yoastseo-dep/CHANGELOG.md

Yoast clientside page analysis

# Change Log

All notable changes to this project will be documented in this file. Releases without a changelog entry contain only minor changes that are irrelevant for users of this library.
We will follow [Semantic Versioning](http://semver.org/) from version 2 and onwards.

## Future Release
### Enhancements
* Adds a missing array entry that caused time-related words in Dutch, English, Indonesian, Russian, and Spanish to not be counted as function words.
* Registers the whole web worker with the scope.
* Replaces assessment shortlinks that are passed to the assessment constructors in the product assessors with variables.
* Adds functionality for the webworker to also set custom options when loading the custom assessors for the SEO analysis (regular/cornerstone), the SEO analysis for related keyphrases (regular/cornerstone), and the readability analysis (regular/cornerstone).
* Adds custom config for TitleWidth assessment in all SEO assessors so that short title is not penalized with a bad score.
* Expands Russian irregular nouns exception list.
* Adds stemming rule for the most common perfective verbs.
* Implements the passive voice assessment for Farsi.
* Implements the consecutive sentences assessment for Farsi to ensure variety in a text.
* Changes the maximum recommended sentence length in Farsi to 25 words.
* Adds Farsi transition words and activates the transition words assessment.
* Expands the list of function words for Farsi.
* Removes the Farsi feature flag.
* Removes hyphens from a keyphrase for the Keyphrase in slug assessment. Also removes the functionality that would look for the keyphrase in the unparsed slug if it was not found in the parsed slug, as it is made redundant by the new functionality.
* Removes feature flag for Greek support to prepare Greek for release.
* Adds a few entries to the list of `multipleWords` for French transition words.
* Removes feature flag for Japanese support to prepare Japanese for release.
* Adds Japanese configuration for meta description length.
* Removes apostrophe before Turkish words are stemmed.
* Improves feedback strings for the Keyphrase Length assessment by making them more explicit.

### Bugfixes
* Moves the initialization of Keyphrase distribution assessment inside `AnalysisWebWorker.js`.
* Counts relative fragment links (`#some-id`) as a link to the same page.
* Removes two items from the list of morphological passive suffixes for Greek due to their overlap with bigger word groups causing false positives detection.
* Removes spaces before and after Japanese full stops when sanitizing strings.
* Removes URLs from Japanese texts before computing text length in the `countCharacters` function.
* Fixes a bug where the results of the SEO analysis of the focus keyphrase were removed when more than two related keyphrases were analyzed.

