refactor: tidy up translation-parser

tools/challenge-parser/translation-parser/index.js

@@ -0,0 +1,145 @@
+const { isEmpty, cloneDeep } = require('lodash');
+
+exports.translateComments = (text, lang, dict, codeLang) => {
+  const knownComments = Object.keys(dict);
+  const config = { knownComments, dict, lang };
+  switch (codeLang) {
+    case 'js':
+    case 'jsx':
+      return transMultiline(transInline(text, config), config);
+    case 'html':
+      return transScript(transHTML(transCSS(text, config), config), config);
+    default:
+      return text;
+  }
+};
+
+exports.translateCommentsInChallenge = (challenge, lang, dict) => {
+  const challClone = cloneDeep(challenge);
+  if (!challClone.files) {
+    console.warn(`Challenge ${challClone.title} has no comments to translate`);
+  } else {
+    Object.keys(challClone.files).forEach(key => {
+      if (challClone.files[key].contents) {
+        challClone.files[key].contents = this.translateComments(
+          challenge.files[key].contents,
+          lang,
+          dict,
+          challClone.files[key].ext
+        );
+      }
+    });
+  }
+  return challClone;
+};
+
+exports.mergeChallenges = (engChal, transChal) => {
+  const hasTests =
+    (engChal.tests && transChal.tests) ||
+    (engChal.question && transChal.question);
+  const challenge = {
+    ...engChal,
+    description: transChal.description,
+    instructions: transChal.instructions,
+    originalTitle: engChal.title,
+    // TODO: throw in production?
+    title: isEmpty(transChal.title) ? engChal.title : transChal.title,
+    forumTopicId: transChal.forumTopicId
+  };
+  if (!hasTests)
+    throw Error(
+      `Both challenges must have tests or questions.
+      title: ${engChal.title}
+      translated title: ${transChal.title}`
+    );
+  // TODO: this should break the build when we go to production, but
+  // not for testing.
+  if (transChal.tests && transChal.tests.length !== engChal.tests.length) {
+    console.error(
+      `Challenges in both languages must have the same number of tests.
+    title: ${engChal.title}
+    translated title: ${transChal.title}`
+    );
+    return challenge;
+  }
+
+  // throw Error(
+  //   `Challenges in both languages must have the same number of tests.
+  // title: ${engChal.title}
+  // translated title: ${transChal.title}`
+  // );
+
+  if (transChal.tests) {
+    const translatedTests =
+      engChal.challengeType === 7
+        ? transChal.tests.map(({ title }, i) => ({
+            title,
+            id: engChal.tests[i].id
+          }))
+        : transChal.tests.map(({ text }, i) => ({
+            text,
+            testString: engChal.tests[i].testString
+          }));
+    challenge.tests = translatedTests;
+  } else {
+    challenge.question = transChal.question;
+  }
+
+  // certificates do not have forumTopicIds
+  if (challenge.challengeType === 7) delete challenge.forumTopicId;
+  return challenge;
+};
+
+// bare urls could be interpreted as comments, so we have to lookbehind for
+// http:// or https://
+function transInline(text, config) {
+  return translateGeneric(text, config, '((?<!https?:)//\\s*)', '(\\s*$)');
+}
+
+function transMultiline(text, config) {
+  return translateGeneric(text, config, '(/\\*\\s*)', '(\\s*\\*/)');
+}
+
+// CSS has to be handled separately since it is looking for comments inside tags
+function transCSS(text, config) {
+  const regex = /<style>.*?<\/style>/gms;
+  const matches = text.matchAll(regex);
+
+  for (const [match] of matches) {
+    text = text.replace(match, transMultiline(match, config));
+  }
+  return text;
+}
+
+function transScript(text, config) {
+  const regex = /<script>.*?<\/script>/gms;
+  const matches = text.matchAll(regex);
+
+  for (const [match] of matches) {
+    text = text.replace(
+      match,
+      transMultiline(transInline(match, config), config)
+    );
+  }
+  return text;
+}
+
+function transHTML(text, config) {
+  return translateGeneric(text, config, '(<!--\\s*)', '(\\s*-->)');
+}
+
+function translateGeneric(text, config, regexBefore, regexAfter) {
+  const { knownComments, dict, lang } = config;
+  const regex = new RegExp(regexBefore + '(.*?)' + regexAfter, 'gms');
+  const matches = text.matchAll(regex);
+
+  for (const [match, before, comment, after] of matches) {
+    if (knownComments.includes(comment)) {
+      text = text.replace(match, `${before}${dict[comment][lang]}${after}`);
+    } else if (comment.trim()) {
+      throw `${comment} does not appear in the comment dictionary`;
+    }
+  }
+
+  return text;
+}