Feature request: Replaceable support for Transliterator

Question

Feature request: Replaceable support for Transliterator

tavianator opened this issue 5 years ago · 10 comments

ICU has a Replaceble interface that can be passed to Transliterator if you want to keep track of where the changes happen. It would be great if custom Replaceable implementations could be provided rather than just strings.

Answer 1 · 2019-08-04T09:22:55.000Z

On Fri, 2 Aug 2019, Tavian Barnes wrote: ICU has a [Replaceble](http://icu-project.org/apiref/icu4c/classReplaceable.html) interface that can be passed to [Transliterator](http://icu-project.org/apiref/icu4c/classicu_1_1Transliterator.html) if you want to keep track of where the changes happen. It would be great if custom Replaceable implementations could be provided rather than just strings.

How would you provide the custom Replaceable implementation to PyICU ?

…

-- You are receiving this because you are subscribed to this thread. Reply to this email directly or view it on GitHub: #107

Answer 2 · 2019-08-28T19:57:42.000Z

Well, I was thinking as a Python class that implements methods for the various Replaceable functions. Open to suggestions.

Answer 3 · 2019-08-28T20:17:49.000Z

Currently, PyICU is a wrapper around existing C/C++ ICU code. Extending ICU C++ classes from Python is not supported by PyICU. That being said, it could be implemented in a reverse-wrapper kind of way, a bit like JCC does with Java: http://lucene.apache.org/pylucene/jcc/index.html You could define a C++ extension of Replaceable that wraps a pure Python implementation. The methods on the C++ extenstion would call the Python methods. Andi..

…

On Wed, 28 Aug 2019, Tavian Barnes wrote: Well, I was thinking as a Python class that implements methods for the various Replaceable functions. Open to suggestions. -- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #107 (comment)

Answer 4 · 2019-10-01T20:38:17.000Z

This feature is now implemented in HEAD. Please, check it out and let me know if it works as expected for you. I, so far, only added PythonReplaceable support in the Transliterator methods that accept it. If there are other such methods accepting Replaceable you'd like support added for, please let me know. Thanks !

Answer 5 · 2019-10-01T20:39:51.000Z

I forgot to mention that I also added a test that shows how to create your own python implementation of a Replaceable, wrap it for ICU with PythonReplaceable and use it with a Transliterator instance.

Answer 6 · 2019-10-02T16:50:02.000Z

Great, thanks! I'll try it out.

Answer 7 · 2019-10-02T18:09:07.000Z

On Wed, 2 Oct 2019, Tavian Barnes wrote: Great, thanks! I'll try it out.

Note that I also just found again something I had implemented years ago that might be relevant for your use case. There is a PythonTransliterator C++ class in transliterator.cpp that lets you define a Transliterator subclass in Python using a similar technique by only implementing the handleTransliterate() method. There is a test for it, showing you how it can be done, in test_Transliterator.py as well.

…

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #107 (comment)

Answer 8 · 2019-10-04T14:54:46.000Z

Just tried this out, it appears to work. Thanks!

Answer 9 · 2019-10-04T17:05:25.000Z

On Fri, 4 Oct 2019, Tavian Barnes wrote: Just tried this out, it appears to work. Thanks!

Great, thanks for checking. I hence consider the feature done.

…

-- You are receiving this because you commented. Reply to this email directly or view it on GitHub: #107 (comment)

Answer 10 · 2019-11-12T00:14:56.000Z

released in PyICU 2.4