### Non user facing
* Fixes a bug where an invalid browser field in package.json would cause Vite to throw an error when building.
* Fixes a typo that caused the `findList` research to not be used in the `ListAssessment`.
* Adds a feature-flag for Farsi support.
* Makes Consecutive sentence, Passive voice and Transition words assessments available for Farsi when the Feature flag is enabled.
* Adds a missing plural variant of a feedback string for the keyphrase length assessment in product pages.
* Replaces Images assessment `countVideos` value that is passed to the assessment constructor in the product assessors with a variable.
* Improves the Farsi passive voice assessment by expanding the list of passive verb forms.
* Adds missing researcher parameter in assessors.
* Updates `README.md` for Farsi support.
* Adjusts Farsi full text tests papers.
* Removes the Farsi feature flag.
* Adjusts the English papers in full text tests for product pages.
* Adds full text tests for product pages and collection pages assessments.
* Implements the passive voice assessment for Farsi.
* Implements the consecutive sentences assessment for Farsi to ensure variety in a text.
* Changes the maximum recommended sentence length in Farsi to 25 words.
* Adds Farsi transition words and activates the transition words assessment.
* Fixes assessment result links that omit tracking parameters.
* Makes it possible to tokenize sentences in Japanese.
* Adds Japanese folder and Researcher.
* Adds a feature-flag for Japanese support.
* Implements the consecutive sentences assessment for Greek to ensure variety in a text.
* Adds Greek transition words and activates the transition words assessment.
* Adds a language folder and Researcher for Greek.
* Moves the initialization of Keyphrase distribution assessment inside `AnalysisWebWorker.js`.
* Makes passive voice and function words in keyphrase assessments available for Greek.
* Adds a paper and unit tests to test the assessments for Greek and edits a typo in the Greek morphologyData file path in `getMorphologyData.js`.
* Adds a `getContentWords` helper and adds function words list for Japanese.
* Updates README.md for Greek support.
* Creates a helper for splitting a Japanese text into words using the TinySegmenter package, and adds Japanese punctuation to the removePunctuation helper. Also changes the global Jest testing environment from `node` to `jsdom`.
* Adds external stemmer for Greek.
* Improves keyphrase recognition in Greek by filtering out function words such as `στον`, `τρίτος`, `τέτοιους`, `ποιανής`.
* Implements passive voice assessment for Greek.
* Adapts `findTransitionWords` research for Japanese.
* Adds Japanese transition words and activates the transition words assessment.
* Adapts `getKeywordDensity` research for Japanese.
* Adapts `functionWordsInKeyphrase` research for Japanese.
* Creates Japanese helper file for matching word in text.
* Changes the helper to match word in text for Japanese in `altTagCount`.
* Adds a helper to create word forms for Japanese.
* Adapts `matchKeywordInSubheadings` research for Japanese.
* Adapts `keyphraseDistribution` research for Japanese and adds Japanese topic length criteria config.
* Adapts `getSubheadingTextLengths` for Japanese.
* Adds the option to pass a language-specific `getWords` helper to the `getSentenceBeginnings` research.
* Adds Japanese sentence beginning exceptions.
* Activates the consecutive sentences assessment for Japanese.
* Adapts `getParagraphLength` research for Japanese.
* Adapts the `sentencesCountFromText` research for Japanese.
* Adapts the `getProminentWordsForInsights` research for Japanese.
* Fixes the prominent words functionality in content-analysis app.
* Adds unit tests that cover Japanese strings in `metaDescriptionLengthSpec` research file.
* Adds the option to pass a custom matchWordInText helper and a helper for matching multiword keyphrases in the findKeywordInPageTitle research.
* Creates a custom helper for finding exact matches of multiword keyphrases in title in Japanese.
* Adds Japanese quotes to the function that checks whether a keyphrase is enclosed in double quotes in the findKeywordInPageTitle research.
* Adds a custom `morphology` research for the Japanese language and edits an HTML tag typo in `matchStringWithRegexSpec.js`.
* Adapts `getProminentWordsForInternalLinking` research for Japanese.
* Adds `customGetStemmer` and `determineStem` helper for Japanese.
* Cleans up spec descriptions for assessing text and paragraph length.
* Updates `README.md` for Japanese support.
* Adapt the `getKeywordDensity` research so that a `getWordsCustomHelper` is used when available and adds unitests for Japanese.
* Adapts `getLinkStatistics` research for Japanese.
* Improves the way keyphrase enclosed in double quotes is matched in Keyphrase in title assessment for Japanese.
* Adapts Keyphrase in title assessment for Japanese and adds a custom Japanese `findKeywordInPageTitle` research file.
* Moves `processExactMatchRequest` helper from `findKeywordInPageTitle.js` to a separate helper file.
* Adds `position` of the matched keyphrase in the returned object of `matchTextWithArray.js` and `findKeywordFormsInString.js`.
* Removes codes that are not relevant anymore which were introduced in [this PR](https://github.com/Yoast/wordpress-seo/pull/17577).
* Adapts subheading distribution assessment for Japanese.
* Fixes highlights for sentence length assessment for other editors.
* Adapts the keyphrase length assessment for Japanese so that it uses custom scoring config and counts keyphrase length in characters instead of words.
* Disables the keyphrase in slug assessment for Japanese.
* Adds custom `keyphraseLength` research for Japanese.
* Adapts the sentence length assessment for Japanese.
* Adapts the text length assessment for Japanese.
* Adapts the Paragraph length assessment for Japanese and adds a Japanese specific configuration file for the paragraph length assessment in non-product and product pages.
* Changes the applicability criteria of the transition words, subheading distribution, and keyphrase density assessments for Japanese so that the minimum required text length is expressed in characters instead of words.
* Adapts the functionality to mark words in a sentence for Japanese.
* Adds a functionality in estimating reading time of a text to use characters per minute formula and adds the reading time score for Japanese.
* Adds a step to exclude Table of Contents in `sanitizeString.js` helper and uses this helper in Japanese `countCharacters` and `getWords` helpers.
* Replaces 'words' with 'characters' in the Japanese feedback strings for the Subheading distribution, Paragraph length, Sentence length, Keyphrase length, and Text length assessments.
* Adds the option to pass a custom `matchWordInText` helper and a helper for matching Japanese keyphrases in the `findKeywordInFirstParagraph` research.
* Fixes the Japanese full text test.
* Adds a Japanese full text test.
* Recognises internal link if the site's host is null but the site domain and the link host are the same.
* Adds documentation of the scoring criteria of the assessments.
* Edits the links to two documentation files and the size of the image displaying the assessments bullets.
* Adds links between newly created readme files and edit two minor typos.
* Refactors the `isInternalLink` helper to improve clarity.
* Removes superfluous full text test papers for all languages in the yoastseo package.
* Adds a spec file for `getLanguagesWithWordFormSupport.js`.


