,MAAM,6BAA6B,CAAC;AAE9D,MAAM,CAAC,MAAM,iBAAiB,GAAG;IAChC,MAAM,EAAE,eAAe;CACvB,CAAC;AAEF,MAAM,CAAC,MAAM,kBAAkB,GAAG,CAAC,QAAiB,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extractors/index.ts"],"names":[],"mappings":"AAAA,2DAA2D;AAC3D,OAAO,EACN,cAAc,GAId,MAAM,kBAAkB,CAAC;AAC1B,OAAO,EAAE,eAAe,EAAE,MAAM,kBAAkB,CAAC"}
|
|
@@ -0,0 +1,7 @@
|
|
|
1
|
+
import type { Nodes as Hast } from "hast";
|
|
2
|
+
import { PRESET_EXTRACTORS } from "../constants";
|
|
3
|
+
import type { ExtractParams, Extractor } from "./types";
|
|
4
|
+
export type PresetExtractors = keyof typeof PRESET_EXTRACTORS;
|
|
5
|
+
export type ExtractorSelector = Extractor | false | PresetExtractors;
|
|
6
|
+
export type ExtractorSelectors = ExtractorSelector | ExtractorSelector[];
|
|
7
|
+
export declare const pipeExtractors: (params: ExtractParams, extractors?: ExtractorSelectors) => Hast;
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"pipeExtractors.js","sourceRoot":"","sources":["../../../src/extractors/pipeExtractors.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,kBAAkB,EAAE,iBAAiB,EAAE,MAAM,cAAc,CAAC;AAOrE,MAAM,CAAC,MAAM,cAAc,GAAG,CAAC,MAAqB,EAAE,aAAiC,kBAAkB,EAAQ,EAAE;IAClH,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IAC9B,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAE1E,MAAM,SAAS,GACd,WAAW,CAAC,MAAM,CAAO,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE;QAC3C,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACzB,OAAO,GAAG,CAAC;QACZ,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,IAAI,iBAAiB,EAAE,CAAC;YACrE,OAAO,iBAAiB,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QAC1D,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;YACrC,OAAO,SAAS,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,sBAAsB,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAElB,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -0,0 +1,10 @@
|
|
|
1
|
+
import type { Nodes as Hast } from "hast";
|
|
2
|
+
import type { ExtractParams } from "../types";
|
|
3
|
+
/**
|
|
4
|
+
* Currently the best Extractor.
|
|
5
|
+
* The word "takumi" is written as 匠 in Japanese, and it refers to a highly skilled artisan or craftsman.
|
|
6
|
+
*
|
|
7
|
+
* @param params - {@link ExtractParams}
|
|
8
|
+
* @returns The HAST tree.
|
|
9
|
+
*/
|
|
10
|
+
export declare const takumiExtractor: (params: ExtractParams) => Hast;
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"takumi.js","sourceRoot":"","sources":["../../../../src/extractors/presets/takumi.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAEhE,MAAM,cAAc,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;AAE5G;;;;;;GAMG;AACH,MAAM,OAAO,GAAG;IACf,MAAM,EAAE,kCAAkC;IAC1C,MAAM,EAAE,4CAA4C;IACpD,yBAAyB,EAAE,8BAA8B;IACzD,kBAAkB,EACjB,wRAAwR;IACzR,oBAAoB,EAAE,mDAAmD;CACzE,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;AAEvG,MAAM,cAAc,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;AAEzF,MAAM,mBAAmB,GAAG;IAC3B,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;CACP,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,IAAU,EAAE,EAAE;IACrC,OAAO,CAAC,CACP,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QAC1C,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CACnH,CAAC;AACH,CAAC,CAAC;AAEF,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAChD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yCAAyC;IACzC,IAAI,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACtE,