Skip to content

Latest commit

 

History

History
101 lines (80 loc) · 3.46 KB

DEVELOPER.md

File metadata and controls

101 lines (80 loc) · 3.46 KB

Instructions for Developers

  1. Clone the repository

  2. Install conda environment from file and activate it. Use the respective environment depending on whether you have a CUDA GPU or not. The cubic spline psf is pre-compiled in a different repo / as a separate package and will be installed automatically.

     # for CUDA (change cpuonly to cudatoolkit in environment.yaml)
 conda env create -f environment.yaml

  3. Test whether everything works as expected. Note that if you run all tests, all files from the gateway file will be downloaded but you can exclude these tests.

         # assuming you are in the repo folder
     
     pytest decode/test  # all tests
     pytest -m "not webbig"  # without downloading the assets
     pytest -m "not plot" decode/test  # without the tests that have graphical output

  4. The package can be used in python as

    import decode

Note

Simulation heavily relies on the Cubic Spline PSF Implementation (Li, Y. et al. Real-time 3D single-molecule localization using experimental point spread functions. Nature Methods 15, 367–369 (2018)). It has been factored out in a seperate Repo to make life easier (see https://github.com/Haydnspass/SplinePSF). It'll be auto-installed.

Building the Docs

Install environment as described above

cd docs
make html

The docs can be found in the build folder.

Building and Deploying

Please follow the instructions as described here. Note that create the wheels only for Google Colab. All users should install via conda.

# optional, only once: create a new conda build environment
conda create --name build_clean conda-build anaconda bump2version -c conda-forge
conda activate build_clean

# bump version so that all versions get updated automatically, creates a git version tag automatically
bump2version [major/minor/patch/release/build]  # --verbose --dry-run to see the effect

# upload git tag
git push --tags

# build wheels
python setup.py bdist_wheel
# edit git release and upload the wheels

# conda release
cd conda
conda-build -c turagalab -c pytorch -c conda-forge decode

anaconda upload -u [your username] [path as provided at the end of the conda-build output]

After this you may test the build in a clean environment. You may update the link to the wheels in the gateway.yaml file in order to let Colab download the respective version. As we want to have this tested first, we did not automate this.

Gateway

Notebooks and Colab depend on gateway.yaml. The only relevant version is the one in the master branch. That means, in master you need to keep all packages that are potentially used.

# hashing a new .zip package
shasum -a 256 [.zip file]

Hdf5 structure

The h5 structure of decode EmitterSets is as follows (example shows a random emitter):

{'data': {
    'bg': array([nan, nan, nan], dtype=float32),
    'bg_cr': None,
    'bg_sig': None,
    'frame_ix': array([0, 0, 0]),
    'id': array([-1, -1, -1]),
    'phot': array([1., 1., 1.], dtype=float32),
    'phot_cr': None,
    'phot_sig': None,
    'prob': array([1., 1., 1.], dtype=float32),
    'xyz': array([[23.236893 , 13.064125 , 25.711937 ],
                  [ 9.199045 ,  7.3632336, 28.120571 ],
                  [21.179623 , 25.227995 , 25.280546 ]], dtype=float32),
    'xyz_cr': None,
    'xyz_sig': None
    },
 'decode': {
    'version': '0.10.0'
    },
 'meta': {
   'px_size': array([100., 100.], dtype=float32), 
   'xy_unit': 'px'
   }
}