## 1.92.0
### Enhancements
* Adds correct shortlinks to product assessors.
* Adjusts recommended text length values for non-cornerstone collections pages so that they are lower than the cornerstone ones.
* Adds a keyphrase distribution assessment to all relevant product-related assessors.
* Adds assessors for collection pages and sets them in the `apps/content-analysis/src/analysis.worker.js` file.
* Adds assessors for blog posts and pages, and sets them in `analysis.worker` in the content analysis app.
* Adds shortlinks to the feedback text of the Lists assessment.
* Creates `ImageAltTagsAssessment` for product pages.
* Adds custom configuration for the keyphrase length assessment when used for product pages.
* Adds language-specific configuration parameters for German, Dutch and Swedish.
* Refactors the `paragraphTooLong` assessment into a class and adds custom config for the assessment when used for product pages.
* Refactors the way that the right `config` is set for the sentence length assessment.
* Passes appropriate `config` to the sentence length assessment class in the product page content assessors.
* Adds an extra check in the `SubheadingsDistributionTooLong` assessment's applicability where we can adjust whether the assessment should appear in a short text analysis or not.
* Adds assessors for store blogs.
* Adds functionality for the webworker to load custom assessors for the SEO analysis (regular/cornerstone), the SEO analysis for related keyphrases (regular/cornerstone), and the readability analysis (regular/cornerstone).
* Implements the consecutive sentence beginnings assessment for Slovak to ensure variety in a text.
* Implements the passive voice assessment for Slovak.
* Implements the transition words assessment for Slovak.
* Improves keyphrase recognition in Slovak by filtering out function words such as `som, a, jedna, že`.
* Adds external stemmer for Slovak.
* Improves keyword detection for Norwegian by expanding the list of function words.
* Implements the passive voice assessment for Norwegian.
* Adds Norwegian transition words and activates the transition words assessment.
* Activates the consecutive sentences assessment for Norwegian and adds a list of exception words to exclude from the assessment.
* Adds `jak wiemy` to the list of Polish transition words.
* Adds Czech stemmer.
* Implements the passive voice assessment for Czech.
* Adds function words for Czech.
* Adds the sentence beginnings assessment for Czech.
* Adds the transition words list for Czech and improves the list of function words with the most common irregular comparatives.
* Changes the recommended maximum result of `portuguesePaper2` in `fullTextTests` folder and deletes a word that is also a preposition from participle list.
* Adds diacritics to the list of full form non-passive exceptions for Turkish.
* Adds more non-passive exception words to the list of full form non-passive exceptions for Turkish.
* Improves code quality of the Spanish stemmer and changes the way we stem words ending in `-mente`.
* Implements the new `Clause` and `SentenceNew` class to English folder, adds English Clause class, and adjusts related specs.
* Adapts the language-specific files for passive voice assessment to the new structure for German, Dutch, French, Spanish, Italian, Portuguese, Polish, Hungarian, and Czech.
* Refactors the `Sentence` and `SentencePart` classes.
* Moves passive voice values and some helpers, and the `ProminentWord` value to different folders in order to improve file organization.
* Moves an exception check specific to English from the general `getClauses` file to the English-specific `getClauses` file.
* Deletes obsolete files and changes the `SentenceNew` class to `Sentence`. Also adds an additional check in `parse.js` to parse classes.
* Fixes the stopwords list for Czech by adding the correct stopwords. Adds punctuation marks as sentence breakers.
* Excludes Table of Contents from the analysis.
* Filters out table block content from the `getSentenceBeginnings` research.
* Makes `transitionWordsAssessment` not applicable when the text has less than 200 words.
* Adds shortlinks to the feedback text of the Images and Image Keyphrase assessments.
* Adds text length score boundaries config for product pages to the product page SEO assessors.
* Adds custom config for `TitleWidth` assessment for product pages and adds extra feedback string in the assessment file for when short title width is not penalized with a bad score.
* Creates a research file that checks if there are lists in the text, and an assessment file that returns a red bullet if there is no list and a green bullet if there is one.
* Adds assessors for product pages.
* Removes the outbound links and internal links assessments from the SEO analysis on product pages.
* Removes the Flesch Reading Ease assessment and the consecutive sentences assessment from the readability analysis on product pages.
* Includes videos in the `ImageCount` analysis when the `countVideos` value is true.
* Passes custom configuration for the images assessment for product pages
* Splits the `TextImagesAssessment` into two, `KeyphraseInImageTextAssessment` and `ImageCountAssessment`.

### Other
* Removes the Flesch Reading Ease assessment from the blog posts and pages assessors.
* Removes the `Images` assessment from related keyphrase analysis in `productPages`.
* Refactors the transition words, passive voice, text presence, and sentence beginnings assessments into classes.

## 1.91.2 March 2nd, 2023
### Bugfixes
* Fixes a potential security issue.

## 1.91.1 January 25th, 2023
### Bugfixes
* Fixes a bug where an invalid browser field in package.json would cause Vite to throw an error when building.

## 1.89.0 January 25th, 2021
### Enhancements
* Adds Turkish stemmer.
* Adds function words for Turkish.
* Adds the consecutive sentences beginnings assessment for Turkish.
* Adds the sentence length assessment for Turkish.
* Implements the passive voice assessment for Turkish.
* Implements the transition words assessment for Turkish.

## 1.88.0 January 12th, 2021
### Enhancements
* Adds function words for Norwegian.
* Adds Norwegian stemmer.

## 1.87.0 December 14th, 2020
### Enhancements
* Implements the Hungarian passive voice.
* Implements Hungarian stemmer.
* Edits typos and add more transition words for Hungarian.
* Adds function words for Hungarian.
* Adds the sentence beginning assessment and specs for Hungarian.
* Adds specs for the sentence length assessment in Hungarian.

## 1.86.0 November 30th, 2020
### Enhancements
* Adds prefixes `وبال, مست, تنن, است, نست, ولل, كال, فال, بال, وال, لن, سن, ست, ات, وس, فل, فب, فس, لل, ول, وب, ال` to the list of ignored prefixes for Arabic. When the Arabic keyphrase is preceded by one of these prefixes, the basic keyphrase forms will also be created.

### Bugfixes
* Fixes a bug where links that contain line terminators (`\n`, `\r`, `\u2028` or `\u2029`) would not be detected as a link.

## 1.85.0 November 16th, 2020
### Enhancements
* Implements Hebrew stemmer.
* Implements the transition words assessment for Hebrew.
* Adds Hebrew papers and specs to check the assessments functionality for Hebrew.
* Adds the consecutive sentences assessment for Hebrew.
* Adds Sentence length assessment for Hebrew.
* Adds passive voice assessment for Hebrew, including a check, a number of specs and a list of Hebrew verb roots, both in their original and (for some) modified version.
* Adds the words 'annoyed', 'depressed', 'disappointed', and 'upset' to the list of English non-passives.
* Adds Swedish, Indonesian, Arabic, Hebrew, and Farsi to the list of languages with function words, to which a different scoring system applies for the keyphraseDistribution assessment.

## 1.84.0 October 26th, 2020
### Enhancements
* Expands list of exceptions for Indonesian words that look like passive voice forms but are not and adds a rule that checks for those exceptions.

