iknow icon indicating copy to clipboard operation
iknow copied to clipboard

Community development repository for iKnow

Results 17 iknow issues
Sort by recently updated
recently updated
newest added

Tested in French, but likely relevant for other languages: the regular expression for the detection of NumberPlusUnit in metadata.csv contains a list of special characters, such as % and ‰....

bug

The following regular expression is used to detect number/number in the French language model: number-fraction;\d+\/\d+ The corresponding lexrep is: ;;{number-fraction};;FRNum; Example sentence: On définit l'hypertension par une valeur supérieure à...

bug

The Japanese language model currently supports DateTime attribute, which mostly covers those expressions that can be either Time or Duration. However, there are certain expressions that can definitely be considered...

enhancement

The implementation for issue #234 addressed expression including the particle の only, i.e., numbered item 1. Currently a new attribute label JPno_join_Con is being used. Additional expressions to be further...

enhancement

I pulled files from the latest Git master branch, built the solution (using the workaround for issue #241), and tried to run iKnowLanguageCompiler.exe by using lang_update.bat. It did not work...

bug

After upgrading to ICU69.1 so that the standalone iKnow language models can be used in IRIS, I ran into errors when building on Japanese OS. This was because there is...

bug

There are no warnings for certain mistakes made in the labels and rules file, which stop processing in IRIS. The following mistakes are currently not picked up by python: -...

bug

In the IRIS implementation, Concept-Relation-Concept triples can be extracted, but there is no feature to do so from iKnowPy. Is it possible to implement a method to extract the CRCs...

enhancement

This issue came out of a conversation with @Rei-hub and Dr. Torikai from Gunma University Hospital on February 9th, 2022 JST. Issue #33 introduced Negation expansion in Japanese for very...

enhancement

This issue came out of a conversation with @Rei-hub and Dr. Torikai from Gunma University Hospital on February 9th, 2022 JST. Certainty detection sounds like a very promising feature, but...

enhancement