angular-cn/packages/compiler/test/i18n/extractor_merger_spec.ts

/**
 * @license
 * Copyright Google Inc. All Rights Reserved.
 *
 * Use of this source code is governed by an MIT-style license that can be
 * found in the LICENSE file at https://angular.io/license
 */

import {DEFAULT_INTERPOLATION_CONFIG, HtmlParser} from '@angular/compiler';

import {digest, serializeNodes as serializeI18nNodes} from '../../src/i18n/digest';
import {extractMessages, mergeTranslations} from '../../src/i18n/extractor_merger';
import * as i18n from '../../src/i18n/i18n_ast';
import {TranslationBundle} from '../../src/i18n/translation_bundle';
import * as html from '../../src/ml_parser/ast';
import {serializeNodes as serializeHtmlNodes} from '../ml_parser/ast_serializer_spec';

export function main() {
  describe('Extractor', () => {
    describe('elements', () => {
      it('should extract from elements', () => {
        expect(extract('<div i18n="m|d|e">text<span>nested</span></div>')).toEqual([
          [
            ['text', '<ph tag name="START_TAG_SPAN">nested</ph name="CLOSE_TAG_SPAN">'], 'm', 'd|e',
            ''
          ],
        ]);
      });

      it('should extract from attributes', () => {
        expect(
            extract(
                '<div i18n="m1|d1"><span i18n-title="m2|d2" title="single child">nested</span></div>'))
            .toEqual([
              [['<ph tag name="START_TAG_SPAN">nested</ph name="CLOSE_TAG_SPAN">'], 'm1', 'd1', ''],
              [['single child'], 'm2', 'd2', ''],
            ]);
      });

      it('should extract from attributes with id', () => {
        expect(
            extract(
                '<div i18n="m1|d1@@i1"><span i18n-title="m2|d2@@i2" title="single child">nested</span></div>'))
            .toEqual([
              [
                ['<ph tag name="START_TAG_SPAN">nested</ph name="CLOSE_TAG_SPAN">'], 'm1', 'd1',
                'i1'
              ],
              [['single child'], 'm2', 'd2', 'i2'],
            ]);
      });

      it('should extract from attributes without meaning and with id', () => {
        expect(
            extract(
                '<div i18n="d1@@i1"><span i18n-title="d2@@i2" title="single child">nested</span></div>'))
            .toEqual([
              [['<ph tag name="START_TAG_SPAN">nested</ph name="CLOSE_TAG_SPAN">'], '', 'd1', 'i1'],
              [['single child'], '', 'd2', 'i2'],
            ]);
      });

      it('should extract from attributes with id only', () => {
        expect(
            extract(
                '<div i18n="@@i1"><span i18n-title="@@i2" title="single child">nested</span></div>'))
            .toEqual([
              [['<ph tag name="START_TAG_SPAN">nested</ph name="CLOSE_TAG_SPAN">'], '', '', 'i1'],
              [['single child'], '', '', 'i2'],
            ]);
      });


      it('should extract from ICU messages', () => {
        expect(
            extract(
                '<div i18n="m|d">{count, plural, =0 { <p i18n-title i18n-desc title="title" desc="desc"></p>}}</div>'))
            .toEqual([
              [
                [
                  '{count, plural, =0 {[<ph tag name="START_PARAGRAPH"></ph name="CLOSE_PARAGRAPH">]}}'
                ],
                'm', 'd', ''
              ],
              [['title'], '', '', ''],
              [['desc'], '', '', ''],
            ]);
      });

      it('should not create a message for empty elements',
         () => { expect(extract('<div i18n="m|d"></div>')).toEqual([]); });

      it('should ignore implicit elements in translatable elements', () => {
        expect(extract('<div i18n="m|d"><p></p></div>', ['p'])).toEqual([
          [['<ph tag name="START_PARAGRAPH"></ph name="CLOSE_PARAGRAPH">'], 'm', 'd', '']
        ]);
      });
    });

    describe('blocks', () => {
      it('should extract from blocks', () => {
        expect(extract(`<!-- i18n: meaning1|desc1 -->message1<!-- /i18n -->
         <!-- i18n: desc2 -->message2<!-- /i18n -->
         <!-- i18n -->message3<!-- /i18n -->
         <!-- i18n: meaning4|desc4@@id4 -->message4<!-- /i18n -->
         <!-- i18n: @@id5 -->message5<!-- /i18n -->`))
            .toEqual([
              [['message1'], 'meaning1', 'desc1', ''], [['message2'], '', 'desc2', ''],
              [['message3'], '', '', ''], [['message4'], 'meaning4', 'desc4', 'id4'],
              [['message5'], '', '', 'id5']
            ]);
      });