### Bugfixes
* Adds various quotation marks to `wordBoundaries` and `removePunctuation`.

## 1.83.0 October 13th, 2020
### Enhancements
* Creates basic keyphrase forms for Farsi when they have the following affixes: prefix `ن`, and suffixes `مان, شان, تان, ش, ت, م, ی` and their variations such as ` ‌اش, ‌ات, یی, یم, یت, یش, ‌ای, ‌ام`.
* Adds function words for Farsi and edits a typo in the Arabic function words file.

## 1.82.0 September 14th, 2020
### Enhancements
* Adds the following words to the list that's used for the English transition word assessment: 'note that', 'not only', 'initially', 'as opposed to'.
* Improves word tokenization when words are followed by specific Arabic or Urdu punctuation marks.
* Adds passive voice assessment for Arabic, including a check, and a list of Arabic passive verbs with long vowel in their root.
* Adds the consecutive sentences assessment for Arabic.
* Implements the transition words assessment for Arabic.
* Adds the prefixes س and أ to the list of ignored prefixes for Arabic. When the Arabic keyphrase is preceded by س and أ, the basic keyphrase forms will also be created.
* Adds and refactors an external stemmer for Arabic.

### Bugfixes
* Fixes the URL of the assessment image in the readme. Props to [Güven Atbakan](https://github.com/shibby).

### Other
* Sets the threshold for words to be recognized as prominent words from 4 to 2, in case a language has no morphology support. This increases the chance that different word forms are saved, leading to better internal linking suggestions for these languages.

## 1.81.0 August 31st, 2020
### Enhancements
* Adds more Polish function words.
* Adds word form functionality for Polish.
* Makes feedback string in the keyphrase in subheadings assessment more explicit.
* Improves keyphrase recognition for Hebrew.
* Adds function words for Hebrew.

### Bugfixes
* Fixes a bug where closing parentheses would always be regarded as sentence endings in RTL languages.
* Fixes a bug where closing parentheses would always be regarded as sentence endings when followed by an upper-case letter.

## 1.80.0 August 17th, 2020
### Enhancements
* Makes it possible to tokenize sentences in languages that are written right-to-left (e.g., Hebrew, Arabic, Farsi and Urdu).
* Improves keyphrase recognition for Arabic.
* Adds function words for Arabic and Hebrew.
* Adds Sentence beginning assessment for Indonesian.
* Adds Transition words assessment for Indonesian.
* Adds Passive voice assessment for Indonesian.
* Adds Flesch reading ease assessment for Portuguese.
* Adds passive voice assessment for Portuguese.
* Fixes inconsistency in feedback strings that are produced by the Keyphrase in SEO Title assessment.

## 1.79.0 August 3rd, 2020
### Enhancements
* Adds some irregular plural forms of Italian words to function words and morphologyData file, including specs and a function to call the exception list.
* Adds a check for whether the word belongs to stemming exceptions to the Russian stemmer.
* Adds an exception check for irregular Italian diminutives to the Italian stemmer.
* Adds a check for exception lists of full forms and words with multiple stems to the Russian stemmer.
* Adds specs for irregular verbs for the Italian full forms exception list.
* Adds check for French verbs on -ions before stemming the verb suffix -ons and does not stem -ons if the word ends on -ions (for these words, the suffix is most likely -s, not -ons).
* Adds Indonesian function words.
* Changes the `getProminentWordsForInternalLinking` research so that it only runs for texts over 400 words or texts over 300 words with a title and/or a metadescription specified.
* Changes the output of the research so that it not only returns the list of prominent words but also information about the presence of metadescription and title, as well as the length of text. This information is later used to return a customized message to the user, within the internal linking suggestions container.

## 1.78.0 July 20th, 2020
### Enhancements
* Adds word form support for Portuguese.
* Adds word form support for Indonesian.
* Adds a check for keywords containing a dash as part of the word (e.g., buku-buku) so that they can be found in the slug.
* Adds "кроме того" to Russian transition words.
* Adds a check for whether the word belongs to stemming exceptions to the Russian stemmer.
* Adds an exception check for irregular Italian diminutives to the Italian stemmer and removes _scrivere_, _scrive_, and _scritto_ from function words.
* Adds a check for exception lists of full forms and words with multiple stems to the Russian stemmer.
* Adds some irregular plural forms of Italian words to function words and morphologyData file, including specs and a function to call the exception list.
* Improves the feedback text for the _keyphrase in title_ assessment to make clear that an exact keyphrase match is necessary.

## 1.77.0 June 22nd, 2020
### Enhancements
* Adds check for stemming -ons suffix in French, and for the exception list of words where only -s should be stemmed in French.
* Adds check for nouns in Russian on _ость_ before removing verb suffixes (including the newly added suffix _ть_).

## 1.76.0 June 8th, 2020
### Enhancements
* Adds a check for the exception list of French verbs with multiple stems and stems them by returning the indicated canonical stem.
* Adds a stemmer for the Italian language.
* Adds an exception check for words ending in -is/us/os where -s should not be stemmed.
* Improves the way keyphrases containing words ending in "ent" are recognized in the text.
* Stems French words that are considered too short to be stemmed according to the stemming rules, but that should nevertheless be stemmed.

## 1.75.0 May 25th, 2020
### Enhancements
* Adds a stemmer for the Russian language.
* Adds checks for exception list with full forms and exception list of words with multiple stems to the French stemmer.
* Checks an exception list of plurals with -x suffix and stems them correctly.
* Transfers data from French stemmer to data file and improves stemming of words in -issement in French.

## 1.74.0 May 11th, 2020
### Added
* Adds French to the list of languages for which we have morphology support.
* Adds word forms support for French in a beta version.

### Changed
* Improves the transition word assessment for Hungarian. Props to [@Zsoru](https://github.com/Zsoru).
* Improves the way diminutive nouns are stemmed.

