fix: encode dataset name with urllib, removing default `safe='/'` config

[mattverantos]

This fixes an issue where the urllib.parse.quote fn is used to encode a path component that has a slash in it. Ex: if you name a dataset test/me, it needs to encode the slash.

>>> import urllib.parse
>>> urllib.parse.quote('test/me')
'test/me'
>>> urllib.parse.quote('test/me', safe='')
'test%2Fme'

You can see here that urllib.parse.quote does not encode the / by default, which is what is used in the _url_encode fn.

Related: https://github.com/langfuse/langfuse/pull/7026 Part of the fix for #5962

---

[!IMPORTANT] Fix dataset name encoding to handle special characters using urllib.parse.quote with safe='' and add tests for special character handling.

• Behavior:
  • Fix encoding of dataset names with special characters in get_dataset in client.py and get in datasets/client.py using urllib.parse.quote with safe=''.
• Tests:
  • Add test_dataset_special_characters in test_datasets.py to verify handling of special and unicode characters in dataset names.

^{This description was created by for 73421d59226d44675a3371c24b38509a6fc5d481. You can customize this summary. It will automatically update as commits are pushed.}

Greptile Summary

Disclaimer: Experimental PR review

This PR fixes URL encoding issues for dataset names in the langfuse-python client, particularly addressing the handling of forward slashes and special characters in URLs.

• Modified _url_encode() in langfuse/_client/client.py to properly encode forward slashes by removing default safe='/' parameter
• Added test case test_dataset_special_characters() in tests/test_datasets.py to validate handling of special and Unicode characters
• Updated get method in datasets/client.py to use urllib.parse.quote(dataset_name, safe='') for consistent encoding
• Potential issue: Some methods still use jsonable_encoder for dataset names which may cause inconsistent encoding behavior

_{💡 (1/5) You can manually trigger the bot by mentioning @greptileai in a comment!}

[mattverantos]

Feels strange conversing with a bot, but I figure someone will read this and make a determination/respond?

[hassiebp]

Thanks again for raising this, this has been already fixed on main 🙏🏾

[mattverantos]

@hassiebp I am still able to reproduce this error. I have had some conversation here if you would like to see my comments: https://github.com/langfuse/langfuse/issues/5962#issuecomment-2962036494 where the issue is both on the client and the server side.

I am still able to reproduce:

>>> special_name = 'Special/Character'
>>> l.create_dataset(name=special_name)
Dataset(REDACTED)
>>> l.get_dataset(name=special_name)
Traceback (most recent call last):
  File "/usr/local/lib/python3.12/site-packages/langfuse/api/resources/datasets/client.py", line 162, in get
    pydantic_v1.parse_obj_as(typing.Any, _response.json())
                                         ^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/site-packages/httpx/_models.py", line 832, in json
    return jsonlib.loads(self.content, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/json/__init__.py", line 346, in loads
    return _default_decoder.decode(s)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/json/decoder.py", line 338, in decode
    obj, end = self.raw_decode(s, idx=_w(s, 0).end())
               ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/usr/local/lib/python3.12/json/decoder.py", line 356, in raw_decode
    raise JSONDecodeError("Expecting value", s, err.value) from None

>>> import langfuse
>>> langfuse.__version__
'3.2.0'