nodeplusplus / xregex-parser / 180

Build # 180

Build Type

circleci

Committed by Unknown Committer

Commit Message

Unknown Commit Message

Pull Request Pull Request #17: Bump json5 from 2.1.2 to 2.2.3

Run Details

160 of 160 branches covered (100.0%)

Branch coverage included in aggregate %.

171 of 171 relevant lines covered (100.0%)

8.82 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

100.0

/src/engines/HTML.ts

import _ from "lodash";
import cheerio from "cheerio";
import { injectable, inject } from "inversify";
import { ILogger } from "@nodeplusplus/xregex-logger";

import {
  XParserReserved,
  IXParser,
  IXParserExecOpts,
  IXParserSchemaItem,
} from "../types/XParser";

@injectable()
export class HTMLParser implements IXParser {
  @inject("LOGGER") private logger!: ILogger;

  public async start() {
    this.logger.info("XPARSER:ENGINE.HTML.STARTED");
  }
  public async stop() {
    this.logger.info("XPARSER:ENGINE.HTML.STOPPED");
  }

  public async exec<T = any>(
    data: any,
    opts: IXParserExecOpts
  ): Promise<T | T[]> {
    if (!data || typeof data !== "string") return "" as any;

    if (!opts.schema?.ref && !opts.schema?.selector) {
      throw new Error("XPARSER:ENGINE.HTML.NO_REF_AND_SELECTOR");
    }
    const schema = opts.schema as IXParserSchemaItem;

    const ref = opts.ref;
    // 1. Return reference data
    if (schema.ref) return _.get(opts.ref, schema.ref);

    // 2. Return root data
    if (schema.selector === XParserReserved.ROOT_SELECTOR) return data as any;

    const $ = cheerio.load(data);

    // 3.1 Remove redundant elements
    if (schema.unselector) {
      const unselector = _.template(schema.unselector)({ ...ref });
      $(unselector).remove();
    }
    // 3.2 We want to get raw HTML intead of property
    if (!schema.prop) {
      return $(schema.selector)
        .toArray()
        .map((e: any) => $.html(e)) as any[];
    }
    // 3.3 Count items;
    if (schema.prop === XParserReserved.PROP_LENGTH) {
      return $(schema.selector).length as any;
    }

    const dom = $.root().find(schema.selector as string);
    if (!dom.length) return "" as any;

    // 3.4 Fix missing some values of .prop()
    if (schema.prop === XParserReserved.PROP_TEXT) return dom.text() as any;

    return (dom.prop(schema.prop) || "") as any;
  }

  public async clean<T = any>(data: any, selectors: string[]): Promise<T> {
    if (!Array.isArray(selectors) || !selectors.length) return data;
    if (!data || typeof data !== "string") return data;

    const $ = cheerio.load(data);
    $(selectors.join(",")).remove();
    return $.html() as any;
  }
}

1	import _ from "lodash";	3✔
2	import cheerio from "cheerio";	3✔
3	import { injectable, inject } from "inversify";	3✔
4	import { ILogger } from "@nodeplusplus/xregex-logger";
5
6	import {	3✔
7	XParserReserved,
8	IXParser,
9	IXParserExecOpts,
10	IXParserSchemaItem,
11	} from "../types/XParser";
12
13	@injectable()
14	export class HTMLParser implements IXParser {	3✔
15	@inject("LOGGER") private logger!: ILogger;	3✔
16
17	public async start() {
18	this.logger.info("XPARSER:ENGINE.HTML.STARTED");	4✔
19	}
20	public async stop() {
21	this.logger.info("XPARSER:ENGINE.HTML.STOPPED");	4✔
22	}
23
24	public async exec<T = any>(
25	data: any,
26	opts: IXParserExecOpts
27	): Promise<T \| T[]> {
28	if (!data \|\| typeof data !== "string") return "" as any;	15✔
29
30	if (!opts.schema?.ref && !opts.schema?.selector) {	11✔
31	throw new Error("XPARSER:ENGINE.HTML.NO_REF_AND_SELECTOR");	2✔
32	}
33	const schema = opts.schema as IXParserSchemaItem;	9✔
34
35	const ref = opts.ref;	9✔
36	// 1. Return reference data
37	if (schema.ref) return _.get(opts.ref, schema.ref);	9✔
38
39	// 2. Return root data
40	if (schema.selector === XParserReserved.ROOT_SELECTOR) return data as any;	8✔
41
42	const $ = cheerio.load(data);	7✔
43
44	// 3.1 Remove redundant elements
45	if (schema.unselector) {	7✔
46	const unselector = _.template(schema.unselector)({ ...ref });	1✔
47	$(unselector).remove();	1✔
48	}
49	// 3.2 We want to get raw HTML intead of property
50	if (!schema.prop) {	7✔
51	return $(schema.selector)	1✔
52	.toArray()
53	.map((e: any) => $.html(e)) as any[];	1✔
54	}
55	// 3.3 Count items;
56	if (schema.prop === XParserReserved.PROP_LENGTH) {	6✔
57	return $(schema.selector).length as any;	1✔
58	}
59
60	const dom = $.root().find(schema.selector as string);	5✔
61	if (!dom.length) return "" as any;	5✔
62
63	// 3.4 Fix missing some values of .prop()
64	if (schema.prop === XParserReserved.PROP_TEXT) return dom.text() as any;	3✔
65
66	return (dom.prop(schema.prop) \|\| "") as any;	2✔
67	}
68
69	public async clean<T = any>(data: any, selectors: string[]): Promise<T> {
70	if (!Array.isArray(selectors) \|\| !selectors.length) return data;	7✔
71	if (!data \|\| typeof data !== "string") return data;	4✔
72
73	const $ = cheerio.load(data);	2✔
74	$(selectors.join(",")).remove();	2✔
75	return $.html() as any;	2✔
76	}
77	}

nodeplusplus / xregex-parser / 180

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous