@comunica/actor-rdf-parse-html 3.0.1-alpha.45.0 → 3.0.1
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/README.md +1 -1
- package/components/context.jsonld +1 -1
- package/lib/ActorRdfParseHtml.js +1 -1
- package/lib/ActorRdfParseHtml.js.map +1 -1
- package/package.json +20 -20
package/README.md
CHANGED
|
@@ -24,7 +24,7 @@ After installing, this package can be added to your engine's configuration as fo
|
|
|
24
24
|
{
|
|
25
25
|
"@context": [
|
|
26
26
|
...
|
|
27
|
-
"https://linkedsoftwaredependencies.org/bundles/npm/@comunica/actor-rdf-parse-html/^3.0.0/components/context.jsonld"
|
|
27
|
+
"https://linkedsoftwaredependencies.org/bundles/npm/@comunica/actor-rdf-parse-html/^3.0.0/components/context.jsonld"
|
|
28
28
|
],
|
|
29
29
|
"actors": [
|
|
30
30
|
...
|
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"@context": [
|
|
3
|
-
"https://linkedsoftwaredependencies.org/bundles/npm/componentsjs/^
|
|
3
|
+
"https://linkedsoftwaredependencies.org/bundles/npm/componentsjs/^6.0.0/components/context.jsonld",
|
|
4
4
|
{
|
|
5
5
|
"npmd": "https://linkedsoftwaredependencies.org/bundles/npm/",
|
|
6
6
|
"carph": "npmd:@comunica/actor-rdf-parse-html/^3.0.0/",
|
package/lib/ActorRdfParseHtml.js
CHANGED
|
@@ -25,7 +25,7 @@ class ActorRdfParseHtml extends bus_rdf_parse_1.ActorRdfParseFixedMediaTypes {
|
|
|
25
25
|
}
|
|
26
26
|
async runHandle(action, mediaType, context) {
|
|
27
27
|
const data = new readable_stream_1.Readable({ objectMode: true });
|
|
28
|
-
data._read =
|
|
28
|
+
data._read = () => {
|
|
29
29
|
// Only initialize once
|
|
30
30
|
data._read = () => {
|
|
31
31
|
// Do nothing
|
|
@@ -1 +1 @@
|
|
|
1
|
-
{"version":3,"file":"ActorRdfParseHtml.js","sourceRoot":"","sources":["ActorRdfParseHtml.ts"],"names":[],"mappings":";;;
|
|
1
|
+
{"version":3,"file":"ActorRdfParseHtml.js","sourceRoot":"","sources":["ActorRdfParseHtml.ts"],"names":[],"mappings":";;;AACA,2DAEiC;AASjC,6CAAqC;AACrC,qDAA2C;AAE3C;;;GAGG;AACH,MAAa,iBAAkB,SAAQ,4CAA4B;IAOjE;;;;;;;;;;OAUG;IACH,YAAmB,IAA4B;QAC7C,KAAK,CAAC,IAAI,CAAC,CAAC;IACd,CAAC;IAEM,KAAK,CAAC,SAAS,CAAC,MAAuB,EAAE,SAAiB,EAAE,OAAuB;QAExF,MAAM,IAAI,GAAG,IAAI,0BAAQ,CAAC,EAAE,UAAU,EAAE,IAAI,EAAE,CAAC,CAAC;QAChD,IAAI,CAAC,KAAK,GAAG,GAAG,EAAE;YAChB,uBAAuB;YACvB,IAAI,CAAC,KAAK,GAAG,GAAG,EAAE;gBAChB,aAAa;YACf,CAAC,CAAC;YAEF,0BAA0B;YAC1B,IAAI,UAAU,GAAG,CAAC,CAAC;YACnB,SAAS,IAAI,CAAC,IAAc;gBAC1B,IAAI,CAAC,IAAI,CAAC,MAAM,EAAE,IAAI,CAAC,CAAC;YAC1B,CAAC;YACD,SAAS,KAAK,CAAC,QAAiB;gBAC9B,IAAI,CAAC,IAAI,CAAC,OAAO,EAAE,QAAQ,CAAC,CAAC;YAC/B,CAAC;YACD,SAAS,GAAG;gBACV,IAAI,EAAE,UAAU,KAAK,CAAC,EAAE,CAAC;oBACvB,IAAI,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC;gBAClB,CAAC;YACH,CAAC;YACD,MAAM,UAAU,GAAwB;gBACtC,OAAO,EAAE,MAAM,CAAC,QAAQ,EAAE,OAAO,IAAI,EAAE;gBACvC,OAAO;gBACP,IAAI;gBACJ,