Skip to content

Commit 4d5841f

Browse files
fmindfmind
authored
fix(datasets): fix dtype backend (#44)
1 parent 09335b5 commit 4d5841f

File tree

3 files changed

+331
-1
lines changed

3 files changed

+331
-1
lines changed

‎python_env.yaml‎

Lines changed: 166 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,166 @@
1+
{
2+
"python": "3.13",
3+
"dependencies": [
4+
"alembic==1.14.1",
5+
"annotated-types==0.7.0",
6+
"antlr4-python3-runtime==4.9.3",
7+
"appnope==0.1.4",
8+
"argcomplete==3.5.3",
9+
"asttokens==3.0.0",
10+
"attrs==25.1.0",
11+
"bandit==1.8.3",
12+
"blinker==1.9.0",
13+
"cachetools==5.5.2",
14+
"certifi==2025.1.31",
15+
"cffi==1.17.1",
16+
"cfgv==3.4.0",
17+
"charset-normalizer==3.4.1",
18+
"click==8.1.8",
19+
"cloudpickle==3.1.1",
20+
"colorama==0.4.6",
21+
"comm==0.2.2",
22+
"commitizen==4.4.1",
23+
"contourpy==1.3.1",
24+
"coverage==7.6.12",
25+
"cycler==0.12.1",
26+
"databricks-sdk==0.44.1",
27+
"debugpy==1.8.12",
28+
"decli==0.6.2",
29+
"decorator==5.2.1",
30+
"deprecated==1.2.18",
31+
"distlib==0.3.9",
32+
"docker==7.1.0",
33+
"execnet==2.1.1",
34+
"executing==2.2.0",
35+
"fastjsonschema==2.21.1",
36+
"filelock==3.17.0",
37+
"flask==3.1.0",
38+
"fonttools==4.56.0",
39+
"gitdb==4.0.12",
40+
"gitpython==3.1.44",
41+
"google-auth==2.38.0",
42+
"graphene==3.4.3",
43+
"graphql-core==3.2.6",
44+
"graphql-relay==3.2.0",
45+
"greenlet==3.1.1",
46+
"gunicorn==23.0.0",
47+
"hatchling==1.27.0",
48+
"identify==2.6.8",
49+
"idna==3.10",
50+
"importlib-metadata==8.6.1",
51+
"iniconfig==2.0.0",
52+
"ipykernel==6.29.5",
53+
"ipython==9.0.0",
54+
"ipython-pygments-lexers==1.1.1",
55+
"itsdangerous==2.2.0",
56+
"jedi==0.19.2",
57+
"jinja2==3.1.5",
58+
"joblib==1.4.2",
59+
"jsonschema==4.23.0",
60+
"jsonschema-specifications==2024.10.1",
61+
"jupyter-client==8.6.3",
62+
"jupyter-core==5.7.2",
63+
"kiwisolver==1.4.8",
64+
"llvmlite==0.44.0",
65+
"loguru==0.7.3",
66+
"mako==1.3.9",
67+
"markdown==3.7",
68+
"markdown-it-py==3.0.0",
69+
"markupsafe==3.0.2",
70+
"matplotlib==3.10.1",
71+
"matplotlib-inline==0.1.7",
72+
"mdurl==0.1.2",
73+
"mlflow==2.20.3",
74+
"mlflow-skinny==2.20.3",
75+
"mypy==1.15.0",
76+
"mypy-extensions==1.0.0",
77+
"narwhals==1.28.0",
78+
"nbformat==5.10.4",
79+
"nest-asyncio==1.6.0",
80+
"nodeenv==1.9.1",
81+
"numba==0.61.0",
82+
"numpy==2.1.3",
83+
"nvidia-ml-py==12.570.86",
84+
"omegaconf==2.3.0",
85+
"opentelemetry-api==1.16.0",
86+
"opentelemetry-sdk==1.16.0",
87+
"opentelemetry-semantic-conventions==0.37b0",
88+
"packaging==24.2",
89+
"pandas==2.2.3",
90+
"pandas-stubs==2.2.3.241126",
91+
"pandera==0.23.0",
92+
"parso==0.8.4",
93+
"pathspec==0.12.1",
94+
"pbr==6.1.1",
95+
"pdoc==15.0.1",
96+
"pexpect==4.9.0",
97+
"pillow==11.1.0",
98+
"platformdirs==4.3.6",
99+
"plotly==6.0.0",
100+
"pluggy==1.5.0",
101+
"plyer==2.1.0",
102+
"pre-commit==4.1.0",
103+
"prompt-toolkit==3.0.50",
104+
"protobuf==5.29.3",
105+
"psutil==7.0.0",
106+
"ptyprocess==0.7.0",
107+
"pure-eval==0.2.3",
108+
"pyarrow==19.0.1",
109+
"pyasn1==0.6.1",
110+
"pyasn1-modules==0.4.1",
111+
"pycparser==2.22",
112+
"pydantic==2.10.6",
113+
"pydantic-core==2.27.2",
114+
"pydantic-settings==2.8.1",
115+
"pygments==2.19.1",
116+
"pynvml==12.0.0",
117+
"pyparsing==3.2.1",
118+
"pytest==8.3.5",
119+
"pytest-cov==6.0.0",
120+
"pytest-mock==3.14.0",
121+
"pytest-xdist==3.6.1",
122+
"python-dateutil==2.9.0.post0",
123+
"python-dotenv==1.0.1",
124+
"pytz==2025.1",
125+
"pyyaml==6.0.2",
126+
"pyzmq==26.2.1",
127+
"questionary==2.1.0",
128+
"referencing==0.36.2",
129+
"requests==2.32.3",
130+
"rich==13.9.4",
131+
"rpds-py==0.23.1",
132+
"rsa==4.9",
133+
"ruff==0.9.9",
134+
"scikit-learn==1.6.1",
135+
"scipy==1.15.2",
136+
"setuptools==75.8.2",
137+
"shap==0.46.0",
138+
"six==1.17.0",
139+
"slicer==0.0.8",
140+
"smmap==5.0.2",
141+
"sqlalchemy==2.0.38",
142+
"sqlparse==0.5.3",
143+
"stack-data==0.6.3",
144+
"stevedore==5.4.1",
145+
"termcolor==2.5.0",
146+
"threadpoolctl==3.5.0",
147+
"tomlkit==0.13.2",
148+
"tornado==6.4.2",
149+
"tqdm==4.67.1",
150+
"traitlets==5.14.3",
151+
"trove-classifiers==2025.3.3.18",
152+
"typeguard==4.4.2",
153+
"types-pytz==2025.1.0.20250204",
154+
"typing-extensions==4.12.2",
155+
"typing-inspect==0.9.0",
156+
"tzdata==2025.1",
157+
"urllib3==2.3.0",
158+
"virtualenv==20.29.2",
159+
"waitress==3.0.2",
160+
"wcwidth==0.2.13",
161+
"werkzeug==3.1.3",
162+
"win32-setctime==1.2.0",
163+
"wrapt==1.17.2",
164+
"zipp==3.21.0"
165+
]
166+
}

