Модуль:data consistency check/doc

This module checks the validity and internal consistency of the language, language family, and script data used on Wiktionary: the modules in Category:Language data modules as well as Module:scripts/data.

Checks performed

For multiple data modules:

• Codes for languages, families and etymology-only languages must be unique and cannot clash with one another.
• Canonical names for languages, families, and etymology-only languages must not be found in the list of other names.
• Each name in the list of other names must appear only once.
• otherNames, if present, must be an array.
• Wikidata item IDs must be a positive integer or a string starting with Q and ending with decimal digits.

The following must be true of the data used by Module:languages:

• Each code must be defined in the correct submodule according to whether it is two-letter, three-letter or exceptional.
• The canonical name (field 1) must be present and must not be the same as the canonical name of another language.
• If field 2 is not nil, it must a valid Wikidata item ID.
• If field 3 or family is given and not nil, it must be a valid family code.
• If field 4 or scripts is given and not nil, it must be an array, and each string in the array must be a valid script code.
• If ancestors is given, it must be an array, and each string in the array must be a valid language or etymology language code.
• If family is given, it must be a valid family code.
• If type is given, it must be one of the recognised values (regular, reconstructed, appendix-constructed).
• If entry_name is given, it must be a table that contains either two arrays (from and to) or a string (remove_diacritics) or both.
• If sort_key is given, it may either be a string, or at table that in turn contains either two arrays (from and to) or a string (remove_diacritics).
• If entry_name or sort_key is given, the from array must be longer or equal in length to the to array.
• If standardChars is given, it must form a valid Lua string pattern when placed between square brackets with ^ before it (lua). (It should match all characters regularly used in the language, but that cannot be tested.)
• If override_translit is set, translit must also be set, because there must be a transliteration module that can override manual transliteration.
• If link_tr is present, it must be true.
• Have no data keys besides these: lua.

Checks not performed:

• If translit is present, it should be the name of a module, and this module should contain a tr function that takes a pagename (and optionally a language code and script code) as arguments.
• If sort_key is a string, it should be the name of a module, and this module should contain a makeSortKey function that takes a pagename (and optionally a language code and script code) as arguments.
• If entry_name or sort_key is a table and contains a field remove_diacritics, the value of the field should be a string that forms a valid Lua pattern when it is placed inside negated set notation ([^...]).

These are not checked here, because module errors will quickly crop up in entries if these conditions are not met, assuming that Module:utilities attempts to generate a sortkey for a category pertaining to the language in question, or lua attempts to use the transliteration module.

Module:languages/code to canonical name and Module:languages/canonical names must contain all the codes and canonical names found in the data submodules of Module:languages, and no more.

The following must be true of the data used by Module:etymology languages:

• canonicalName must be given.
• parent must be given must be a valid language, family or etymology-only language code.
• If ancestors is given, it must be an array, and each string in the array must be a valid language or etymology language code. The etymology language should also be listed as the ancestor of a regular language.
• Have no data keys besides these: lua.

Codes in Module:families data must:

• Have canonicalName, which must not be the same as the canonical name of another family.
• If family is given, it must be a valid family code.
• Have at least one language or subfamily belonging to it.
• Have no data keys besides these: lua.

Codes in Module:scripts data must:

• Have canonicalName.
• Have at least one language that lists it as one of its scripts.
• Have a characters pattern for script autodetection, and this must form a valid Lua string pattern when placed between square brackets (lua). (It should match all characters in the script, but that cannot be tested.)
• Have no data keys besides these: lua.