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QAC7E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QAC9E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,KAAK,QAAQ,IAAI,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAChG,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,KAAK,CAAC;IACd,CAAC;IAED,wBAAwB;IACxB,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,CAAC,IAAc,CAAC,EAAE,CAAC;QAC3F,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,qBAAqB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC5C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,6BAA6B;IAC7B,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/G,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,KAAK,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;IAEnC,IAAI,OAAO,CAAC,yBAAyB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,6BAA6B;IAC7B,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzF,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,iBAAiB,GAAG,CAAC,IAAU,EAAE,KAAa,EAAE,EAAE;IACvD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF;;;;;;GAMG;AACH,MAAM,CAAC,MAAM,eAAe,GAAG,CAAC,MAAqB,EAAQ,EAAE;IAC9D,MAAM,EAAE,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,MAAM,CAAC;IACrC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAE1C,MAAM,oBAAoB,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAY,CAAC,CAAC,CAAC;IAClF,MAAM,wBAAwB,GAAG,oBAAoB,IAAI,YAAY,CAAC,oBAAoB,CAAC,CAAC;IAC5F,IAAI,CAAC,CAAC,oBAAoB,IAAI,wBAAwB,CAAC,EAAE,CAAC;QACzD,OAAO,IAAI,CAAC;IACb,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,CAAC,oBAAoB,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,sBAAsB,CAAC,IAAY,CAAC,CAAC,CAAC;IACjG,MAAM,eAAe,GAAG,WAAW,CAAC,CAAC,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACrE,MAAM,CAAC,QAAQ,EAAE,QAAQ,CAAC,GACzB,eAAe,CAAC,MAAM,GAAG,wBAAwB,CAAC,MAAM,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,GAAG,IAAI;QAC5F,CAAC,CAAE,CAAC,WAAmB,EAAE,eAAe,CAAW;QACnD,CAAC,CAAE,CAAC,oBAA4B,EAAE,wBAAwB,CAAW,CAAC;IAExE,IAAI,aAAa,GAAG,IAAI,IAAI,mBAAmB,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAwC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;IACtH,IAAI,QAAQ,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QACrC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IAED,IAAI,aAAa,GAAS,QAAQ,CAAC;IACnC,IAAI,aAAa,GAAG,QAAQ,CAAC;IAE7B,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,EAAE,IAAI,EAAE,MAAe,EAAE,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,WAAW,CAAC,EAAE,CAAC;QACtF,MAAM,WAAW,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;QAE1C,IAAI,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC7B,SAAS;QACV,CAAC;QAED,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC;QACzD,IAAI,WAAW,GAAG,GAAG,EAAE,CAAC;YACvB,SAAS;QACV,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;YACxC,aAAa,GAAG,OAAO,CAAC;YACxB,aAAa,GAAG,WAAW,CAAC;YAC5B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,aAAa,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,CAAC,iBAAiB,CAAC,IAAY,EAAE,IAAI,CAAC,EAAE,CAAC;YAC5C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,qBAAqB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QAED,OAAO,IAAI,CAAC;IACb,CAAC,CAAS,CAAC;IAEX,MAAM,SAAS,GACd,YAAY,CAAC,iBAAiB,CAAC,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,aAAa,CAAC;IACvG,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../../src/extractors/presets/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC"}
|
|