GAAG;gBACH,KAAK;gBACL,OAAO,EAAE,MAAM,CAAC,OAAO;aACxB,CAAC;YAEF,gCAAgC;YAChC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,eAAe,CAAC,OAAO,CAAC,UAAU,CAAC,CAAC;iBAClD,IAAI,CAAC,KAAK,EAAC,OAAO,EAAE,EAAE;gBACrB,UAAU,IAAI,OAAO,CAAC,MAAM,CAAC;gBAE7B,MAAM,kBAAkB,GAAyB,EAAE,CAAC;gBACpD,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;oBAC7B,MAAM,EAAE,iBAAiB,EAAE,GAAG,MAAM,MAAM,CAAC,KAAK,CAAC,GAAG,CAAC,UAAU,CAAC,CAAC;oBACjE,kBAAkB,CAAC,IAAI,CAAC,iBAAiB,CAAC,CAAC;gBAC7C,CAAC;gBAED,gBAAgB;gBAChB,MAAM,MAAM,GAAG,IAAI,oBAAM,CAAC;oBACxB,UAAU;wBACR,IAAI,CAAC;4BACH,KAAK,MAAM,iBAAiB,IAAI,kBAAkB,EAAE,CAAC;gCACnD,iBAAiB,CAAC,UAAU,EAAE,CAAC;4BACjC,CAAC;wBACH,CAAC;wBAAC,OAAO,MAAe,EAAE,CAAC;4BACzB,KAAK,CAAC,MAAM,CAAC,CAAC;wBAChB,CAAC;oBACH,CAAC;oBACD,KAAK;wBACH,IAAI,CAAC;4BACH,KAAK,MAAM,iBAAiB,IAAI,kBAAkB,EAAE,CAAC;gCACnD,iBAAiB,CAAC,KAAK,EAAE,CAAC;4BAC5B,CAAC;wBACH,CAAC;wBAAC,OAAO,MAAe,EAAE,CAAC;4BACzB,KAAK,CAAC,MAAM,CAAC,CAAC;wBAChB,CAAC;wBACD,GAAG,EAAE,CAAC;oBACR,CAAC;oBACD,SAAS,CAAC,IAAY,EAAE,UAAkC;wBACxD,IAAI,CAAC;4BACH,KAAK,MAAM,iBAAiB,IAAI,kBAAkB,EAAE,CAAC;gCACnD,iBAAiB,CAAC,SAAS,CAAC,IAAI,EAAE,UAAU,CAAC,CAAC;4BAChD,CAAC;wBACH,CAAC;wBAAC,OAAO,MAAe,EAAE,CAAC;4BACzB,KAAK,CAAC,MAAM,CAAC,CAAC;wBAChB,CAAC;oBACH,CAAC;oBACD,MAAM,CAAC,IAAY;wBACjB,IAAI,CAAC;4BACH,KAAK,MAAM,iBAAiB,IAAI,kBAAkB,EAAE,CAAC;gCACnD,iBAAiB,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;4BACjC,CAAC;wBACH,CAAC;wBAAC,OAAO,MAAe,EAAE,CAAC;4BACzB,KAAK,CAAC,MAAM,CAAC,CAAC;wBAChB,CAAC;oBACH,CAAC;iBACF,EAAE;oBACD,cAAc,EAAE,IAAI;oBACpB,oBAAoB,EAAE,IAAI;oBAC1B,OAAO,EAAE,KAAK;iBACf,CAAC,CAAC;gBAEH,wBAAwB;gBACxB,MAAM,CAAC,IAAI;qBACR,EAAE,CAAC,OAAO,EAAE,KAAK,CAAC;qBAClB,EAAE,CAAC,MAAM,EAAE,KAAK,CAAC,EAAE,CAAC,MAAM,CAAC,KAAK,CAAC,KAAK,CAAC,QAAQ,EAAE,CAAC,CAAC;qBACnD,EAAE,CAAC,KAAK,EAAE,GAAG,EAAE,CAAC,MAAM,CAAC,GAAG,EAAE,CAAC,CAAC;YACnC,CAAC,CAAC,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;QACpB,CAAC,CAAC;QAEF,OAAO,EAAE,IAAI,EAAE,CAAC;IAClB,CAAC;CACF;AAvHD,8CAuHC","sourcesContent":["import type { IActionRdfParse, IActorRdfParseFixedMediaTypesArgs, IActorRdfParseOutput } from '@comunica/bus-rdf-parse';\nimport {\n ActorRdfParseFixedMediaTypes,\n} from '@comunica/bus-rdf-parse';\nimport type {\n IActionRdfParseHtml,\n IActorRdfParseHtmlOutput,\n IHtmlParseListener,\n} from '@comunica/bus-rdf-parse-html';\nimport type { Actor, Bus, IActorTest } from '@comunica/core';\nimport type { IActionContext } from '@comunica/types';\nimport type * as RDF from '@rdfjs/types';\nimport { Parser } from 'htmlparser2';\nimport { Readable } from 'readable-stream';\n\n/**\n * A comunica HTML RDF Parse Actor.\n * It creates an HTML parser, and delegates its events via the bus-rdf-parse-html bus to other HTML parsing actors.