Description
-
I have checked that this issue has not already been reported.
-
I have confirmed this bug exists on the latest version of pandas.
-
I have confirmed this bug exists on the master branch of pandas.
Reproducible Example
.
Issue Description
I tried to create a new development environment today.
Resolving the dependencies using pip install -r requirements-dev.txt
took two hours before I cancelled it.
I then tried to use pip-compile
from pip install pip-tools
and I received the message:
There are incompatible versions in the resolved dependencies:
botocore>=1.11 (from -r requirements-dev.in (line 35))
botocore>=1.12.201 (from moto==2.2.8->-r requirements-dev.in (line 37))
botocore<1.22.0,>=1.21.51 (from boto3==1.18.51->-r requirements-dev.in (line 34))
botocore<1.20.107,>=1.20.106 (from aiobotocore==1.4.1->-r requirements-dev.in (line 72))
Having pinned boto3>=1.17.106,<1.17.107
since that is what aiobotocore
packages with, the next error was:
There are incompatible versions in the resolved dependencies:
fsspec<2021.6.0,>=0.7.4 (from -r requirements-dev.in (line 73))
fsspec==2021.09.0 (from gcsfs==2021.9.0->-r requirements-dev.in (line 74))
fsspec==2021.09.0 (from s3fs==2021.9.0->-r requirements-dev.in (line 71))
fsspec (from fastparquet==0.7.1->-r requirements-dev.in (line 67))
fsspec>=0.6.0 (from dask==2021.9.1->-r requirements-dev.in (line 26))
So then I pinned gcsfs==2021.5.0
and s3f2==2021.5.0
.
It then compiled and I was able to install packages.
(adding a third complaint: new pyright checks need numpy >1.21.0 but numba needs less than that)
Expected Behavior
consistent deps.
Installed Versions
master as of 29-sep-21. python 3.8, mac os.