Support Python 3.x - Githubissues

datamade / usaddress

:us: a python library for parsing unstructured United States address strings into address components

https://parserator.datamade.us/usaddress

MIT License

1.52k stars 304 forks source link

Support Python 3.x #320

Closed qidanrui closed 2 years ago

qidanrui commented 2 years ago

The usaddress library only supports Python 2.7, which is because the probableparsing library only support Python 2.7 (https://pypi.org/project/probableparsing/). Remove this library or incorporate the code of probableparsing will solve this problem and make usaddress support Python 3.x.

mlissner commented 2 years ago

It's pretty sad around here:

This issue is asking for python 3 support (I can confirm that it works in 3.8 FWTW)
A PR to add Python 3.7
A bug explaining that python 3.10 doesn't work and suggesting there's a PR to fix it (but no PR seems to exist).
Forks adding better 3.x support
A forks fixing the dependency on python-crfsuite

This was a really great library and one that my organization depended on for parsing attorney addresses. Unfortunately, unless it's maintained somehow, we'll have to drop support for the feature, which is a bummer.

Is there any way we can get this moving again? @fgregg, maybe you have some ideas? I'd offer to take over the package at FLP, but we don't really have the expertise.

I think all it really needs is a fix to the crfsuite thing and a few tweaks to anywhere it mentions python 2.7, and it'd be ready to roll.

fgregg commented 2 years ago

hi @mlissner , i like and have appreciated our personal interactions , so your message that this project is disappointing you hurt my feelings. the package has worked for years with python 3. if it has stopped recently, that would be good to know, but your message has been very demotivating

mlissner commented 2 years ago

Sorry, I don't mean to lay out all these problems and insult the good work here, but at the same time, it does look like the project is abandoned. There are a number of bugs, PRs, and forks related to python 3, and there hasn't been a commit here for four years aside from a readme tweak.

I'd love to get this fixed up again, but it's not an area where I have expertise. Usually that's what I do when something I rely on breaks or needs a tweak, but I can't do that here, I don't think.

What's a better way forward if the project is still maintained?

fgregg commented 2 years ago

If you don't like this state of the project, you are always welcome to fork the project. If that does not work for you, Court Listener can sign a contract with DataMade to get some of your high priority issues addressed.

mlissner commented 2 years ago

That might work. I'll get in touch via email.

fgregg commented 2 years ago

3,10 works now