Derek Anderson

> Dear associate, > > We regret to inform you that we are unable to accept your request for access to the Product Advertising API. We have reviewed your site...

i don't think this is related to 3.11. 1. can you try changing the catch from `OSError` to `FileNotFoundError`? 2. does `ctypes.CDLL("libc.dylib")` run successfully on your windows box?

That's concerning. Did this happen to you? Is there an example somewhere? This is complicated a bit because people can supply their own model files. We'd need to not break...

Hey Shi, Not really sure. For those words I'm seeing: ``` >>> import wordninja >>> wordninja.split('occupational') ['occupational'] >>> wordninja.split('particulate') ['particulate'] ``` Are those the exact strings you're trying? The algo...

so this would be as a no-op check during deployment, right? if so, yeah, makes sense. but not crazy about `always_confirm_no`. maybe `confirm_noop`? `fail_if_not_current`? let's brainstorm names, open to other...

re local testing, where did it fail w/ that? at `createdb peeweedbevolve_test`? can you open a bug w/ the stacktrace? at min i should doc steps or catch and give...

they are in descending probability. i didn't add this as a feature because i didn't have any good ideas as to how to express that probability when adding. but likely...

Yeah I'm pretty sure the existing model and code is case blind. I'd be open to a PR and a language model trained on case. The statistical model should be...

true. but i wanted to keep the default model small (

two notes: i'd go w/ `add_words` that takes an iterable. `add` should not save / overwrite the original file, just work on the in memory copy. if we're going to...