@@ -0,0 +1 @@
|
|
|
1
|
+
{"version":3,"file":"types.js","sourceRoot":"","sources":["../../../src/extractors/types.ts"],"names":[],"mappings":""}
|
|
@@ -1,12 +1,12 @@
|
|
|
1
1
|
import type { Nodes as Hast } from "hast";
|
|
2
2
|
import type { Nodes as Mdast } from "mdast";
|
|
3
|
-
import { type
|
|
3
|
+
import { type ExtractorSelectors } from "./extractors";
|
|
4
4
|
export type HtmlToMdastOptions = {
|
|
5
5
|
/**
|
|
6
6
|
* An array of extractors to extract specific elements from the HTML.
|
|
7
7
|
* You can define your own functions in addition to the Extractor provided as a preset.
|
|
8
8
|
*/
|
|
9
|
-
extractors?:
|
|
9
|
+
extractors?: ExtractorSelectors;
|
|
10
10
|
/** Whether to convert links to plain text. */
|
|
11
11
|
linkAsText?: boolean;
|
|
12
12
|
/** Whether to convert tables to plain text. */
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,
|
|
1
|
+
{"version":3,"file":"html-to-mdast.js","sourceRoot":"","sources":["../../src/html-to-mdast.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,QAAQ,EAAE,MAAM,qBAAqB,CAAC;AAC/C,OAAO,EAAE,OAAO,EAAE,MAAM,oBAAoB,CAAC;AAG7C,OAAO,EAAE,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAC/C,OAAO,EAA2B,cAAc,EAAE,MAAM,cAAc,CAAC;AACvE,OAAO,EAAE,cAAc,EAAE,MAAM,mCAAmC,CAAC;AACnE,OAAO,EAAE,iBAAiB,EAAE,MAAM,sCAAsC,CAAC;AACzE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,gBAAgB,EAAE,MAAM,qCAAqC,CAAC;AACvE,OAAO,EAAE,kBAAkB,EAAE,MAAM,uCAAuC,CAAC;AAC3E,OAAO,EAAE,WAAW,EAAE,MAAM,+BAA+B,CAAC;AAC5D,OAAO,EAAE,eAAe,EAAE,cAAc,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AAoBrF;;;;;;;;;;;;;;;;GAgBG;AACH,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,UAAyB,EAAE,OAA4B,EAAS,EAAE;IAC7F,MAAM,EAAE,UAAU,EAAE,GAAG,EAAE,UAAU,EAAE,IAAI,EAAE,WAAW,EAAE,GAAG,OAAO,IAAI,EAAE,CAAC;IAEzE,MAAM,CAAC,IAAI,EAAE,IAAI,CAAC,GACjB,OAAO,UAAU,KAAK,QAAQ;QAC7B,CAAC,CAAC,CAAC,WAAW,IAAI,cAAc,CAAC,UAAU,CAAC,EAAE,QAAQ,CAAC,UAAU,EAAE,EAAE,QAAQ,EAAE,IAAI,EAAE,CAAC,CAAC;QACvF,CAAC,CAAC,CAAC,WAAW,IAAI,eAAe,CAAC,UAAU,CAAC,EAAE,UAAU,CAAC,CAAC;IAE7D,MAAM,GAAG,GAAG,UAAU,IAAI,cAAc,CAAC,IAAI,CAAC,CAAC;IAE/C,MAAM,aAAa,GAAG,cAAc,CAAC,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,EAAE,EAAE,UAAU,CAAC,CAAC;IAEtE,MAAM,KAAK,GAAG,OAAO,CAAC,aAAa,EAAE;QACpC,QAAQ,EAAE;YACT,IAAI,EAAE,WAAW;YACjB,GAAG,EAAE,gBAAgB;YACrB,GAAG,EAAE,iBAAiB;YACtB,CAAC,EAAE,cAAc,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,UAAU,EAAE,CAAC;YAClD,GAAG,EAAE,gBAAgB,CAAC,EAAE,SAAS,EAAE,OAAO,EAAE,SAAS,EAAE,CAAC;YACxD,KAAK,EAAE,kBAAkB,CAAC,EAAE,MAAM,EAAE,OAAO,EAAE,WAAW,EAAE,CAAC;SAC3D;KACD,CAAC,CAAC;IAEH,MAAM,cAAc,GAAG,YAAY,CAAC,KAAK,CAAC,CAAC;IAE3C,OAAO,cAAc,CAAC;AACvB,CAAC,CAAC"}
|
package/dist/types/index.d.ts
CHANGED
|
@@ -2,4 +2,4 @@ export { htmlToMarkdown, type HtmlToMarkdownOptions } from "./html-to-markdown";
|
|
|
2
2
|
export { mdastSplitter } from "./md-splitter";
|
|
3
3
|
export { htmlToMdast, type HtmlToMdastOptions } from "./html-to-mdast";
|
|
4
4
|
export { mdastToMarkdown } from "./mdast-to-markdown";
|
|
5
|
-
export {
|
|
5
|
+
export { pipeExtractors, takumiExtractor, type ExtractorSelectors, type ExtractorSelector, type PresetExtractors, type ExtractParams, type Extractor, } from "./extractors";
|
package/dist/types/index.js.map
CHANGED
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,2DAA2D;AAC3D,OAAO,EAAE,cAAc,EAA8B,MAAM,oBAAoB,CAAC;AAChF,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAC9C,OAAO,EAAE,WAAW,EAA2B,MAAM,iBAAiB,CAAC;AACvE,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,
