To run the cloud-tracking algorithm, use
python run_tracker.py [Input_dir]
where [Input_dir] is the address of the input directory. The model parameters will automatically be retrieved from the given files.
The cloud-tracking requires Python 3.5.
- Make run_tracker.py run with command arg
- Get rid of config.cfg
- Read model parameters from xray dataset
- Write model parameters to HDF5 dataset
- Implement/benchmark asyncio operations on ~ 30,000 cloudlet computations
- Numba/Cython for the expansion algorithm (cloudtracker.utility_functions)
- Wrap up with dask -> HDF5 output
- Filter noisy cloudlets out for speedup
- Threaded HDF5 file read for speed
- Wrap file I/O with dask
To profile the cloud-tracking algorithm using line_profiler, do
python -O -B -m kernprof -l -v run_tracker.py > line_stats.txt
or, to run with a memory profiler instead,
python -O -B -m memory_profiler run_tracker.py > memory_stats.txt