### Fixed
* Makes sure that lists with single words don't skew the keyphrase distribution score.

## 1.72.0 April 9th, 2020
### Added
* Adds word-form support for Spanish.

### Fixed
* Fixes a bug where the text analysis would break if the text contains the word "Ying".

## 1.71.0 March 30th, 2020
### Enhancements:
* Adds word form support for Dutch.

## 1.66.0 January 6th, 2019
### Other
* Drops IE11 support through configuring Babel to use the preset environment with the own list of supported browsers specified.
* Adds a way to register a custom parser for parsing a paper to a structured tree representation, ready for further analysis.
* Adds a new attribute to the `Paper`: `wpBlocks`, to be used to send WordPress block editor data to the analysis.

## 1.65.1 December 11th, 2019
* Fixes a bug where the metabox would be broken when a relative URL was configured as `WP_CONTENT_URL`. Props to [FPCSJames](https://github.com/FPCSJames).

## 1.63.0 November 13th, 2019
### Fixed
 * Fixes a bug where the verb form `landscape` was not recognized in the text when the keyphrase contained the verb `landscaping`.

## 1.62.0 October 29th, 2019
### Enhancements:
 * Adds a new Readability Score of 0 and "Not Available". Changes Readability for empty content from "Needs Improvement" to "Not Available". Props to [emilyatmobtown](https://github.com/emilyatmobtown)

## 1.61.0 October 14th, 2019
### Added
* Adds the transition word assessment for Hungarian, props to [9abor](https://github.com/9abor)

## 1.60.0 September 30th, 2019
No user-facing changes.

## 1.59.0 September 16th, 2019
No user-facing changes.

## 1.58.0 September 3rd, 2019
### Enhancements:
* Implements the assessment that checks whether multiple sentences begin with the same word for Portuguese, props to [amesdigital](https://github.com/amesdigital).
* Increases the recommended sentence length limit for Portuguese, props to [amesdigital](https://github.com/amesdigital).

## 1.57.0 July 22nd, 2019
### Changed
* No user-facing changes.

## 1.56.0 July 8th, 2019
### Changed
* Changes the following improved internal linking functionality (which is disabled by a feature flag by default):
  * Bumps the minimum number of required word occurrences from 2 to 4 (when extracting prominent words for internal linking).

## 1.55.0 June 24th, 2019
### Changed
* Updated the CSS autoprefixer configuration to drop support for old Internet Explorer versions.

## 1.54.0 June 11th, 2019
### Added
* Adds an improved internal linking functionality behind a feature flag. For now, this new functionality is disabled by default. Specifically, the new internal linking functionality has the following improvements:
  * Changes the way prominent words are extracted for the Internal linking suggestions and Insights features. No multiple-word combinations are considered anymore, but single words only. For languages with morphological support (currently English and German) different morphological forms of words (e.g., `link`, `linking`, `links`, `linked`) are recognized and collapsed together, which improves the performance of the said features.
  * Changes the Insights functionality in the following way: no words that occurred less than 5 times in the text of the post are displayed. For languages with morphological support (for now, English and German) different word forms of the same word are collapsed based on their stem/base form (e.g., the stem of words `link`, `linking`, `links`, and `linked` is `link`). If the stem coincides with an actual word used in the text, it is displayed to the user in the list of Insights. If the stem does not coincide with any words used in the text, the first word form of the stem that occurred in the text will be used for display.
  * Changes the Internal linking suggestions functionality in the following way: not only the text of the post but also different meta data fields are used for analysis, i.e., title, keyphrase, synonyms, meta description, subheadings (H2 and H3 levels). One occurrence of a word in the said meta data fields is deemed more important than one occurrence of this word inside the text. Stems are further saved in the database and used for matching of prominent words between posts and the internal linking functionality.

## 1.52.0 May 14th, 2019
### Added
* Adds a dependency to `@yoast/feature-flag`.
* Adds createRegexFromArray export to yoastseo index.js.

### Fixed
* Improves keyphrase recognition for additional irregular verbs in German (e.g., `brennen`, `senden`, `kennen`).

## 1.51.0 April 29th, 2019
### Added
* Adds functionality to specify a custom premium-configuration branch to use for tests, locally and on Travis.

### Fixed
* Adds error handling to the YoastSEO development tool, when building the tree for visualization purposes.
* Fixes the parsing of paragraphs within headings when using the tree parser. Previously, it crashed the building of the tree.

### Changed
* Improves the recognition of German keyphrases that include words with an `i` or `e` in between vowels (e.g., `schrieen`, `schreien`, `speie`).
* Rewrites the SEO and readability analysis such that it uses both the original assessors as well the new tree assessors. The results of both are combined and their scores aggregated.

## 1.50.0 April 1st, 2019
### Added
* Adds word form recognition for German.
* Adds more transition words for Swedish.
* Adds visualization of the tree to the dev tool.

### Changed
* Increases the recommended sentence length limit for Spanish and Catalan, props to [Sílvia Fustegueres](https://www.ampersand.net/en/).
* Improves list of Catalan transition words, props to [Sílvia Fustegueres](https://www.ampersand.net/en/).

### Fixed
* Fixes a bug that impeded recognition of word forms for keyphrases on taxonomy pages.

## 1.49.0 March 11th, 2019
### Added
* The recalibrated analysis is out of its beta phase and is now the default for the SEO analysis. Thanks for testing and giving us your valuable feedback! You are awesome!

### Fixed
* Changes the text read out by a screen reader from `Bad SEO score.` to `Needs improvement.` when focused on the SEO score indicator in the menu bar and the traffic light in the snippet preview.