      it('should ignore implicit elements in blocks', () => {
        expect(extract('<!-- i18n:m|d --><p></p><!-- /i18n -->', ['p'])).toEqual([
          [['<ph tag name="START_PARAGRAPH"></ph name="CLOSE_PARAGRAPH">'], 'm', 'd', '']
        ]);
      });


      it('should extract siblings', () => {
        expect(
            extract(
                `<!-- i18n -->text<p>html<b>nested</b></p>{count, plural, =0 {<span>html</span>}}{{interp}}<!-- /i18n -->`))
            .toEqual([
              [
                [
                  '{count, plural, =0 {[<ph tag name="START_TAG_SPAN">html</ph name="CLOSE_TAG_SPAN">]}}'
                ],
                '', '', ''
              ],
              [
                [
                  'text', '<ph tag name="START_PARAGRAPH">html, <ph tag' +
                      ' name="START_BOLD_TEXT">nested</ph name="CLOSE_BOLD_TEXT"></ph name="CLOSE_PARAGRAPH">',
                  '<ph icu name="ICU">{count, plural, =0 {[<ph tag' +
                      ' name="START_TAG_SPAN">html</ph name="CLOSE_TAG_SPAN">]}}</ph>',
                  '[<ph name="INTERPOLATION">interp</ph>]'
                ],
                '', '', ''
              ],
            ]);
      });

      it('should ignore other comments', () => {
        expect(extract(`<!-- i18n: meaning1|desc1@@id1 --><!-- other -->message1<!-- /i18n -->`))
            .toEqual([
              [['message1'], 'meaning1', 'desc1', 'id1'],
            ]);
      });

      it('should not create a message for empty blocks',
         () => { expect(extract(`<!-- i18n: meaning1|desc1 --><!-- /i18n -->`)).toEqual([]); });
    });

    describe('ICU messages', () => {
      it('should extract ICU messages from translatable elements', () => {
        // single message when ICU is the only children
        expect(extract('<div i18n="m|d">{count, plural, =0 {text}}</div>')).toEqual([
          [['{count, plural, =0 {[text]}}'], 'm', 'd', ''],
        ]);

        // single message when ICU is the only (implicit) children
        expect(extract('<div>{count, plural, =0 {text}}</div>', ['div'])).toEqual([
          [['{count, plural, =0 {[text]}}'], '', '', ''],
        ]);

        // one message for the element content and one message for the ICU
        expect(extract('<div i18n="m|d@@i">before{count, plural, =0 {text}}after</div>')).toEqual([
          [
            ['before', '<ph icu name="ICU">{count, plural, =0 {[text]}}</ph>', 'after'], 'm', 'd',
            'i'
          ],
          [['{count, plural, =0 {[text]}}'], '', '', ''],
        ]);
      });

      it('should extract ICU messages from translatable block', () => {
        // single message when ICU is the only children
        expect(extract('<!-- i18n:m|d -->{count, plural, =0 {text}}<!-- /i18n -->')).toEqual([
          [['{count, plural, =0 {[text]}}'], 'm', 'd', ''],
        ]);

        // one message for the block content and one message for the ICU
        expect(extract('<!-- i18n:m|d -->before{count, plural, =0 {text}}after<!-- /i18n -->'))
            .toEqual([
              [['{count, plural, =0 {[text]}}'], '', '', ''],
              [
                ['before', '<ph icu name="ICU">{count, plural, =0 {[text]}}</ph>', 'after'], 'm',
                'd', ''
              ],
            ]);
      });

      it('should not extract ICU messages outside of i18n sections',
         () => { expect(extract('{count, plural, =0 {text}}')).toEqual([]); });

      it('should ignore nested ICU messages', () => {
        expect(extract('<div i18n="m|d">{count, plural, =0 { {sex, select, male {m}} }}</div>'))
            .toEqual([
              [['{count, plural, =0 {[{sex, select, male {[m]}},  ]}}'], 'm', 'd', ''],
            ]);
      });