|
|
1
|
+
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../src/index.ts"],"names":[],"mappings":"AAAA,2DAA2D;AAC3D,OAAO,EAAE,cAAc,EAA8B,MAAM,oBAAoB,CAAC;AAChF,OAAO,EAAE,aAAa,EAAE,MAAM,eAAe,CAAC;AAC9C,OAAO,EAAE,WAAW,EAA2B,MAAM,iBAAiB,CAAC;AACvE,OAAO,EAAE,eAAe,EAAE,MAAM,qBAAqB,CAAC;AACtD,OAAO,EACN,cAAc,EACd,eAAe,GAMf,MAAM,cAAc,CAAC"}
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"puppeteer.js","sourceRoot":"","sources":["../../../src/loaders/puppeteer.ts"],"names":[],"mappings":"AAAA,OAAO,SAAS,MAAM,WAAW,CAAC;AAGlC,MAAM,KAAK,GAAG,CAAC,EAAU,EAAE,EAAE,CAAC,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;AAEhF,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,GAAW,EAAE,GAA4B,EAAE,EAAE;IAC3E,MAAM,OAAO,GAAG,MAAM,SAAS,CAAC,MAAM,CACrC,GAAG,IAAI;QACN,QAAQ,EAAE,IAAI;QACd,IAAI,EAAE,CAAC,cAAc,EAAE,0BAA0B,CAAC;KAClD,CACD,CAAC;IACF,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;IACrC,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAErB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;IAElC,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,kBAAkB,EAAE,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAE9D,MAAM,
|
|
1
|
+
{"version":3,"file":"puppeteer.js","sourceRoot":"","sources":["../../../src/loaders/puppeteer.ts"],"names":[],"mappings":"AAAA,OAAO,SAAS,MAAM,WAAW,CAAC;AAGlC,MAAM,KAAK,GAAG,CAAC,EAAU,EAAE,EAAE,CAAC,IAAI,OAAO,CAAC,CAAC,OAAO,EAAE,EAAE,CAAC,UAAU,CAAC,OAAO,EAAE,EAAE,CAAC,CAAC,CAAC;AAEhF,MAAM,CAAC,MAAM,QAAQ,GAAG,KAAK,EAAE,GAAW,EAAE,GAA4B,EAAE,EAAE;IAC3E,MAAM,OAAO,GAAG,MAAM,SAAS,CAAC,MAAM,CACrC,GAAG,IAAI;QACN,QAAQ,EAAE,IAAI;QACd,IAAI,EAAE,CAAC,cAAc,EAAE,0BAA0B,CAAC;KAClD,CACD,CAAC;IACF,MAAM,IAAI,GAAG,MAAM,OAAO,CAAC,OAAO,EAAE,CAAC;IACrC,MAAM,IAAI,CAAC,IAAI,CAAC,GAAG,CAAC,CAAC;IAErB,MAAM,IAAI,GAAG,MAAM,IAAI,CAAC,OAAO,EAAE,CAAC;IAElC,MAAM,OAAO,CAAC,IAAI,CAAC,CAAC,IAAI,CAAC,kBAAkB,EAAE,EAAE,KAAK,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC;IAE9D,MAAM,OAAO,CAAC,KAAK,EAAE,CAAC;IAEtB,OAAO,IAAI,CAAC;AACb,CAAC,CAAC"}
|
package/package.json
CHANGED
|
@@ -1,9 +1,9 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "webforai",
|
|
3
|
-
"version": "
|
|
3
|
+
"version": "2.0.0",
|
|
4
4
|
"description": "A library that provides a web interface for AI",
|
|
5
5
|
"author": "inaridiy",
|
|
6
|
-
"license": "
|
|
6
|
+
"license": "Apache-2.0",
|
|
7
7
|
"keywords": [
|
|
8
8
|
"web",
|
|
9
9
|
"ai",
|
|
@@ -135,7 +135,7 @@
|
|
|
135
135
|
"copy:package.cjs.json": "pnpm ncp ./package.cjs.json ./dist/cjs/package.json && pnpm ncp ./package.cjs.json ./dist/types/package.json ",
|
|
136
136
|
"clean": "rimraf dist",
|
|
137
137
|
"build": "pnpm clean && tsx build.ts && pnpm copy:package.cjs.json",
|
|
138
|
-
"
|
|
138
|
+
"typecheck": "tsc --noEmit",
|
|
139
139
|
"prerelease": "pnpm build",
|
|
140
140
|
"release": "np"
|
|
141
141
|
}
|
|
@@ -1,57 +0,0 @@
|
|
|
1
|
-
"use strict";
|
|
2
|
-
var __defProp = Object.defineProperty;
|
|
3
|
-
var __getOwnPropDesc = Object.getOwnPropertyDescriptor;
|
|
4
|
-
var __getOwnPropNames = Object.getOwnPropertyNames;
|
|
5
|
-
var __hasOwnProp = Object.prototype.hasOwnProperty;
|
|
6
|
-
var __export = (target, all) => {
|
|
7
|
-
for (var name in all)
|
|
8
|
-
__defProp(target, name, { get: all[name], enumerable: true });
|
|
9
|
-
};
|
|
10
|
-
var __copyProps = (to, from, except, desc) => {
|
|
11
|
-
if (from && typeof from === "object" || typeof from === "function") {
|
|
12
|
-
for (let key of __getOwnPropNames(from))
|
|
13
|
-
if (!__hasOwnProp.call(to, key) && key !== except)
|
|
14
|
-
__defProp(to, key, { get: () => from[key], enumerable: !(desc = __getOwnPropDesc(from, key)) || desc.enumerable });
|
|
15
|
-
}
|
|
16
|
-
return to;
|
|
17
|
-
};
|
|
18
|
-
var __toCommonJS = (mod) => __copyProps(__defProp({}, "__esModule", { value: true }), mod);
|
|
19
|
-
var strongly_exports = {};
|
|
20
|
-
__export(strongly_exports, {
|
|
21
|
-
stronglyExtractHast: () => stronglyExtractHast
|
|
22
|
-
});
|
|
23
|
-
module.exports = __toCommonJS(strongly_exports);
|
|
24
|
-
var import_hast_util_select = require("hast-util-select");
|
|
25
|
-
var import_unist_util_filter = require("unist-util-filter");
|
|
26
|
-
const REMOVE_REGEX = /fallback|icon|header|nav|speechify-ignore|assist|-ad-|banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|assist|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i;
|
|
27
|
-
const stronglyExtractHast = (params) => {
|
|
28
|
-
const { hast } = params;
|
|
29
|
-
const selectors = ["article", "#article", "#content", "main", "div.container"];
|
|
30
|
-
let selectedElement = hast;
|
|
31
|
-
for (const selector of selectors) {
|
|
32
|
-
const element = (0, import_hast_util_select.select)(selector, hast);
|
|
33
|
-
if (element) {
|
|
34
|
-
selectedElement = element;
|
|
35
|
-
break;
|
|
36
|
-
}
|
|
37
|
-
}
|
|
38
|
-
const removedElements = (0, import_unist_util_filter.filter)(selectedElement, (node) => {
|
|
39
|
-
if (node.type === "comment") {
|
|
40
|
-
return false;
|
|
41
|
-
}
|
|
42
|
-
if (node.type !== "element") {
|
|
43
|
-
return true;
|
|
44
|
-
}
|
|
45
|
-
const element = node;
|
|
46
|
-
const elementName = `${element.tagName} ${element.properties.id} ${element.properties.className}`;
|
|
47
|
-
if (REMOVE_REGEX.test(elementName)) {
|
|
48
|
-
return false;
|
|
49
|
-
}
|
|
50
|
-
return true;
|
|
51
|
-
});
|
|
52
|
-
return removedElements;
|
|
53
|
-
};
|
|
54
|
-
// Annotate the CommonJS export names for ESM import in node:
|
|
55
|
-
0 && (module.exports = {
|
|
56
|
-
stronglyExtractHast
|
|
57
|
-
});
|
|
@@ -1,34 +0,0 @@
|
|
|
1
|
-
// src/extract-hast/strongly.ts
|
|
2
|
-
import { select } from "hast-util-select";
|
|
3
|
-
import { filter } from "unist-util-filter";
|
|
4
|
-
var REMOVE_REGEX = /fallback|icon|header|nav|speechify-ignore|assist|-ad-|banner|combx|comment|com-|contact|foot|footer|footnote|gdpr|masthead|media|meta|outbrain|promo|related|scroll|share|assist|shoutbox|sidebar|skyscraper|sponsor|shopping|tags|tool|widget/i;
|
|
5
|
-
var stronglyExtractHast = (params) => {
|
|
6
|
-
const { hast } = params;
|
|
7
|
-
const selectors = ["article", "#article", "#content", "main", "div.container"];
|
|
8
|
-
let selectedElement = hast;
|
|
9
|
-
for (const selector of selectors) {
|
|
10
|
-
const element = select(selector, hast);
|
|
11
|
-
if (element) {
|
|
12
|
-
selectedElement = element;
|
|
13
|
-
break;
|
|
14
|
-
}
|
|
15
|
-
}
|
|
16
|
-
const removedElements = filter(selectedElement, (node) => {
|
|
17
|
-
if (node.type === "comment") {
|
|
18
|
-
return false;
|
|
19
|
-
}
|
|
20
|
-
if (node.type !== "element") {
|
|
21
|
-
return true;
|
|
22
|
-
}
|
|
23
|
-
const element = node;
|
|
24
|
-
const elementName = `${element.tagName} ${element.properties.id} ${element.properties.className}`;
