Limiting compatibility with numpy 2.1

[hmaarrfk]

Comment:

Is it worth limiting compatibility with numpy 2.1?

On my personal package, I already limit to numpy 1.X so I don't hit it so much, but I feel like some others might be hitting
pydata/xarray#9403

We personally hit it and added monkey patches around it....

[keewis]

looks like the reason we don't see that in CI (except upstream-dev) is that a dependency already pins to numpy<2.1

[hmaarrfk]

hmmm, i'm not sure how my previous environments were working with numpy 2.1 + xarray

I tried:

mamba create --name xr xarray>=2022 numpy=2.1 --channel conda-forge --override-channels

and its not resolving.... so maybe something changed....

[hmaarrfk]

Hmm, i just created a fresh envrionment and I noticed my monkeypatching code (that fixex the issue referenced) got triggered (i print a warning).

$ mamba list | grep -E "(xarray|numpy)"
numpy                     2.1.1           py310hd6e36ab_0    conda-forge
xarray                    2024.9.0           pyhd8ed1ab_0    conda-forge

seems to be compatible....

$ mamba create --name xr "xarray>=2024" "numpy>=2.1.0" python=3.10 --channel conda-forge --override-channels

Looking for: ["xarray[version='>=2024']", "numpy[version='>=2.1.0']", 'python=3.10']

warning  libmamba Could not parse mod/etag header
warning  libmamba Could not parse mod/etag header
conda-forge/noarch                                  16.5MB @   3.5MB/s  4.7s
conda-forge/linux-64                                38.2MB @   7.3MB/s  5.3s
Transaction

  Prefix: /home/mark/miniforge3/envs/xr

  Updating specs:

   - xarray[version='>=2024']
   - numpy[version='>=2.1.0']
   - python=3.10


  Package               Version  Build                Channel           Size
──────────────────────────────────────────────────────────────────────────────
  Install:
──────────────────────────────────────────────────────────────────────────────

  + _libgcc_mutex           0.1  conda_forge          conda-forge     Cached
  + ld_impl_linux-64       2.40  hf3520f5_7           conda-forge     Cached
  + python_abi             3.10  5_cp310              conda-forge     Cached
  + ca-certificates   2024.8.30  hbcca054_0           conda-forge     Cached
  + libgomp              14.1.0  h77fa898_1           conda-forge     Cached
  + _openmp_mutex           4.5  2_gnu                conda-forge     Cached
  + libgcc               14.1.0  h77fa898_1           conda-forge     Cached
  + libgfortran5         14.1.0  hc5f4f2c_1           conda-forge     Cached
  + libstdcxx            14.1.0  hc0a3c3a_1           conda-forge     Cached
  + libgcc-ng            14.1.0  h69a702a_1           conda-forge     Cached
  + openssl               3.3.2  hb9d3cd8_0           conda-forge     Cached
  + libgfortran          14.1.0  h69a702a_1           conda-forge     Cached
  + libzlib               1.3.1  h4ab18f5_1           conda-forge     Cached
  + libxcrypt            4.4.36  hd590300_1           conda-forge     Cached
  + libffi                3.4.2  h7f98852_5           conda-forge     Cached
  + bzip2                 1.0.8  h4bc722e_7           conda-forge     Cached
  + ncurses                 6.5  he02047a_1           conda-forge     Cached
  + libuuid              2.38.1  h0b41bf4_0           conda-forge     Cached
  + libnsl                2.0.1  hd590300_0           conda-forge     Cached
  + xz                    5.2.6  h166bdaf_0           conda-forge     Cached
  + libgfortran-ng       14.1.0  h69a702a_1           conda-forge     Cached
  + tk                   8.6.13  noxft_h4845f30_101   conda-forge     Cached
  + libsqlite            3.46.1  hadc24fc_0           conda-forge     Cached
  + readline                8.2  h8228510_1           conda-forge     Cached
  + libopenblas          0.3.27  pthreads_hac2b453_1  conda-forge     Cached
  + libblas               3.9.0  24_linux64_openblas  conda-forge     Cached
  + libcblas              3.9.0  24_linux64_openblas  conda-forge     Cached
  + liblapack             3.9.0  24_linux64_openblas  conda-forge     Cached
  + tzdata                2024a  h8827d51_1           conda-forge     Cached
  + python              3.10.15  h4a871b0_0_cpython   conda-forge       25MB
  + wheel                0.44.0  pyhd8ed1ab_0         conda-forge     Cached
  + setuptools           74.1.2  pyhd8ed1ab_0         conda-forge     Cached
  + pip                    24.2  pyh8b19718_1         conda-forge     Cached
  + six                  1.16.0  pyh6c4a22f_0         conda-forge     Cached
  + pytz                 2024.1  pyhd8ed1ab_0         conda-forge     Cached
  + python-tzdata        2024.1  pyhd8ed1ab_0         conda-forge     Cached
  + packaging              24.1  pyhd8ed1ab_0         conda-forge     Cached
  + python-dateutil       2.9.0  pyhd8ed1ab_0         conda-forge     Cached
  + numpy                 2.1.1  py310hd6e36ab_0      conda-forge     Cached
  + pandas                2.2.3  py310h5eaa309_1      conda-forge     Cached
  + xarray             2024.9.0  pyhd8ed1ab_0         conda-forge     Cached

  Summary:

  Install: 41 packages

  Total download: 25MB

──────────────────────────────────────────────────────────────────────────────


Confirm changes: [Y/n] y
python                                              25.3MB @  10.0MB/s  2.5s

Downloading and Extracting Packages:

Preparing transaction: done
Verifying transaction: done
Executing transaction: done

i don't know why i couldn't do this before