      it('should ignore implicit elements in non translatable ICU messages', () => {
        expect(extract(
                   '<div i18n="m|d@@i">{count, plural, =0 { {sex, select, male {<p>ignore</p>}}' +
                       ' }}</div>',
                   ['p']))
            .toEqual([[
              [
                '{count, plural, =0 {[{sex, select, male {[<ph tag name="START_PARAGRAPH">ignore</ph name="CLOSE_PARAGRAPH">]}},  ]}}'
              ],
              'm', 'd', 'i'
            ]]);
      });

      it('should ignore implicit elements in non translatable ICU messages', () => {
        expect(extract('{count, plural, =0 { {sex, select, male {<p>ignore</p>}} }}', ['p']))
            .toEqual([]);
      });
    });

    describe('attributes', () => {
      it('should extract from attributes outside of translatable sections', () => {
        expect(extract('<div i18n-title="m|d@@i" title="msg"></div>')).toEqual([
          [['msg'], 'm', 'd', 'i'],
        ]);
      });

      it('should extract from attributes in translatable elements', () => {
        expect(extract('<div i18n><p><b i18n-title="m|d@@i" title="msg"></b></p></div>')).toEqual([
          [
            ['<ph tag name="START_PARAGRAPH"><ph tag name="START_BOLD_TEXT"></ph' +
             ' name="CLOSE_BOLD_TEXT"></ph name="CLOSE_PARAGRAPH">'],
            '', '', ''
          ],
          [['msg'], 'm', 'd', 'i'],
        ]);
      });

      it('should extract from attributes in translatable blocks', () => {
        expect(extract('<!-- i18n --><p><b i18n-title="m|d" title="msg"></b></p><!-- /i18n -->'))
            .toEqual([
              [['msg'], 'm', 'd', ''],
              [
                ['<ph tag name="START_PARAGRAPH"><ph tag name="START_BOLD_TEXT"></ph' +
                 ' name="CLOSE_BOLD_TEXT"></ph name="CLOSE_PARAGRAPH">'],
                '', '', ''
              ],
            ]);
      });

      it('should extract from attributes in translatable ICUs', () => {
        expect(extract(`<!-- i18n -->{count, plural, =0 {<p><b i18n-title="m|d@@i" 
                 title="msg"></b></p>}}<!-- /i18n -->`))
            .toEqual([
              [['msg'], 'm', 'd', 'i'],
              [
                [
                  '{count, plural, =0 {[<ph tag name="START_PARAGRAPH"><ph tag' +
                  ' name="START_BOLD_TEXT"></ph name="CLOSE_BOLD_TEXT"></ph name="CLOSE_PARAGRAPH">]}}'
                ],
                '', '', ''
              ],
            ]);
      });

      it('should extract from attributes in non translatable ICUs', () => {
        expect(extract('{count, plural, =0 {<p><b i18n-title="m|d" title="msg"></b></p>}}'))
            .toEqual([
              [['msg'], 'm', 'd', ''],
            ]);
      });

      it('should not create a message for empty attributes',
         () => { expect(extract('<div i18n-title="m|d" title></div>')).toEqual([]); });
    });

    describe('implicit elements', () => {
      it('should extract from implicit elements', () => {
        expect(extract('<b>bold</b><i>italic</i>', ['b'])).toEqual([
          [['bold'], '', '', ''],
        ]);
      });

      it('should allow nested implicit elements', () => {
        let result: any[];

        expect(() => {
          result = extract('<div>outer<div>inner</div></div>', ['div']);
        }).not.toThrow();

        expect(result).toEqual([
          [['outer', '<ph tag name="START_TAG_DIV">inner</ph name="CLOSE_TAG_DIV">'], '', '', ''],
        ]);
      });

    });

    describe('implicit attributes', () => {
      it('should extract implicit attributes', () => {
        expect(extract('<b title="bb">bold</b><i title="ii">italic</i>', [], {'b': ['title']}))
            .toEqual([
              [['bb'], '', '', ''],
            ]);
      });
    });

    describe('errors', () => {
      describe('elements', () => {
        it('should report nested translatable elements', () => {
          expect(extractErrors(`<p i18n><b i18n></b></p>`)).toEqual([
            ['Could not mark an element as translatable inside a translatable section', '<b i18n>'],
          ]);
        });

        it('should report translatable elements in implicit elements', () => {
          expect(extractErrors(`<p><b i18n></b></p>`, ['p'])).toEqual([
            ['Could not mark an element as translatable inside a translatable section', '<b i18n>'],
          ]);
        });

        it('should report translatable elements in translatable blocks', () => {
          expect(extractErrors(`<!-- i18n --><b i18n></b><!-- /i18n -->`)).toEqual([
            ['Could not mark an element as translatable inside a translatable section', '<b i18n>'],
          ]);
        });
      });

      describe('blocks', () => {
        it('should report nested blocks', () => {
          expect(extractErrors(`<!-- i18n --><!-- i18n --><!-- /i18n --><!-- /i18n -->`)).toEqual([
            ['Could not start a block inside a translatable section', '<!--'],
            ['Trying to close an unopened block', '<!--'],
          ]);

