bash_profile_pipeline.sh

# BASH PROFILE COMMAND OPTION FOR PIPELINE (RECOMMENDED)

# 1) Copy this command into your bash profile ($ open ~/.bash_profile)
# 2) Change swd (line 10) to the absolute path on your computer where this folder is stored
# 3) Source your bash profile ($ source ~/.bash_profile)
# 4) Run spiderFish ($ spiderFish)
# It will prompt you for the rest of the input
# ALTERNATIVE: see shell_pipeline.sh

spiderFish() {
	# swd = ABSOLUTE/PATH/TO/CLONED/REPO
	swd=/Users/hannah/Dropbox/Westneat_Lab/spiderFish

	# prompt fish family
	echo -e "\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  Fish family?"
	read family

	# prompt destination for results
	echo "Folder destination?"
	read fishPath

	# make folder for storing all results
	mkdir -p $fishPath/$family

	# cd to that directory and store the absolute filepath (since relative filepath may be provided)
	cd $fishPath
	cwd=$(pwd)

	# cd to directory where local spider is stored and initiate search
	cd $swd/fishbase
	echo -e "\xf0\x9f\x8e\xa3 Gone fishin'... \xf0\x9f\x8e\xa3"
	sleep 1 # did I make it pause for a second so you have time to read the ‘Gone fishin’ with the fishing rod emoticons? you bet I did
	scrapy crawl fish -a family=$family -o $family.json

	# remove SHA1 tagged files/keep only the renamed ones
	# this is janky, I know there's probably a way to say 'get rid of anything that is exactly 20 characters' but I futzed with it for ages and couldn't get it to work so this is fine for now…
	find ./fishbase/output/full -type f -name "????????????????????*.*" -exec rm -rf {} \;

	# move all downloaded images and output JSON file to results folder
	mv ./fishbase/output/full/*.jpg $cwd/$family
	rm ./fishbase/output/full/*.gif
	mv $family.json $cwd/$family

	# initiate image classification
	python $swd/Sorting.py --dataset $cwd/$family

	# make folder for storing unsorted copies of images and move images into it
	mkdir $cwd/$family/All
	mv $cwd/$family/*.jpg $cwd/$family/All

	# reads JSON file and images in results folder and compares them to list of species on fishbase
	# output is up to 3 CSV files:

	# 1) family_speciesURLs.csv gives the picture URL and the species it corresponds to (ex: caaur_u1.jpg = 'Carassius auratus').

	# 2) family_missingPics.csv gives a list of any species for which no pictures are downloaded but which are listed as valid species on Fishbase (if 0, this file is not generated).

	# 3) family_failOnly.csv gives a list of species for which any fishbase photos were rejected by the image classifier (if 0, this file is not generated).
	Rscript $swd/speciesNames.R $family $cwd

	# go to results directory and prompt for display of images that passed image classifier
	cd $cwd
	echo "Finished. Display results? (y/n)"
	read response
	case $response in
		y|Y) open $cwd/$family/Pass/* ;;
	esac

	echo -e "\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  	\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  	\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  	\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  	\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1  	\xf0\x9f\x90\x9f  \xf0\x9f\x90\xA0  \xf0\x9f\x90\xA1"
}