|
|
25
|
-
if (REMOVE_REGEX.test(elementName)) {
|
|
26
|
-
return false;
|
|
27
|
-
}
|
|
28
|
-
return true;
|
|
29
|
-
});
|
|
30
|
-
return removedElements;
|
|
31
|
-
};
|
|
32
|
-
export {
|
|
33
|
-
stronglyExtractHast
|
|
34
|
-
};
|
|
@@ -1,15 +0,0 @@
|
|
|
1
|
-
import type { Nodes as Hast } from "hast";
|
|
2
|
-
export type ExtractPresets = "strongly" | "readability";
|
|
3
|
-
export type ExtractParams = {
|
|
4
|
-
hast: Hast;
|
|
5
|
-
lang?: string;
|
|
6
|
-
url?: string;
|
|
7
|
-
};
|
|
8
|
-
export type Extractor = ((param: ExtractParams) => Hast) | false | ExtractPresets;
|
|
9
|
-
export type Extracotrs = Extractor | Extractor[];
|
|
10
|
-
export declare const PRESET_EXTRACT_HAST: {
|
|
11
|
-
strongly: (params: ExtractParams) => Hast;
|
|
12
|
-
readability: (params: ExtractParams) => Hast;
|
|
13
|
-
};
|
|
14
|
-
export declare const DEFAULT_EXTRACT_HAST: Extracotrs;
|
|
15
|
-
export declare const extractHast: (params: ExtractParams, extractors?: Extracotrs) => Hast;
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"index.js","sourceRoot":"","sources":["../../../src/extract-hast/index.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,sBAAsB,EAAE,MAAM,eAAe,CAAC;AACvD,OAAO,EAAE,mBAAmB,EAAE,MAAM,YAAY,CAAC;AAOjD,MAAM,CAAC,MAAM,mBAAmB,GAAG;IAClC,QAAQ,EAAE,mBAAmB;IAC7B,WAAW,EAAE,sBAAsB;CACnC,CAAC;AAEF,MAAM,CAAC,MAAM,oBAAoB,GAAe,CAAC,aAAa,CAAC,CAAC;AAEhE,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,MAAqB,EAAE,aAAyB,oBAAoB,EAAQ,EAAE;IACzG,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IAC9B,MAAM,WAAW,GAAG,KAAK,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC;IAE1E,MAAM,SAAS,GACd,WAAW,CAAC,MAAM,CAAO,CAAC,GAAG,EAAE,SAAS,EAAE,EAAE;QAC3C,IAAI,SAAS,KAAK,KAAK,EAAE,CAAC;YACzB,OAAO,GAAG,CAAC;QACZ,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,QAAQ,IAAI,SAAS,IAAI,mBAAmB,EAAE,CAAC;YACvE,OAAO,mBAAmB,CAAC,SAAS,CAAC,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QAC5D,CAAC;QACD,IAAI,OAAO,SAAS,KAAK,UAAU,EAAE,CAAC;YACrC,OAAO,SAAS,CAAC,EAAE,IAAI,EAAE,GAAG,EAAE,IAAI,EAAE,CAAC,CAAC;QACvC,CAAC;QACD,MAAM,IAAI,KAAK,CAAC,sBAAsB,SAAS,EAAE,CAAC,CAAC;IACpD,CAAC,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAElB,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"readability.js","sourceRoot":"","sources":["../../../src/extract-hast/readability.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,SAAS,EAAE,MAAM,kBAAkB,CAAC;AACrD,OAAO,EAAE,QAAQ,IAAI,YAAY,EAAE,MAAM,qBAAqB,CAAC;AAC/D,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAE3C,OAAO,EAAE,UAAU,EAAE,YAAY,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAEhE,MAAM,cAAc,GAAG,CAAC,MAAM,EAAE,SAAS,EAAE,eAAe,EAAE,YAAY,EAAE,OAAO,EAAE,aAAa,EAAE,QAAQ,CAAC,CAAC;AAE5G,MAAM,OAAO,GAAG;IACf,MAAM,EAAE,kCAAkC;IAC1C,MAAM,EAAE,4CAA4C;IACpD,yBAAyB,EAAE,8BAA8B;IACzD,kBAAkB,EACjB,wRAAwR;IACzR,oBAAoB,EAAE,mDAAmD;CACzE,CAAC;AAEF,MAAM,cAAc,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,eAAe,EAAE,eAAe,EAAE,UAAU,EAAE,QAAQ,CAAC,CAAC;AAEvG,MAAM,cAAc,GAAG,CAAC,GAAG,EAAE,GAAG,EAAE,KAAK,EAAE,SAAS,EAAE,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,CAAC,CAAC;AAEzF,MAAM,mBAAmB,GAAG;IAC3B,EAAE,EAAE,GAAG;IACP,EAAE,EAAE,GAAG;CACP,CAAC;AAEF,yBAAyB;AACzB,UAAU;AACV,WAAW;AACX,KAAK;AAEL,MAAM,cAAc,GAAG,CAAC,IAAU,EAAE,EAAE;IACrC,OAAO,CAAC,CACP,CAAC,SAAS,EAAE,SAAS,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,IAAI,CAAC;QAC1C,CAAC,IAAI,CAAC,IAAI,KAAK,