Fix test_long_examples_validator

[tuliren]

Currently test_long_examples_validator fails with the following error message:

>       assert prepared_data_cmd_output.stderr == ""
E       assert ('Traceback (most recent call last):\n'\n '  File "/Users/lirentu/.pyenv/versions/3.9.11/bin/openai", line 8, in '\n '<module>\n'\n '    sys.exit(main())\n'\n '  File "/Users/lirentu/git/openai-python/openai/_openai_scripts.py", line '\n '78, in main\n'\n '    args.func(args)\n'\n '  File "/Users/lirentu/git/openai-python/openai/cli.py", line 592, in '\n 'prepare_data\n'\n '    apply_validators(\n'\n '  File "/Users/lirentu/git/openai-python/openai/validators.py", line 843, in '\n 'apply_validators\n'\n '    df, optional_applied = apply_optional_remediation(\n'\n '  File "/Users/lirentu/git/openai-python/openai/validators.py", line 611, in '\n 'apply_optional_remediation\n'\n '    df = remediation.optional_fn(df)\n'\n '  File "/Users/lirentu/git/openai-python/openai/validators.py", line 425, in '\n 'add_space_start\n'\n '    x["completion"] = x["completion"].apply(\n'\n '  File '\n '"/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/series.py", '\n 'line 4433, in apply\n'\n '    return SeriesApply(self, func, convert_dtype, args, kwargs).apply()\n'\n '  File '\n '"/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/apply.py", '\n 'line 1088, in apply\n'\n '    return self.apply_standard()\n'\n '  File '\n '"/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/apply.py", '\n 'line 1143, in apply_standard\n'\n '    mapped = lib.map_infer(\n'\n '  File "pandas/_libs/lib.pyx", line 2870, in pandas._libs.lib.map_infer\n'\n '  File "/Users/lirentu/git/openai-python/openai/validators.py", line 426, in '\n '<lambda>\n'\n '    lambda x: ("" if x[0] == " " else " ") + x\n'\n 'IndexError: string index out of range\n') == ''
E         + Traceback (most recent call last):
E         +   File "/Users/lirentu/.pyenv/versions/3.9.11/bin/openai", line 8, in <module>
E         +     sys.exit(main())
E         +   File "/Users/lirentu/git/openai-python/openai/_openai_scripts.py", line 78, in main
E         +     args.func(args)
E         +   File "/Users/lirentu/git/openai-python/openai/cli.py", line 592, in prepare_data
E         +     apply_validators(
E         +   File "/Users/lirentu/git/openai-python/openai/validators.py", line 843, in apply_validators
E         +     df, optional_applied = apply_optional_remediation(
E         +   File "/Users/lirentu/git/openai-python/openai/validators.py", line 611, in apply_optional_remediation
E         +     df = remediation.optional_fn(df)
E         +   File "/Users/lirentu/git/openai-python/openai/validators.py", line 425, in add_space_start
E         +     x["completion"] = x["completion"].apply(
E         +   File "/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/series.py", line 4433, in apply
E         +     return SeriesApply(self, func, convert_dtype, args, kwargs).apply()
E         +   File "/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/apply.py", line 1088, in apply
E         +     return self.apply_standard()
E         +   File "/Users/lirentu/.pyenv/versions/3.9.11/lib/python3.9/site-packages/pandas/core/apply.py", line 1143, in apply_standard
E         +     mapped = lib.map_infer(
E         +   File "pandas/_libs/lib.pyx", line 2870, in pandas._libs.lib.map_infer
E         +   File "/Users/lirentu/git/openai-python/openai/validators.py", line 426, in <lambda>
E         +     lambda x: ("" if x[0] == " " else " ") + x
E         + IndexError: string index out of range

openai/tests/test_long_examples_validator.py:50: AssertionError

The root cause is that the x in the lambda can be an empty string.

This PR fixes this issue by adding an extra check. It assumes that when the string is empty, no extra space is needed.

Such issue can be prevented by adding a CI to run the unit tests automatically. You may want to merge #203. I have also put up a CI PR tuliren#1 with all tests passing on my fork, and can re-create the PR against this repo if that's helpful.

[aelnosu]

Both work but change is better
"Both functions aim to add a space at the start of a string if it is not already present. However, there is a subtle difference in their implementation.

The first function checks if the first character of the string is a space using x[0] == " ". If it is not, it adds a space at the beginning using "" if x[0] == " " else " ". Note that this function assumes that the input string is not empty. If the input string is empty, this function will throw an IndexError.

The second function also checks if the first character of the string is a space using x[0] == " ". However, it also checks if the string is empty using x == "". If the string is empty, it simply returns the string as is. Otherwise, it adds a space at the beginning using "" if x == "" or x[0] == " " else " ".

Therefore, the second function is more robust as it handles the case where the input string is empty."

[tuliren]

Thank you for the suggestion. I have updated the PR. Also I changed the x in the lambda to s so that it does not clash with the input parameter x.

[EliahKagan]

Looks great! I've verified that (as before) this makes the test pass when applied on top of b82a3f7.

[rattrayalex]

Thanks for this!

We've since rewritten the library entirely, so this change is no longer relevant. I'm sorry we didn't get to it sooner.

[rattrayalex]

Ah, actually this change is still relevant, the code has just moved. I'll be applying this change in a separate PR shortly.