Available metrics and rules
This catalog documents what arText calculates, how each result is interpreted and the limitations the integration must respect.
The catalog originates from the 21 plain-language recommendations of the arText claro system and currently comprises 25 suggestion rules, 22 of which form the plain-language set and 3 of which are used only by specialized genres. They are organized into the discourse, morphosyntactic and lexical levels, and each genre applies its own. Every report records the rules evaluated and their implementation version.
Metrics are not selected per request: POST /v1/reports snapshots the ones the requested genre (domain_slug + text_type_slug) enables for that language. GET /v1/metrics is the discovery and documentation catalog, and accepts the same pair to show what applies to each genre.
- 1
The report is created with language, domain_slug and text_type_slug.
- 2
The metric_id appears in measurements or suggestions.
- 3
The metric page documents unit, output and limitations.
Document
Paragraph count
paragraph-countNumber of non-empty paragraphs detected.
Segment count
segment-countNumber of recognized intra-sentence segments.
Sentence count
sentence-countNumber of sentences segmented by the analyzer.
Token count
token-countNumber of lexical units produced by the analyzer.
Word count
word-countNumber of words in the normalized text.
Readability
Average sentence words
average-sentence-wordsAverage number of words per sentence.
Maximum paragraph words
maximum-paragraph-wordsWord length of the longest paragraph.
Maximum sentence words
maximum-sentence-wordsWord length of the longest sentence.
Lexical
Use acronyms consistently
acronym-consistencyDetects repeated use of an expanded term after its acronym has already been introduced.
Introduce acronyms
acronym-introductionDetects the first occurrence of uppercase acronyms that do not appear with an expanded form.
Use of definitions
definition-usageCompares the number of definition markers in the text with the genre's reference value.
Replace difficult expressions
difficult-expressionsFinds difficult-to-understand expressions in the catalog and returns alternatives when available.
Replace ambiguous words
imprecise-wordsFinds imprecise words or sequences listed in the editable catalog and returns their alternatives.
Review long words
long-wordsFinds long words in the editorial catalog and proposes shorter alternatives when available.
Lack of definitions
missing-definitionsWarns when the text contains no definition marker although the genre usually includes definitions.
Remove redundant expressions
redundant-expressionsFinds catalogued redundant expressions so they can be removed or rephrased.
Use of subjectivity markers
subjectivity-markersCombines lexical items, sequences, morphological categories and exclamations to find possible subjectivity markers.
Use more transparent terms
transparent-termsDetects catalogued difficult terms and proposes more transparent wording.
Review repeated words
word-repetitionFlags content words repeated in the text above the threshold configured for the genre.
Syntax
Remove archaic verb forms
archaic-verb-formsDetects forms tagged as future subjunctive, generally considered archaic in current usage.
Review gerunds
gerundsFinds standalone gerund forms worth reviewing, excluding recognized auxiliaries and compound constructions.
Rephrase negative ideas
negative-formulationsFinds catalogued negative words or sequences to encourage direct affirmative wording.
Review participles
participlesFinds standalone participles while excluding recognized compound tenses and passive constructions.
Use of passive voice
passive-voiceFinds constructions identified by morphosyntactic analysis as possible passive voice.
Use verbs consistently
verb-person-consistencyWarns when first-person singular and plural verb forms coexist in the same document.
Review verbal nominalizations
verbal-nominalizationsDetects nouns matching productive verbal-nominalization patterns and excludes catalogued exceptions.
Discourse
Vary connectors
connector-varietyDetects discourse connectors repeated above the threshold and offers alternatives expressing the same relation.
Include lists
list-usageWarns, with a single document-level occurrence, when no structured list has been recognized in the text.
Review long paragraphs
long-paragraphsFlags paragraphs exceeding the word limit configured for the text genre.
Split long sentences
long-segmented-sentencesFlags long sentences with several segments and exposes structural points where they might be split.
Review long sentences
long-sentencesFlags sentences exceeding the word limit configured for the text genre and, when the sentence can be divided, the points where it might be split.
Review long sentences
long-unsegmented-sentencesFlags long sentences that DiSeg has not divided into internal segments.
Introduce connectors at the beginning of paragraphs
paragraph-initial-connectorsDetects paragraphs that do not begin with a recognized discourse connector.
Review sentence-paragraphs
single-sentence-paragraphsFinds paragraphs made up of a single sentence so they can be reviewed for excessive density or insufficient development.