### Changed
* Props to [Kingdutch](https://github.com/Kingdutch) for helping us to improve our open source content analysis library.

## 1.48.0 February 25th, 2019
### Added
* Improves the feedback for the assessment that checks the length of the text in cornerstone articles.

### Fixed
* Fixes a bug where a Flesch reading ease score of exactly 90 would incorrectly trigger a red bullet.

## 1.47.0 February 11th, 2019
### Fixed
* Fixes accidental removal of the SEO Assessor export from the index. Props [Kingdutch](https://github.com/Kingdutch).

## 1.46.0 January 21st, 2019
### Added
* Adds readability analysis for Swedish.
* Adds prominent words for Swedish.
* Improves keyword recognition in Swedish by filtering function words.
* Improves the transition word assessment for German.
* Improves the error logging when there is an error in an SEO or readability assessment.

## 1.45.0 January 7th, 2019
### Fixed
* Fixes a bug where special characters from certain word lists weren't correctly escaped when matched with a regex. This resulted in `eggs` being incorrectly matched as the transition word `e.g.`, for example.
* Fixes a crash when loading the createWorker module because the window was accessed in the global scope immediately.

### Added
* When the recalibration feature flag is switched on:
  * The single title assessment is added. This assessment makes sure that you don't use superfluous H1s in your text.
  * The following assessments are not used anymore:
    * The assessment checking the length or your URL.
    * The assessment checking whether your URL contains stopwords.
  * Assessments changes:
    * Keyphrase density: changes scoring schema to account for the length of the keyphrase and changes feedback strings so that we give feedback about the number of occurrences rather than a percentage.
    * Outbound links assessment: changes the scoring schema so that red bullet instead of an orange bullet is shown when you have no outbound links.
    * Image alt attributes: if there are at least 5 images, checks whether the alt tags contain the keyphrase or synoynyms in 30-70% of all images. If there are less than 5 images, 1 image with the keyphrase or synonym in the alt tag is still scored as good.
    * Keyphrase in title: function words preceding the exact match keyphrase are ignored when determining the position of the keyphrase in the title.
    * Keyphrase length: makes the scoring scheme less strict for languages that don't have function word support, so that for these languages keyphrases with 1-6 words are scored as green, 7-9 as orange, and more than 9 as red.
    * Keyphrase in subheading: only takes H2 and H3 level subheadings into account and changes the scoring schema so that 30%-75% of these subheadings need to include the keyphrase or its synonyms. In languages without function word support, a match is only counted if all the words from the keyphrase/synonym appear in the subheading.
    * Text length: on taxonomy pages, the recommended minimum text length is increased from 150 to 250 words.
* The browser console now shows more descriptive error messages when something went wrong during analyses in the web worker.

### Changed
* Improved README Usage to detail the Web Worker API.

## 1.44.0 December 14th, 2018
### Fixed
* Fixes a bug where keyphrases weren't recognized in the URL when the words in the URL were separated by underscore characters instead of hyphens.
* Fixes a bug that caused numbers to be stripped when marking a keyphrase containing a number, e.g. `Yoast SEO 9.3`.

### Added
* Adds relevant words from the browserified example to the dev tool.
* Adds the option to use local morphology data in the dev tool.

### Changed
* Improves error handling in the analysis web worker by rejecting the last request instead of just throwing an error.

## 1.43.0 November 19th, 2018
### Fixed
* Fixes assessments failing when using a `<` sign in the content.
* Fixes a bug where paragraphs were sometimes not correctly detected because paragraph tags were not automatically added in WordPress-like fashion.

### Added
* Adds toggles to use the different assessors in the webpack example (e.g., for all the different combinations for cornerstone and taxonomy pages and related keyphrase).
* Introduce logger in the AnalysisWebWorker to replace the development console log.

### Changed
* Refactor SEO assessment file names and exports. Props [Kingdutch](https://github.com/Kingdutch).
* Disables the non-functioning markers for the subheading distribution assessment.

## 1.42.0 November 5th, 2018

### Fixed
* Improves keyword recognition in the first paragraph on texts which start with images and captions.

### Other
* Removes non-functioning eye-markers from the link keyphrase assessment.

## 1.41.1 October 29th, 2018

### Fixed
* Fixes a bug where the Chrome browser tab would crash on Windows when a French or Italian text contains sentences in passive voice, props [CarloCannas](https://github.com/CarloCannas).
* Fixes a bug where the Yoast SEO analysis would error if used together with the DelightfulDownloads plugin.

## 1.41.0 October 22nd, 2018

### Added
* Introduces two new principles for keyword recognition:
  * Makes keyphrase recognition flexible with regards to word order. This means that the keyphrase `SEO WordPress plugin` will be found in the sentence `This is the most popular SEO plugin for WordPress.` In order to use exact matches, the keyphrase can be enclosed in quotation marks.
  * When matching keyphrases for the various assessments, the analysis only targets content words and ignores function words (e.g., `the` or `and`). This functionality is available in English, German, Dutch, French, Spanish, Italian, Portuguese, Russian and Polish.
* Implements support for word form recognition for keyphrases in English (requires Premium configuration).
* Improves the feedback texts for all SEO and readability assessments.
* Adds functionality to append a query string to the assessment links through the analysis worker.
* Adds an assessment that checks whether your keyword consists only of function words.