‎requirements.txt‎

Lines changed: 164 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,164 @@
1+
# This file was autogenerated by uv via the following command:
2+
# uv export --format=requirements-txt --no-dev --no-hashes --no-editable --no-emit-project --output-file=requirements.txt
3+
alembic==1.14.1
4+
annotated-types==0.7.0
5+
antlr4-python3-runtime==4.9.3
6+
appnope==0.1.4 ; platform_system == 'Darwin'
7+
argcomplete==3.5.3
8+
asttokens==3.0.0
9+
attrs==25.1.0
10+
bandit==1.8.3
11+
blinker==1.9.0
12+
cachetools==5.5.2
13+
certifi==2025.1.31
14+
cffi==1.17.1 ; implementation_name == 'pypy'
15+
cfgv==3.4.0
16+
charset-normalizer==3.4.1
17+
click==8.1.8
18+
cloudpickle==3.1.1
19+
colorama==0.4.6
20+
comm==0.2.2
21+
commitizen==4.4.1
22+
contourpy==1.3.1
23+
coverage==7.6.12
24+
cycler==0.12.1
25+
databricks-sdk==0.44.1
26+
debugpy==1.8.12
27+
decli==0.6.2
28+
decorator==5.2.1
29+
deprecated==1.2.18
30+
distlib==0.3.9
31+
docker==7.1.0
32+
execnet==2.1.1
33+
executing==2.2.0
34+
fastjsonschema==2.21.1
35+
filelock==3.17.0
36+
flask==3.1.0
37+
fonttools==4.56.0
38+
gitdb==4.0.12
39+
gitpython==3.1.44
40+
google-auth==2.38.0
41+
graphene==3.4.3
42+
graphql-core==3.2.6
43+
graphql-relay==3.2.0
44+
greenlet==3.1.1 ; (python_full_version == '3.13.*' and platform_machine == 'AMD64') or (python_full_version == '3.13.*' and platform_machine == 'WIN32') or (python_full_version == '3.13.*' and platform_machine == 'aarch64') or (python_full_version == '3.13.*' and platform_machine == 'amd64') or (python_full_version == '3.13.*' and platform_machine == 'ppc64le') or (python_full_version == '3.13.*' and platform_machine == 'win32') or (python_full_version == '3.13.*' and platform_machine == 'x86_64')
45+
gunicorn==23.0.0 ; platform_system != 'Windows'
46+
hatchling==1.27.0
47+
identify==2.6.8
48+
idna==3.10
49+
importlib-metadata==8.6.1
50+
iniconfig==2.0.0
51+
ipykernel==6.29.5
52+
ipython==9.0.0
53+
ipython-pygments-lexers==1.1.1
54+
itsdangerous==2.2.0
55+
jedi==0.19.2
56+
jinja2==3.1.5
57+
joblib==1.4.2
58+
jsonschema==4.23.0
59+
jsonschema-specifications==2024.10.1
60+
jupyter-client==8.6.3
61+
jupyter-core==5.7.2
62+
kiwisolver==1.4.8
63+
llvmlite==0.44.0
64+
loguru==0.7.3
65+
mako==1.3.9
66+
markdown==3.7
67+
markdown-it-py==3.0.0
68+
markupsafe==3.0.2
69+
matplotlib==3.10.1
70+
matplotlib-inline==0.1.7
71+
mdurl==0.1.2
72+
mlflow==2.20.3
73+
mlflow-skinny==2.20.3
74+
mypy==1.15.0
75+
mypy-extensions==1.0.0
76+
narwhals==1.28.0
77+
nbformat==5.10.4
78+
nest-asyncio==1.6.0
79+
nodeenv==1.9.1
80+
numba==0.61.0
81+
numpy==2.1.3
82+
nvidia-ml-py==12.570.86
