[BUG] load_solar failing on main, with pandas._libs.tslibs.parsing.DateParseError

[yarnabrina]

MCVE

>>> from sktime.datasets import load_solar
>>> load_solar()
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
  File "/home/anirban/sktime-fork/sktime/datasets/_single_problem_loaders.py", line 1272, in load_solar
    return _load_solar(
  File "/home/anirban/sktime-fork/sktime/datasets/_single_problem_loaders.py", line 1257, in _load_solar
    df = df.asfreq("30T")
  File "/home/anirban/conda-environments/sktime/lib/python3.10/site-packages/pandas/core/frame.py", line 10971, in asfreq
    return super().asfreq(
  File "/home/anirban/conda-environments/sktime/lib/python3.10/site-packages/pandas/core/generic.py", line 8347, in asfreq
    return asfreq(
  File "/home/anirban/conda-environments/sktime/lib/python3.10/site-packages/pandas/core/resample.py", line 2232, in asfreq
    dti = date_range(obj.index.min(), obj.index.max(), freq=freq)
  File "/home/anirban/conda-environments/sktime/lib/python3.10/site-packages/pandas/core/indexes/datetimes.py", line 945, in date_range
    dtarr = DatetimeArray._generate_range(
  File "/home/anirban/conda-environments/sktime/lib/python3.10/site-packages/pandas/core/arrays/datetimes.py", line 401, in _generate_range
    start = Timestamp(start)
  File "pandas/_libs/tslibs/timestamps.pyx", line 1667, in pandas._libs.tslibs.timestamps.Timestamp.__new__
  File "pandas/_libs/tslibs/conversion.pyx", line 280, in pandas._libs.tslibs.conversion.convert_to_tsobject
  File "pandas/_libs/tslibs/conversion.pyx", line 557, in pandas._libs.tslibs.conversion.convert_str_to_tsobject
  File "pandas/_libs/tslibs/parsing.pyx", line 329, in pandas._libs.tslibs.parsing.parse_datetime_string
  File "pandas/_libs/tslibs/parsing.pyx", line 658, in pandas._libs.tslibs.parsing.dateutil_parse
pandas._libs.tslibs.parsing.DateParseError: Unknown datetime string format, unable to parse: %Y-%m-%dT%H:%i:%SZ

Version

>>> from sktime import show_versions; show_versions()

System:
    python: 3.10.12 (main, Jul  5 2023, 18:54:27) [GCC 11.2.0]
executable: /home/anirban/conda-environments/sktime/bin/python
   machine: Linux-5.15.90.1-microsoft-standard-WSL2-x86_64-with-glibc2.35

Python dependencies:
          pip: 23.3.1
       sktime: 0.24.0
      sklearn: 1.3.0
       skbase: 0.4.6
        numpy: 1.24.4
        scipy: 1.11.1
       pandas: 2.0.3
   matplotlib: 3.7.2
       joblib: 1.3.1
        numba: 0.57.1
  statsmodels: 0.14.0
     pmdarima: 2.0.3
statsforecast: 1.6.0
      tsfresh: None
      tslearn: None
        torch: None
   tensorflow: None
tensorflow_probability: None

[yarnabrina]

@fkiraly I don't think it's coming from #5437, as I just replicated #5004 changes. But if I missed something, please let me know.

[fkiraly]

FYI @ciaran-g

[fkiraly]

We have the same in PeakTimeFeatures, see below.

My first suspicion was pandas, but there is no new release. Neither for numpy.

133     Examples
134     --------
135     >>> from sktime.transformations.series.peak import PeakTimeFeature
136     >>> from sktime.datasets import load_solar
137     >>> y = load_solar()
UNEXPECTED EXCEPTION: ValueError('could not convert string to Timestamp')
Traceback (most recent call last):
  File "pandas/_libs/tslibs/conversion.pyx", line 530, in pandas._libs.tslibs.conversion._convert_str_to_tsobject
  File "pandas/_libs/tslibs/parsing.pyx", line 318, in pandas._libs.tslibs.parsing.parse_datetime_string
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/dateutil/parser/_parser.py", line 1368, in parse
    return DEFAULTPARSER.parse(timestr, **kwargs)
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/dateutil/parser/_parser.py", line 643, in parse
    raise ParserError("Unknown string format: %s", timestr)
dateutil.parser._parser.ParserError: Unknown string format: %Y-%m-%dT%H:%i:%SZ
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/doctest.py", line 1350, in __run
    exec(compile(example.source, filename, "single",
  File "<doctest sktime.transformations.series.peak.PeakTimeFeature[2]>", line 1, in <module>
  File "/home/runner/work/sktime/sktime/sktime/datasets/_single_problem_loaders.py", line 1277, in load_solar
    return _load_solar(
  File "/home/runner/work/sktime/sktime/sktime/datasets/_single_problem_loaders.py", line 1262, in _load_solar
    df = df.asfreq("30T")
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/pandas/core/frame.py", line 11367, in asfreq
    return super().asfreq(
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/pandas/core/generic.py", line 8235, in asfreq
    return asfreq(
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/pandas/core/resample.py", line 2229, in asfreq
    dti = date_range(obj.index.min(), obj.index.max(), freq=freq)
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/pandas/core/indexes/datetimes.py", line 1125, in date_range
    dtarr = DatetimeArray._generate_range(
  File "/opt/hostedtoolcache/Python/3.10.13/x64/lib/python3.10/site-packages/pandas/core/arrays/datetimes.py", line 361, in _generate_range
    start = Timestamp(start)
  File "pandas/_libs/tslibs/timestamps.pyx", line 1698, in pandas._libs.tslibs.timestamps.Timestamp.__new__
  File "pandas/_libs/tslibs/conversion.pyx", line 249, in pandas._libs.tslibs.conversion.convert_to_tsobject
  File "pandas/_libs/tslibs/conversion.pyx", line 533, in pandas._libs.tslibs.conversion._convert_str_to_tsobject
ValueError: could not convert string to Timestamp
/home/runner/work/sktime/sktime/sktime/transformations/series/peak.py:137: UnexpectedException

[fkiraly]

this is actually also coming from load_solar.

A data download should not run in a docstring, that should be easier to fix.

[yarnabrina]

Can not reproduce the originally reported issue on current main, so closing. It must be fixed by sometime in last 2 months, don't know when.