This project is based on U-net
To run color filling, you need the following module installed:
- numpy
- opencv-python
- tqdm
- pillow
- cython
- aiohttp
- scikit-image
- torch
- torchvision
You can install these dependencies via Anaconda or Miniconda. Miniconda is faster to install. (On Windows, choose the 64-bit Python 3.x version. Launch the Anaconda shell from the Start menu and navigate to this directory.) Then:
conda env create -f environment.yml
conda activate flatting
To update an already created environment if the environment.yml file changes or to change environments, activate and then run conda env update --file environment.yml --prune.
Download the pretrained network model and unzip checkpoints.zip into ./src/flatting/.
You can run our backend directly by:
cd src
python -m flatting
If you just want to run the backend only and don't want to touch the code. We provide a portable backend (Windows only) which packaged by the pyinstaller (see sec 4b.) You can download it and unzip to any place, then run:
cd flatting_server
flatting_server.exe
Issues: Although briefcase can output a cleaner package of our backend but it seems also hide the running log as well, we currently don't have a good solution for this issue yet.
Use briefcase commands for packaging. Briefcase can't compile Cython modules, so you must first do that. There is only one. Compile it via cythonize -i src/flatting/trapped_ball/adjacency_matrix.pyx.
To start the process, run:
briefcase create
briefcase build
To run the standalone program:
briefcase run
To create an installer:
briefcase package
To update the standalone program when your code or dependencies change:
briefcase update -r -d
You can also simply run briefcase run -u.
To debug this process, you can run your code from the entrypoint briefcase uses:
briefcase dev
This reveals some issues important to debug. It doesn't reveal dependency issues, because it's not using briefcase's python installation.
On my setup, I have to manually edit edit macOS/app/Flatting/Flatting.app/Contents/Resources/app_packages/torch/distributed/rpc/api.py to insert a line if docstring is None: continue after line 443:
assert docstring is not None, "RRef user-facing methods should all have docstrings."
If briefcase doesn't work, you can use pyinstaller:
pyinstaller --noconfirm flatting_server.spec
Download the flatting plugin and unzip it to any place. Download the backend server by following the instructions inside the "flatting plugin.zip"