\n */\nexport class ActorRdfParseHtml extends ActorRdfParseFixedMediaTypes {\n private readonly busRdfParseHtml: Bus<Actor<\n IActionRdfParseHtml,\nIActorTest,\nIActorRdfParseHtmlOutput\n>, IActionRdfParseHtml, IActorTest, IActorRdfParseHtmlOutput>;\n\n /**\n * @param args -\n * \\ @defaultNested {{\n * \"text/html\": 1.0,\n * \"application/xhtml+xml\": 0.9\n * }} mediaTypePriorities\n * \\ @defaultNested {{\n * \"text/html\": \"http://www.w3.org/ns/formats/HTML\",\n * \"application/xhtml+xml\": \"http://www.w3.org/ns/formats/HTML\"\n * }} mediaTypeFormats\n */\n public constructor(args: IActorRdfParseHtmlArgs) {\n super(args);\n }\n\n public async runHandle(action: IActionRdfParse, mediaType: string, context: IActionContext):\n Promise<IActorRdfParseOutput> {\n const data = new Readable({ objectMode: true });\n data._read = () => {\n // Only initialize once\n data._read = () => {\n // Do nothing\n };\n\n // Create callbacks action\n let endBarrier = 1;\n function emit(quad: RDF.Quad): void {\n data.emit('data', quad);\n }\n function error(subError: unknown): void {\n data.emit('error', subError);\n }\n function end(): void {\n if (--endBarrier === 0) {\n data.push(null);\n }\n }\n const htmlAction: IActionRdfParseHtml = {\n baseIRI: action.metadata?.baseIRI ?? '',\n context,\n emit,\n end,\n error,\n headers: action.headers,\n };\n\n // Register html parse listeners\n Promise.all(this.busRdfParseHtml.publish(htmlAction))\n .then(async(outputs) => {\n endBarrier += outputs.length;\n\n const htmlParseListeners: IHtmlParseListener[] = [];\n for (const output of outputs) {\n const { htmlParseListener } = await output.actor.run(htmlAction);\n htmlParseListeners.push(htmlParseListener);\n }\n\n // Create parser\n const parser = new Parser({\n onclosetag() {\n try {\n for (const htmlParseListener of htmlParseListeners) {\n htmlParseListener.onTagClose();\n }\n } catch (error_: unknown) {\n error(error_);\n }\n },\n onend() {\n try {\n for (const htmlParseListener of htmlParseListeners) {\n htmlParseListener.onEnd();\n }\n } catch (error_: unknown) {\n error(error_);\n }\n end();\n },\n onopentag(name: string, attributes: Record<string, string>) {\n try {\n for (const htmlParseListener of htmlParseListeners) {\n htmlParseListener.onTagOpen(name, attributes);\n }\n } catch (error_: unknown) {\n error(error_);\n }\n },\n ontext(text: string) {\n try {\n for (const htmlParseListener of htmlParseListeners) {\n htmlParseListener.onText(text);\n }\n } catch (error_: unknown) {\n error(error_);\n }\n },\n }, {\n decodeEntities: true,\n recognizeSelfClosing: true,\n xmlMode: false,\n });\n\n // Push stream to parser\n action.data\n .on('error', error)\n .on('data', chunk => parser.write(chunk.toString()))\n .on('end', () => parser.end());\n }).catch(error);\n };\n\n return { data };\n }\n}\n\nexport