"Any results or figures that you put into the report, we should be able to generate, using the instructions in the README, the more automated, the better."
Example project beta has a Makefile to automatically run a bunch of substeps for a step; e.g., running make preprocess_data will run all the preprocessing code. I'm thinking we can also do this for generating figures. For example, I'll probably have two figures in the report for segmentation - one from k-means and one from MRF-EM. In the Makefile I can write
so if someone runs make segmentation in the terminal, python will run kmeans_figure.py and mrf_em_figure.py and generate two figures. The Makefile would be in the code folder.
On piazza Matthew says:
"Any results or figures that you put into the report, we should be able to generate, using the instructions in the README, the more automated, the better."
Example project beta has a Makefile to automatically run a bunch of substeps for a step; e.g., running
make preprocess_data
will run all the preprocessing code. I'm thinking we can also do this for generating figures. For example, I'll probably have two figures in the report for segmentation - one from k-means and one from MRF-EM. In the Makefile I can writeso if someone runs
make segmentation
in the terminal, python will run kmeans_figure.py and mrf_em_figure.py and generate two figures. The Makefile would be in thecode
folder.Does this sound good?