maoii-workflow

Workflow for massive analysis of internet images

Installing

1. Create a Python 3.10 virtualenv (not tested with other versions)
2. Clone this repository (e.g. gh repo clone eamonnbell-dur/maoii-workflow)
3. Initialize and update submodules, if not already done (e.g. git submodule update --init --recursive)
4. Install backend dependencies into the Python virtualenv using maoii-backend/requirements.txt
   • If you encounter an error installing datrie using pip, you may have to install this separately, and tell pip to use an earlier version of gcc. This is a known issue for datrie.
5. Install frontend dependencies in maoii-frontend/package.json, using npm or similar (e.g. yarn)
6. Build the frontend using webpack
7. Install jq

Using

1. Run the workflow using the following command:
   • snakemake -c NUM_CORES --resources gpu_workers=NUM_GPUS
   • where NUM_CORES is the number of available CPU cores, and NUM_GPUS is the number of CUDA devices available where snakemake is running.
2. When done, run a webserver from the root of this repository and navigate to the application.