        });

        it('should report unclosed blocks', () => {
          expect(extractErrors(`<!-- i18n -->`)).toEqual([
            ['Unclosed block', '<!--'],
          ]);
        });

        it('should report translatable blocks in translatable elements', () => {
          expect(extractErrors(`<p i18n><!-- i18n --><!-- /i18n --></p>`)).toEqual([
            ['Could not start a block inside a translatable section', '<!--'],
            ['Trying to close an unopened block', '<!--'],
          ]);
        });

        it('should report translatable blocks in implicit elements', () => {
          expect(extractErrors(`<p><!-- i18n --><!-- /i18n --></p>`, ['p'])).toEqual([
            ['Could not start a block inside a translatable section', '<!--'],
            ['Trying to close an unopened block', '<!--'],
          ]);
        });

        it('should report when start and end of a block are not at the same level', () => {
          expect(extractErrors(`<!-- i18n --><p><!-- /i18n --></p>`)).toEqual([
            ['I18N blocks should not cross element boundaries', '<!--'],
            ['Unclosed block', '<p>'],
          ]);

          expect(extractErrors(`<p><!-- i18n --></p><!-- /i18n -->`)).toEqual([
            ['I18N blocks should not cross element boundaries', '<!--'],
            ['Unclosed block', '<!--'],
          ]);
        });
      });
    });
  });

  describe('Merger', () => {
    describe('elements', () => {
      it('should merge elements', () => {
        const HTML = `<p i18n="m|d">foo</p>`;
        expect(fakeTranslate(HTML)).toEqual('<p>**foo**</p>');
      });

      it('should merge nested elements', () => {
        const HTML = `<div>before<p i18n="m|d">foo</p><!-- comment --></div>`;
        expect(fakeTranslate(HTML)).toEqual('<div>before<p>**foo**</p></div>');
      });

      it('should merge empty messages', () => {
        const HTML = `<div i18n>some element</div>`;
        const htmlNodes: html.Node[] = parseHtml(HTML);
        const messages: i18n.Message[] =
            extractMessages(htmlNodes, DEFAULT_INTERPOLATION_CONFIG, [], {}).messages;

        expect(messages.length).toEqual(1);
        const i18nMsgMap: {[id: string]: i18n.Node[]} = {};
        i18nMsgMap[digest(messages[0])] = [];
        const translations = new TranslationBundle(i18nMsgMap, null, digest);

        const output =
            mergeTranslations(htmlNodes, translations, DEFAULT_INTERPOLATION_CONFIG, [], {});
        expect(output.errors).toEqual([]);

        expect(serializeHtmlNodes(output.rootNodes).join('')).toEqual(`<div></div>`);
      });
    });

    describe('blocks', () => {
      it('should merge blocks', () => {
        const HTML = `before<!-- i18n --><p>foo</p><span><i>bar</i></span><!-- /i18n -->after`;
        expect(fakeTranslate(HTML))
            .toEqual(
                'before**[ph tag name="START_PARAGRAPH">foo[/ph name="CLOSE_PARAGRAPH">[ph tag' +
                ' name="START_TAG_SPAN">[ph tag name="START_ITALIC_TEXT">bar[/ph' +
                ' name="CLOSE_ITALIC_TEXT">[/ph name="CLOSE_TAG_SPAN">**after');
      });

      it('should merge nested blocks', () => {
        const HTML =
            `<div>before<!-- i18n --><p>foo</p><span><i>bar</i></span><!-- /i18n -->after</div>`;
        expect(fakeTranslate(HTML))
            .toEqual(
                '<div>before**[ph tag name="START_PARAGRAPH">foo[/ph name="CLOSE_PARAGRAPH">[ph' +
                ' tag name="START_TAG_SPAN">[ph tag name="START_ITALIC_TEXT">bar[/ph' +
                ' name="CLOSE_ITALIC_TEXT">[/ph name="CLOSE_TAG_SPAN">**after</div>');
      });
    });