### Changed
* The analysis of the following assessments incorporates the new keyword recognition principles:
  * Image alt attributes: checks whether there’s at least one image with an alt tag that contains words from the keyphrase. An exact match isn’t required anymore.
  * Keyphrase in introduction: checks whether words from the keyphrase are matched within one sentence in the introduction or, if not, whether they are present in the first paragraph at all. An exact match isn’t required anymore.
  * Keyphrase in title: still checks whether an exact match of the keyphrase is found in (the beginning of) the title, but now also recommends improvement if all words from the keyphrase are found in the title.
  * Keyphrase length: has new boundaries to check whether the keyphrase is not too long. For languages that have support for function word stripping (see above), only content words are taken into account.
  * Keyphrase in meta description: checks how often all words from the keyphrase are matched within the meta description.
  * Keyphrase in subheading: checks whether at least one subheading contains more than half of the words from the keyphrase. An exact match isn’t required anymore.
  * Keyphrase in slug: checks whether a sufficient number of words from the keyphrase is used in the slug. The number of words required depends on the length of the keyphrase.
  * Keyphrase density: checks whether there are enough keyphrase matches; a match is defined as a sentence that contains all words from the keyphrase. The upper boundary for a good score is higher when word form recognition is available, since in that case the analysis is able to pick up more matches.
  * Link focus keyphrase: the assessment that checks whether you’re using your keyphrase to link to a different article doesn't require an exact match anymore.
  * Keyphrase distribution uses an improved algorithm that checks whether the keyphrase is evenly distributed throughout the text.
* The following assessments will also count synonym matches as keyphrase matches:
  * Image alt attributes
  * Keyphrase in introduction
  * Keyphrase in meta description
  * Keyphrase in subheading
  * Keyphrase distribution
* Deprecates the assessment that checks if stopwords are used within the keyphrase.
* The analysis for related keyphrases only shows assessments relevant for the specific keyphrase. It omits assessments that are non-keyphrase-related and assessments that should only be applied to the focus keyphrase.

### Fixed
* Fixes a bug where the keyword would not be found in the slug when containing punctuation, for example the keyphrase `apples & pears` in the slug `apples-pears`.

## 1.40.0 September 24th, 2018

### Bugs:

* Fixes a bug that caused keywords beginning with the Turkish characters `İ` / `i` and `I` / `ı` not to be recognized when changing that character from lowercase to uppercase and vice versa.

### Enhancements:

* Exposes word boundaries for use in other libraries or applications.

### Changed:

* Drops TypeScript support.
* Changes all usage of `lodash` to `lodash-es`.
* `index.js` has been rewritten to use ES6 Module import and export logic, instead of Node's `require` logic.

## 1.39.3 September 19th, 2018

### Bugs:

* Fixes a bug that would cause a browser crash in Chrome and Opera on Windows when the site language was Polish and the readability analysis was switched on.

## 1.39.2 September 6th, 2018

### Bugs:

* Fixes a bug where the readability analysis would not show the correct scores for cornerstone content.
* Fixes a bug where switching off the SEO analysis would stop the readability analysis from loading.

## 1.39.1 September 5th, 2018

* Fixes a bug where our JavaScript memory usage would increase indefinitely. This could result in a browser crash after a long enough period.

## 1.39 August 28th, 2018

### Enhancements
* Adds readability analysis for Polish.
* Adds prominent words for Polish.

## 1.38.4 September 5th, 2018

### Bugs:

* Fixes a bug where our JavaScript memory usage would increase indefinitely. This could result in a browser crash after a long enough period.

## 1.38.3 August 24th, 2018

### Bugs:

* Fixes an issue where the worker communication wouldn't work when the script was minified.

## 1.38.2 August 24th, 2018

### Bugs:

* Fixes an issue where we would show bad results for an empty Paper.

### Enhancements:

* Automatically parse and serialize all value objects send to and from the worker. This means `Paper` objects can be passed to the analysis wrapper.
* Add a method to the worker that can run any arbitrary research, `runResearch`.
* Add a priority system to the Scheduler, this means extensions run before the analysis to make sure they are loaded.
* Adds support for related keywords.

## 1.38.1 August 21st, 2018
* Improves web worker functionality.

## 1.38.0 August 21st, 2018
* Adds basic web worker functionality.

## 1.37.0 August 13th, 2018
* Updates the font size of the snippet title measure element to correctly mimic Google's desktop snippet title.
* Deprecates the switch assessor.

## 1.36.1 August 6th, 2018
### Changed
* Increased the debounce delay in the App to make sure the refresh is triggered less often.

## 1.36.0 July 24th, 2018
### Added
* Adds a link to a relevant article about re-using keywords to the feedback of the assessment that checks if the keyword was used previously.
* Exposes all assessments and more as a public API.
* Adds the passive voice assessment for Dutch.

### Changed
* Improves the order in which assessments are triggered. The keyword in the title is only checked once there's a title, the keyword in the introduction is only checked once there's a text, and the keyword in the meta description is only checked once there's a meta description.

### Fixed
* Fixes a bug that caused keywords to be incorrectly recognized within possessive forms (e.g. `Natalia` in `Natalia's fix`).
* Improves the recognition of keywords with special diacritics in the URL.
* Improves keyword recognition through adding Spanish inverted exclamation and question marks to the rules that determine word boundaries.

## 1.35.5 July 16th, 2018

## 1.35.4 July 16th, 2018
### Changed
* Changes feedback in the keyword density assessment to make it more explicit that synonyms are not taken into consideration when calculating the score.

## 1.35.3 July 9th, 2018
### Changed
* Makes keyword distribution boundaries the same regardless of whether or not synonyms are set, enables markings also when the bullet is green, and shows a gray bullet when there are not enough keyword occurrences to calculate distribution.

## 1.35.2 July 5th, 2018
### Changed
* Removes the topic density assessment and uses the keyword density assessment also when synonyms are set.

