Skip to content

.net library with Unicode data and algorithms to detect writing scripts used in a given text

License

Notifications You must be signed in to change notification settings

DaniRK/UnicodeScriptDetectorNet

Repository files navigation

UnicodeScriptDetectorNet

This solution consists of two parts:

  • UnicodeScriptDetector: the actual .net library.
  • ImportDataToCSharp: Produces a C# code file (Initialize.cs) based on newest Unicode consortium data, that is then used in the UnicodeScriptDetector project.

NuGet:

UnicodeScriptDetectorNet

Usage:

There are various detection methods and data extraction methods.

See here for a complete list of methods and more examples.

As an example, the UnicodeScriptDetector.GetUsedScripts method returns a list of possible Unicode "writing scripts" (like 'Latin', 'Arabic', 'Cyrillic') that might have been used to write the specified text, together with a probablity for each.

The list may contain 0, 1 or multiple results:

  • 0 result: there is no identication of any script (for example for a string "123 +")
  • 1 result: only 1 script was detected
  • multiple results: this usually happens if the text was written using mixed scripts (e.g.: "Hello Привет", using Latin and Cyrillic script)

The probablity for each script (>0, <=1) is calculated as the number of characters in the text belonging to that script devided by the number of all characters that could be related to any script.

Returned script identifiers are the official Unicode short and long names.

Example:

results = UnicodeScriptDetector.GetUsedScripts("Hello translates in Hebrew to: שלום");  
foreach(var r in results)  
{  
     Console.WriteLine($"Script short code: {r.scriptNameShort}, long code {r.scriptNameLong}, probablity:{r.propabilty}");  
}  

Output:  
Script short code: Latn, long code Latin, probablity:0.862069  
Script short code: Hebr, long code Hebrew, probablity:0.137931  
See here for explanations and more examples.

This library is based on data from official data files version 12.0 from the Unicode consortium.

For more details of theory and logic: http://www.unicode.org/reports/tr24/

About

.net library with Unicode data and algorithms to detect writing scripts used in a given text

Topics

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Languages