SAAS,IAAI,CAAC,QAAQ,EAAE,OAAO,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,KAAK,EAAE,OAAO,CAAC,CAAC,QAAQ,CAAC,IAAI,CAAC,OAAO,CAAC,CAAC,CACnH,CAAC;AACH,CAAC,CAAC;AAEF,MAAM,sBAAsB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC7C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,CAAC,OAAO,EAAE,KAAK,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAChD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yCAAyC;IACzC,IAAI,CAAC,QAAQ,EAAE,aAAa,CAAC,CAAC,IAAI,CAAC,CAAC,GAAG,EAAE,EAAE,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,CAAC,CAAC,EAAE,CAAC;QACtE,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,CAAC,SAAS,EAAE,EAAE,CAAC,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,SAAS,CAAC,CAAC,EAAE,CAAC;QAC7E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QAClC,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,KAAK,QAAQ,IAAI,OAAO,CAAC,UAAU,CAAC,YAAY,CAAC,EAAE,CAAC;QAC9E,OAAO,KAAK,CAAC;IACd,CAAC;IAED,yBAAyB;IACzB,IAAI,OAAO,CAAC,UAAU,CAAC,GAAG,KAAK,QAAQ,IAAI,YAAY,CAAC,OAAO,CAAC,UAAU,CAAC,QAAQ,EAAE,QAAQ,CAAC,EAAE,CAAC;QAChG,OAAO,KAAK,CAAC;IACd,CAAC;IACD,IAAI,OAAO,CAAC,MAAM,CAAC,IAAI,CAAC,WAAW,CAAC,OAAO,CAAC,CAAC,EAAE,CAAC;QAC/C,OAAO,KAAK,CAAC;IACd,CAAC;IAED,wBAAwB;IACxB,IAAI,OAAO,CAAC,UAAU,CAAC,IAAI,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,UAAU,CAAC,IAAc,CAAC,EAAE,CAAC;QAC3F,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,qBAAqB,GAAG,CAAC,IAAU,EAAE,EAAE;IAC5C,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,6BAA6B;IAC7B,IAAI,CAAC,MAAM,EAAE,SAAS,EAAE,MAAM,EAAE,SAAS,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,IAAI,EAAE,GAAG,CAAC,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC/G,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,KAAK,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC;IAEnC,IAAI,OAAO,CAAC,yBAAyB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACnD,OAAO,KAAK,CAAC;IACd,CAAC;IAED,6BAA6B;IAC7B,IAAI,OAAO,CAAC,kBAAkB,CAAC,IAAI,CAAC,KAAK,CAAC,IAAI,CAAC,OAAO,CAAC,oBAAoB,CAAC,IAAI,CAAC,KAAK,CAAC,EAAE,CAAC;QACzF,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,iBAAiB,GAAG,CAAC,IAAU,EAAE,KAAa,EAAE,EAAE;IACvD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;QAC7B,OAAO,IAAI,CAAC;IACb,CAAC;IACD,MAAM,OAAO,GAAG,IAAe,CAAC;IAEhC,IAAI,cAAc,CAAC,QAAQ,CAAC,OAAO,CAAC,OAAO,CAAC,EAAE,CAAC;QAC9C,OAAO,IAAI,CAAC;IACb,CAAC;IAED,IAAI,OAAO,CAAC,OAAO,KAAK,KAAK,IAAI,CAAC,OAAO,CAAC,UAAU,CAAC,GAAG,EAAE,CAAC;QAC1D,OAAO,KAAK,CAAC;IACd,CAAC;IAED,OAAO,IAAI,CAAC;AACb,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,sBAAsB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IACrE,MAAM,EAAE,IAAI,EAAE,IAAI,GAAG,IAAI,EAAE,GAAG,MAAM,CAAC;IACrC,MAAM,IAAI,GAAG,MAAM,CAAC,MAAM,EAAE,IAAI,CAAC,IAAI,IAAI,CAAC;IAE1C,MAAM,oBAAoB,GAAG,MAAM,CAAC,IAAI,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,cAAc,CAAC,IAAY,CAAC,CAAC,CAAC;IAClF,MAAM,wBAAwB,GAAG,oBAAoB,IAAI,YAAY,CAAC,oBAAoB,CAAC,CAAC;IAC5F,IAAI,CAAC,CAAC,oBAAoB,IAAI,wBAAwB,CAAC,EAAE,CAAC;QACzD,OAAO,IAAI,CAAC;IACb,CAAC;IAED,MAAM,WAAW,GAAG,MAAM,CAAC,oBAAoB,EAAE,CAAC,IAAI,EAAE,EAAE,CAAC,sBAAsB,CAAC,IAAY,CAAC,CAAC,CAAC;IACjG,MAAM,eAAe,GAAG,WAAW,CAAC,CAAC,CAAC,YAAY,CAAC,WAAW,CAAC,CAAC,CAAC,CAAC,EAAE,CAAC;IACrE,MAAM,CAAC,QAAQ,EAAE,QAAQ,CAAC,GACzB,eAAe,CAAC,MAAM,GAAG,wBAAwB,CAAC,MAAM,GAAG,CAAC,IAAI,eAAe,CAAC,MAAM,GAAG,IAAI;QAC5F,CAAC,CAAE,CAAC,WAAmB,EAAE,eAAe,CAAW;QACnD,CAAC,CAAE,CAAC,oBAA4B,EAAE,wBAAwB,CAAW,CAAC;IAExE,IAAI,aAAa,GAAG,IAAI,IAAI,mBAAmB,CAAC,CAAC,CAAC,mBAAmB,CAAC,IAAwC,CAAC,CAAC,CAAC,CAAC,GAAG,CAAC;IACtH,IAAI,QAAQ,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;QACrC,aAAa,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,GAAG,GAAG,CAAC,CAAC;IACpD,CAAC;IAED,IAAI,aAAa,GAAS,QAAQ,CAAC;IACnC,IAAI,aAAa,GAAG,QAAQ,CAAC;IAE7B,KAAK,MAAM,QAAQ,IAAI,cAAc,EAAE,CAAC;QACvC,MAAM,OAAO,GAAG,EAAE,IAAI,EAAE,MAAe,EAAE,QAAQ,EAAE,SAAS,CAAC,QAAQ,EAAE,WAAW,CAAC,EAAE,CAAC;QACtF,MAAM,WAAW,GAAG,YAAY,CAAC,OAAO,CAAC,CAAC;QAE1C,IAAI,WAAW,CAAC,MAAM,GAAG,EAAE,EAAE,CAAC;YAC7B,SAAS;QACV,CAAC;QAED,MAAM,KAAK,GAAG,SAAS,CAAC,GAAG,EAAE,OAAO,CAAC,CAAC;QACtC,MAAM,QAAQ,GAAG,KAAK,CAAC,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC;QAElE,MAAM,WAAW,GAAG,QAAQ,CAAC,MAAM,GAAG,WAAW,CAAC,MAAM,CAAC;QACzD,IAAI,WAAW,GAAG,GAAG,EAAE,CAAC;YACvB,SAAS;QACV,CAAC;QAED,IAAI,WAAW,CAAC,MAAM,GAAG,aAAa,EAAE,CAAC;YACxC,aAAa,GAAG,OAAO,CAAC;YACxB,aAAa,GAAG,WAAW,CAAC;YAC5B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,iBAAiB,GAAG,MAAM,CAAC,aAAa,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,CAAC,iBAAiB,CAAC,IAAY,EAAE,IAAI,CAAC,EAAE,CAAC;YAC5C,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,CAAC,qBAAqB,CAAC,IAAY,CAAC,EAAE,CAAC;YAC1C,OAAO,KAAK,CAAC;QACd,CAAC;QAED,OAAO,IAAI,CAAC;IACb,CAAC,CAAS,CAAC;IAEX,MAAM,SAAS,GACd,YAAY,CAAC,iBAAiB,CAAC,CAAC,MAAM,GAAG,aAAa,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,iBAAiB,CAAC,CAAC,CAAC,aAAa,CAAC;IACvG,OAAO,SAAS,CAAC;AAClB,CAAC,CAAC"}
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"strongly.js","sourceRoot":"","sources":["../../../src/extract-hast/strongly.ts"],"names":[],"mappings":"AACA,OAAO,EAAE,MAAM,EAAE,MAAM,kBAAkB,CAAC;AAC1C,OAAO,EAAE,MAAM,EAAE,MAAM,mBAAmB,CAAC;AAG3C,MAAM,YAAY,GACjB,iPAAiP,CAAC;AAEnP,MAAM,CAAC,MAAM,mBAAmB,GAAG,CAAC,MAAqB,EAAQ,EAAE;IAClE,MAAM,EAAE,IAAI,EAAE,GAAG,MAAM,CAAC;IACxB,MAAM,SAAS,GAAG,CAAC,SAAS,EAAE,UAAU,EAAE,UAAU,EAAE,MAAM,EAAE,eAAe,CAAC,CAAC;IAC/E,IAAI,eAAe,GAAG,IAAI,CAAC;IAC3B,KAAK,MAAM,QAAQ,IAAI,SAAS,EAAE,CAAC;QAClC,MAAM,OAAO,GAAG,MAAM,CAAC,QAAQ,EAAE,IAAI,CAAC,CAAC;QACvC,IAAI,OAAO,EAAE,CAAC;YACb,eAAe,GAAG,OAAO,CAAC;YAC1B,MAAM;QACP,CAAC;IACF,CAAC;IAED,MAAM,eAAe,GAAG,MAAM,CAAC,eAAe,EAAE,CAAC,IAAI,EAAE,EAAE;QACxD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,KAAK,CAAC;QACd,CAAC;QACD,IAAI,IAAI,CAAC,IAAI,KAAK,SAAS,EAAE,CAAC;YAC7B,OAAO,IAAI,CAAC;QACb,CAAC;QACD,MAAM,OAAO,GAAG,IAAe,CAAC;QAChC,MAAM,WAAW,GAAG,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,OAAO,CAAC,UAAU,CAAC,SAAS,EAAE,CAAC;QAClG,IAAI,YAAY,CAAC,IAAI,CAAC,WAAW,CAAC,EAAE,CAAC;YACpC,OAAO,KAAK,CAAC;QACd,CAAC;QACD,OAAO,IAAI,CAAC;IACb,CAAC,CAAC,CAAC;IAEH,OAAO,eAAuB,CAAC;AAChC,CAAC,CAAC"}
|
|
@@ -1 +0,0 @@
|
|
|
1
|
-
{"version":3,"file":"utils.js","sourceRoot":"","sources":["../../../src/extract-hast/utils.ts"],"names":[],"mappings":"AAEA,MAAM,CAAC,MAAM,WAAW,GAAG,CAAC,OAAgB,EAAE,EAAE,CAC/C,GAAG,OAAO,CAAC,OAAO,IAAI,OAAO,CAAC,UAAU,CAAC,EAAE,IAAI,UAAU,CAAC,OAAO,CAAC,CAAC,IAAI,CAAC,GAAG,CAAC,EAAE,CAAC;AAEhF,MAAM,CAAC,MAAM,UAAU,GAAG,CAAC,OAAgB,EAAE,EAAE;IAC9C,IAAI,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,UAAU,CAAC,SAAS,CAAC,EAAE,CAAC;QACjD,OAAO,OAAO,CAAC,UAAU,CAAC,SAAqB,CAAC;IACjD,CAAC;IACD,OAAO,EAAE,CAAC;AACX,CAAC,CAAC;AAEF,MAAM,CAAC,MAAM,YAAY,GAAG,CAAC,KAAc,EAAE,KAAa,EAAE,EAAE;IAC7D,IAAI,OAAO,KAAK,KAAK,QAAQ,EAAE,CAAC;QAC/B,OAAO,KAAK,CAAC,QAAQ,CAAC,KAAK,CAAC,CAAC;IAC9B,CAAC;IACD,OAAO,KAAK,CAAC;AACd,CAAC,CAAC"}
|
|
File without changes
|
|
File without changes
|