christimperley/comby-python

Python bindings for Comby

comby-python's Issues

Use loguru for logging

Add __repr__ to classes to aid debugging

• Environment
• Match
• Location
• LocationRange
• BoundTerm

Add dual support for server and binary via CombyInterface

Rename CombyException to CombyError

Improve logging levels

When set to logging.INFO, rooibos still produces far too much information, most of which should be reported at a logging.DEBUG level.

Add ability to specify language when matching/substituting/rewriting

Add optional installation of binary / server binary to setup.cfg

Implement basic binary interface via CombyBinary

Unexpected Escape Characters (Backslash '\') In Comby-Python Output

Pls refer to the bug-report for details of the issue faced and reproduction route
report.pdf

Download comby binary as part of installation process

Add version property to CombyInterface

Add diff option to rewrite to return diff rather than updated text

Use sourcelocation to provide location data structures

Feature request: Support comby timeouts

For large numbers of comby operations on large source bases, it can take upwards of an hour to process all the files in my testing. This comes down dramatically (5mins!) if I parallelise my calls to comby with multiprocessing up to 32 ways. Unfortunately comby has a default 3 second timeout, and complicated files / very busy periods means that some files trip the timeout and don't get processed.

I've added the ability to set the comby timeout in my local branch and it now gives me correct results with the performance improvements.

Attaching the diff I used here as part of the feature request in case it's something interesting for the main project.
comby-timestamps.diff.txt

matches should still return an Iterator when no match is found

I'm trying to rewrite a bunch of code, however a few source files will not match the pattern and matches simply explodes on jsn = json.loads(self.call(cmd_s, text=source)) because call returns None. Is this the expected behaviour? To bypass this I'm catching the JSONDecodeError and going to the next file, but it would be nice if we simply had an empty Iterator

Add stream flag to matches method to toggle lazy/eager loading

Add basic examples to README

Add logging to Client