SHELL=/bin/bash

.SECONDARY:

# $< - piersza zaleznosc
# $@ - cel

predict: test-A/out.tsv dev-0/out.tsv

test-A/out.tsv: test-A/out.num.tsv num2label.py
	./num2label.py < $< > $@

test-A/out.num.tsv: test-A/in.vw.txt train/train.vw.model
	vw -t $< -i train/train.vw.model --loss_function logistic --probabilities -p /dev/stdout > $@

test-A/in.vw.txt: test-A/in.tsv
	./tsv2vw.py < $< > $@

dev-0/out.tsv: dev-0/out.num.tsv num2label.py
	./num2label.py < $< > $@

dev-0/out.num.tsv: dev-0/in.vw.txt train/train.vw.model
	vw -t $< -i train/train.vw.model --loss_function logistic --probabilities -p /dev/stdout > $@

dev-0/in.vw.txt: dev-0/in.tsv
	./tsv2vw.py < $< > $@

train/train.vw.model: train/train.vw.txt
	vw $< -f $@ --passes 50 -b 20 --random_seed 2020 --oaa 8 --loss_function logistic --probabilities -k --cache_file vw-meta-cache 

train/train.vw.txt : train/train.tsv.gz tsv2vw.py
	zcat $< | ./tsv2vw.py > $@

clean:
	rm -rf dev-0/out.tsv dev-0/in.vw.txt train/train.vw.model train/train.vw.txt dev-0/out.num.tsv test-A/out.tsv test-A/out.num.tsv test-A/in.vw.txt