Makefile

# Make the whole analysis
# The prequisite for all is the end of your pipeline (in this case, your model predictions)
.PHONY: all
all: models/predictions.pkl
# You can preview the sequence of commands in the pipeline with: make all -B --recon
# I recommend putting that in your README

# 1. Download raw data
### replace species.csv, species2.csv, etc with names of your raw data files
### the ampersand-colon &: tells make that this rule has multiple grouped targets
data/raw/species.csv data/raw/species2.csv &: 
# replace this with something like: python src/data/download_data.py
	wget -O data/raw/species.csv https://figshare.com/ndownloader/files/3299483
	wget -O data/raw/species2.csv https://figshare.com/ndownloader/files/3299483

# 2. Process data
data/processed/features.csv: data/raw/species.csv
	python src/features/generate_features.py

# 3. Fit model
models/simple_features.pkl: data/processed/features.csv
	python src/models/train_model.py

# 4. Predict
models/predictions.pkl: models/simple_features.pkl
	python src/models/predict_model.py

# Utilities
# clean up
.PHONY: clean
clean:
	rm -rf data/raw/*

# notebook
notebook:
	jupyter notebook