Closed rohandahale closed 1 year ago
You are correct.
Please see https://github.com/achael/eht-imaging/pull/163 -- it's merged in dev, and @achael has not made a release to the main branch since.
Is this still active? There's a number of bugs and code duplication I could fix but it won't mean much if it doesn't get merged
This particular problem is already fixed in the dev branch. @achael prefers that you make bugfixes (hopefully tested!) against the dev branch.
If you have additional bugs or see code duplication, please open separate issues.
Not referring to this, asking about the project as a whole. I already opened #164
As a whole, the project is quite active. I was about to comment on that PR.
Copy, thanks
This is issue is closed now but I will keep it open for people to see this.
To install eht-imaging 1.2.4 use pip with numpy 1.21.1:
pip install ehtim numpy==1.21.1
To install eht-imaging 1.2.5 use the dev branch on GitHub:
git clone --branch dev https://github.com/achael/eht-imaging.git
conda create --name ehtim_dev
conda activate ehtim_dev
conda install python=3.8.16
cd eht-imaging
python setup.py build
python setup.py install
pip install pandas==2.0.1 matplotlib==3.7.1 packaging astropy==5.2.2 hdrhistogram skyfield
conda install scikit-image networkx
conda install -c conda-forge pynfft
For M1 Pro and M1 Max chips, follow the issue #161 for NFFT installation and not the pynfft on conda.
thank you @rohandahale and apologies for the late action on this. I will investigate next week if we can easily make ehtim compatible with numpy 1.2.4. and make a new release
eht-imaging 1.2.5 seems to be working with numpy version 1.24 at the moment (though please raise an issue if anything else comes up). Since @rohandahale's installation instructions for M1 macs has been added to the README, I am now closing this issue.
pip install ehtim
will install numpy 1.24.2. In a python 3.8 shell when you doimport ehtim
, you will get an error:ImportError: cannot import name 'bool' from 'numpy'
The correct way to install ehtim should be:
pip install ehtim numpy==1.21.1