    describe('attributes', () => {
      it('should merge attributes', () => {
        const HTML = `<p i18n-title="m|d" title="foo"></p>`;
        expect(fakeTranslate(HTML)).toEqual('<p title="**foo**"></p>');
      });

      it('should merge attributes with ids', () => {
        const HTML = `<p i18n-title="@@id" title="foo"></p>`;
        expect(fakeTranslate(HTML)).toEqual('<p title="**foo**"></p>');
      });

      it('should merge nested attributes', () => {
        const HTML = `<div>{count, plural, =0 {<p i18n-title title="foo"></p>}}</div>`;
        expect(fakeTranslate(HTML))
            .toEqual('<div>{count, plural, =0 {<p title="**foo**"></p>}}</div>');
      });

      it('should merge attributes without values', () => {
        const HTML = `<p i18n-title="m|d" title=""></p>`;
        expect(fakeTranslate(HTML)).toEqual('<p title=""></p>');
      });

      it('should merge empty attributes', () => {
        const HTML = `<div i18n-title title="some attribute">some element</div>`;
        const htmlNodes: html.Node[] = parseHtml(HTML);
        const messages: i18n.Message[] =
            extractMessages(htmlNodes, DEFAULT_INTERPOLATION_CONFIG, [], {}).messages;

        expect(messages.length).toEqual(1);
        const i18nMsgMap: {[id: string]: i18n.Node[]} = {};
        i18nMsgMap[digest(messages[0])] = [];
        const translations = new TranslationBundle(i18nMsgMap, null, digest);

        const output =
            mergeTranslations(htmlNodes, translations, DEFAULT_INTERPOLATION_CONFIG, [], {});
        expect(output.errors).toEqual([]);

        expect(serializeHtmlNodes(output.rootNodes).join(''))
            .toEqual(`<div title="">some element</div>`);
      });
    });
  });
}

function parseHtml(html: string): html.Node[] {
  const htmlParser = new HtmlParser();
  const parseResult = htmlParser.parse(html, 'extractor spec', true);
  if (parseResult.errors.length > 1) {
    throw new Error(`unexpected parse errors: ${parseResult.errors.join('\n')}`);
  }
  return parseResult.rootNodes;
}

function fakeTranslate(
    content: string, implicitTags: string[] = [],
    implicitAttrs: {[k: string]: string[]} = {}): string {
  const htmlNodes: html.Node[] = parseHtml(content);
  const messages: i18n.Message[] =
      extractMessages(htmlNodes, DEFAULT_INTERPOLATION_CONFIG, implicitTags, implicitAttrs)
          .messages;

  const i18nMsgMap: {[id: string]: i18n.Node[]} = {};

  messages.forEach(message => {
    const id = digest(message);
    const text = serializeI18nNodes(message.nodes).join('').replace(/</g, '[');
    i18nMsgMap[id] = [new i18n.Text(`**${text}**`, null)];
  });

  const translations = new TranslationBundle(i18nMsgMap, null, digest);

  const output = mergeTranslations(
      htmlNodes, translations, DEFAULT_INTERPOLATION_CONFIG, implicitTags, implicitAttrs);
  expect(output.errors).toEqual([]);

  return serializeHtmlNodes(output.rootNodes).join('');
}

function extract(
    html: string, implicitTags: string[] = [],
    implicitAttrs: {[k: string]: string[]} = {}): [string[], string, string][] {
  const result =
      extractMessages(parseHtml(html), DEFAULT_INTERPOLATION_CONFIG, implicitTags, implicitAttrs);

  if (result.errors.length > 0) {
    throw new Error(`unexpected errors: ${result.errors.join('\n')}`);
  }

  // clang-format off
  // https://github.com/angular/clang-format/issues/35
  return result.messages.map(
    message => [serializeI18nNodes(message.nodes), message.meaning, message.description, message.id]) as [string[], string, string][];
  // clang-format on
}

function extractErrors(
    html: string, implicitTags: string[] = [], implicitAttrs: {[k: string]: string[]} = {}): any[] {
  const errors =
      extractMessages(parseHtml(html), DEFAULT_INTERPOLATION_CONFIG, implicitTags, implicitAttrs)
          .errors;

  return errors.map((e): [string, string] => [e.msg, e.span.toString()]);
}