## 1.35.1 July 4th, 2018
### Fixed
* Fixes a bug where a custom callback would not work correctly.

## 1.35.0 July 3rd, 2018
### Added
* Adds an assessment the checks the distribution of the focus keyword in the text.
* Adds a topic distribution and topic density assessment to use with synonyms.
* Adds links to relevant articles to all SEO assessments.
* Adds Flesch Reading Ease assessment for French.
* Adds Flesch Reading Ease assessment for Spanish.
* Add passive voice assessment for Italian.

### Fixed
* Instances of the same keyword with different kinds of apostrophes (e.g., brain’s and brain's) are now recognized as the same in the following assessments: keyword in meta description, keyword in subheading, keyword in first paragraph, keyword in title, keyword in URL.
* Filters out prominent word combinations ending in 's in English. Props to [swekkiekekkie](https://github.com/swekkiekekkie).

## 1.34.0 June 12th, 2018
### Fixed
* Fixes title width measurements by adding font styles to the hidden input field.

### Changed
* Improves the lists of French transition words, stopwords, and function words, props [Laurent-1971](https://github.com/Laurent-1971).

## 1.33.1 May 25th, 2018
### Fixed
* Fixes a bug where Flesch Reading Ease translation strings were not fully translated.

## 1.33.0 May 23rd, 2018
### Added
* Adds Catalan transition words.
* Added the title width to rawData in app.js to be used by the title width assessment.

### Changed
* Changed max meta description length from 320 to 156.

## 1.32.0 May 16th, 2018
### Added
* Adds Flesh Reading Ease for Russian.

### Changed
* Sequences of symbols which do not contain a single letter or digit are no longer considered a valid keyword.
* Question marks and hashes are stripped from the snippet preview URL.
* Makes the snippetPreview optional in the App. This can be enabled by putting the hasSnippetPreview argument to false.

## 1.31.0 April 26th, 2018
### Added
* Adds readability analysis for Russian.
* Adds prominent words for Russian.

### Changed
* Improves SVG image accessibility.
* Updates the language support table in the README.

### Fixed
* Fixes a bug where sentences ending in multiple sentence marks, exclamation marks or ellipses were treated as multiple sentences.

## 1.30.2  March 16th, 2018
### Changed
* Reverted the default view of the snippet preview to desktop.

## 1.30.1 March 12th, 2018
### Fixed
* Fixes a bug that broke a filter which marks Spanish and French sentences as non-passive when certain exception words occur between the auxiliary and the participle.

## 1.30.0 March 5th, 2018
### Added
* Adds a setter for titleWidth to the snippet preview.
* Adds a researcher to calculate the reading time for a given paper.
* Adds a filter to mark Spanish sentences as non-passive when certain exception words occur between the auxiliary and the participle. The list of exception words includes all forms of the copula 'estar'.
* Adds transition words assessment for Portuguese, props [amesdigital](https://github.com/amesdigital).
* Adds prominent words for Portuguese, props [amesdigital](https://github.com/amesdigital).

### Changed
* The snippet preview now shows the mobile preview by default.
* Fixes a bug where division by zero errors in the passive voice assessment would cause `NaN%` to show up in the feedback.
* Fixes a bug where multiple `rel` arguments prevented correct `nofollow` detection.
* Slightly increased the height of the meta description box so it matches the maximum amount of characters without needing a scrollbar.
* Improves the list of Portuguese function words.

## 1.29.0 January 15th, 2018
### Added
* Adds language support table to the README.
* Adds a performance analysis tool.
* Adds the passive voice assessment for French.
* Adds the passive voice assessment for Spanish.

### Changed
* Simplifies the message for the SubheadingsKeywordAssessment.
* Reduces the number of times the content analysis is refreshed on page load.
* Fixes a bug where relative URLs were not counted as internal links in the internal link assessment.

## 1.28.0 December 18th, 2017
### Changed
* Includes the link to the post about using your focus keyword multiple times also to the feedback text that is shown when using a focus keyword twice.
* Changes the anchor text of the link to the post about using your focus keyword multiple times to a more accessible one.

## 1.27.0 December 15th, 2017
### Changed
* The upper boundary of the meta description length has been changed from 156 to 320 characters.

## 1.26.0 December 13th, 2017
### Changed
* Updates the copy for the `previouslyUsedKeywords` assessment by referring to a new blogpost on why it might be a bad idea to use the same keyword more than once.

## 1.25.0 December 11th, 2017
### Added
* Adds a link to the subheadings article in the readability analysis to better explain the advantages of using subheadings in your text. #1317

## 1.24 November 29th, 2017
### Changed
* Changes the feedback string for a good meta description length from `The length of the meta description is sufficient.` to `The meta description has a nice length.` #1307

## 1.23.1 November 7th, 2017
### Added
* Adds a sentence research. #1278
* Adds escaping of the focus keyword in the url generated by the `PreviouslyUsedKeywords` assessment. #1281

## 1.23.0 October 31st, 2017
* Errors now give a score of -1. `ScoreToRating` interpreter will convert `-1` to `error`. #1272
* Adds `Excited` to exception -ed verbs for passive voice assessment. #1269
* Adds `release` script to `.travis.yml` for publishing to npm. #1265
* Adds a filter to exclude `code` tags and their enclosed content from the content analysis assessments. #1250
* Adds a filter to exclude `pre` tags and their enclosed content from the content analysis assessments, props [chrisboo](https://github.com/chrisboo). #1258
* Switches testing framework to `jest`. #1266

## 1.22.0 October 3rd, 2017
* Added typescript support.
* Add filter for exception words between auxiliary and passi