interface IActorRdfParseHtmlArgs extends IActorRdfParseFixedMediaTypesArgs {\n /* eslint-disable max-len */\n /**\n * The RDF Parse HTML bus for fetching HTML listeners\n * @default {<npmd:@comunica/bus-rdf-parse-html/^3.0.0/components/ActorRdfParseHtml.jsonld#ActorRdfParseHtml_default_bus>}\n */\n busRdfParseHtml: Bus<Actor<IActionRdfParseHtml, IActorTest, IActorRdfParseHtmlOutput>, IActionRdfParseHtml, IActorTest, IActorRdfParseHtmlOutput>;\n /* eslint-enable max-len */\n}\n"]}
|
package/package.json
CHANGED
|
@@ -1,49 +1,49 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@comunica/actor-rdf-parse-html",
|
|
3
|
-
"version": "3.0.1
|
|
3
|
+
"version": "3.0.1",
|
|
4
4
|
"description": "A html rdf-parse actor",
|
|
5
5
|
"lsd:module": true,
|
|
6
|
-
"
|
|
7
|
-
"
|
|
6
|
+
"license": "MIT",
|
|
7
|
+
"homepage": "https://comunica.dev/",
|
|
8
8
|
"repository": {
|
|
9
9
|
"type": "git",
|
|
10
10
|
"url": "https://github.com/comunica/comunica.git",
|
|
11
11
|
"directory": "packages/actor-rdf-parse-html"
|
|
12
12
|
},
|
|
13
|
-
"
|
|
14
|
-
"
|
|
13
|
+
"bugs": {
|
|
14
|
+
"url": "https://github.com/comunica/comunica/issues"
|
|
15
15
|
},
|
|
16
|
-
"sideEffects": false,
|
|
17
16
|
"keywords": [
|
|
18
17
|
"comunica",
|
|
19
18
|
"actor",
|
|
20
19
|
"rdf-parse",
|
|
21
20
|
"html"
|
|
22
21
|
],
|
|
23
|
-
"
|
|
24
|
-
"
|
|
25
|
-
|
|
22
|
+
"sideEffects": false,
|
|
23
|
+
"main": "lib/index.js",
|
|
24
|
+
"typings": "lib/index",
|
|
25
|
+
"publishConfig": {
|
|
26
|
+
"access": "public"
|
|
26
27
|
},
|
|
27
|
-
"homepage": "https://comunica.dev/",
|
|
28
28
|
"files": [
|
|
29
29
|
"components",
|
|
30
30
|
"lib/**/*.d.ts",
|
|
31
31
|
"lib/**/*.js",
|
|
32
32
|
"lib/**/*.js.map"
|
|
33
33
|
],
|
|
34
|
-
"dependencies": {
|
|
35
|
-
"@comunica/bus-rdf-parse": "3.0.1-alpha.45.0",
|
|
36
|
-
"@comunica/bus-rdf-parse-html": "3.0.1-alpha.45.0",
|
|
37
|
-
"@comunica/core": "3.0.1-alpha.45.0",
|
|
38
|
-
"@comunica/types": "3.0.1-alpha.45.0",
|
|
39
|
-
"@rdfjs/types": "*",
|
|
40
|
-
"htmlparser2": "^9.0.0",
|
|
41
|
-
"readable-stream": "^4.4.2"
|
|
42
|
-
},
|
|
43
34
|
"scripts": {
|
|
44
35
|
"build": "npm run build:ts && npm run build:components",
|
|
45
36
|
"build:ts": "node \"../../node_modules/typescript/bin/tsc\"",
|
|
46
37
|
"build:components": "componentsjs-generator"
|
|
47
38
|
},
|
|
48
|
-
"
|
|
39
|
+
"dependencies": {
|
|
40
|
+
"@comunica/bus-rdf-parse": "^3.0.1",
|
|
41
|
+
"@comunica/bus-rdf-parse-html": "^3.0.1",
|
|
42
|
+
"@comunica/core": "^3.0.1",
|
|
43
|
+
"@comunica/types": "^3.0.1",
|
|
44
|
+
"@rdfjs/types": "*",
|
|
45
|
+
"htmlparser2": "^9.0.0",
|
|
46
|
+
"readable-stream": "^4.4.2"
|
|
47
|
+
},
|
|
48
|
+
"gitHead": "9c0b537cd927d29590c287fe882185d88a65461b"
|
|
49
49
|
}
|