83+
omegaconf==2.3.0
84+
opentelemetry-api==1.16.0
85+
opentelemetry-sdk==1.16.0
86+
opentelemetry-semantic-conventions==0.37b0
87+
packaging==24.2
88+
pandas==2.2.3
89+
pandas-stubs==2.2.3.241126
90+
pandera==0.23.0
91+
parso==0.8.4
92+
pathspec==0.12.1
93+
pbr==6.1.1
94+
pdoc==15.0.1
95+
pexpect==4.9.0 ; sys_platform != 'emscripten' and sys_platform != 'win32'
96+
pillow==11.1.0
97+
platformdirs==4.3.6
98+
plotly==6.0.0
99+
pluggy==1.5.0
100+
plyer==2.1.0
101+
pre-commit==4.1.0
102+
prompt-toolkit==3.0.50
103+
protobuf==5.29.3
104+
psutil==7.0.0
105+
ptyprocess==0.7.0 ; sys_platform != 'emscripten' and sys_platform != 'win32'
106+
pure-eval==0.2.3
107+
pyarrow==19.0.1
108+
pyasn1==0.6.1
109+
pyasn1-modules==0.4.1
110+
pycparser==2.22 ; implementation_name == 'pypy'
111+
pydantic==2.10.6
112+
pydantic-core==2.27.2
113+
pydantic-settings==2.8.1
114+
pygments==2.19.1
115+
pynvml==12.0.0
116+
pyparsing==3.2.1
117+
pytest==8.3.5
118+
pytest-cov==6.0.0
119+
pytest-mock==3.14.0
120+
pytest-xdist==3.6.1
121+
python-dateutil==2.9.0.post0
122+
python-dotenv==1.0.1
123+
pytz==2025.1
124+
pywin32==308 ; sys_platform == 'win32'
125+
pyyaml==6.0.2
126+
pyzmq==26.2.1
127+
questionary==2.1.0
128+
referencing==0.36.2
129+
requests==2.32.3
130+
rich==13.9.4
131+
rpds-py==0.23.1
132+
rsa==4.9
133+
ruff==0.9.9
134+
scikit-learn==1.6.1
135+
scipy==1.15.2
136+
setuptools==75.8.2
137+
shap==0.46.0
138+
six==1.17.0
139+
slicer==0.0.8
140+
smmap==5.0.2
141+
sqlalchemy==2.0.38
142+
sqlparse==0.5.3
143+
stack-data==0.6.3
144+
stevedore==5.4.1
145+
termcolor==2.5.0
146+
threadpoolctl==3.5.0
147+
tomlkit==0.13.2
148+
tornado==6.4.2
149+
tqdm==4.67.1
150+
traitlets==5.14.3
151+
trove-classifiers==2025.3.3.18
152+
typeguard==4.4.2
153+
types-pytz==2025.1.0.20250204
154+
typing-extensions==4.12.2
155+
typing-inspect==0.9.0
156+
tzdata==2025.1
157+
urllib3==2.3.0
158+
virtualenv==20.29.2
159+
waitress==3.0.2 ; platform_system == 'Windows'
160+
wcwidth==0.2.13
161+
werkzeug==3.1.3
162+
win32-setctime==1.2.0 ; sys_platform == 'win32'
163+
wrapt==1.17.2
164+
zipp==3.21.0

‎src/bikes/io/datasets.py‎

Lines changed: 1 addition & 1 deletion
Original file line numberDiff line numberDiff line change
@@ -74,7 +74,7 @@ class ParquetReader(Reader):
7474
@T.override
7575
def read(self) -> pd.DataFrame:
7676
# can't limit rows at read time
77-
data = pd.read_parquet(self.path, dtype_backend="pyarrow")
77+
data = pd.read_parquet(self.path, dtype_backend=self.backend)
7878
if self.limit is not None:
7979
data = data.head(self.limit)
8080
return data

0 commit comments

Comments
 (0)