Parse an NLM MeSH Trees file into something usable as a Drupal taxonomy

mesh-tree-parse.py

from __future__ import print_function

# The data comes from the NLM's MeSH Trees file, which can be downloaded here: https://www.nlm.nih.gov/mesh/filelist.html
tree_file = open('mtrees2013.bin')
tree_outfile = open('mtrees2013-parsed.txt', 'w')
tree_array = {}
for row in tree_file.readlines():
	#print(row)
	# The 'index' of the term is whatever follows the semicolon
	term_index = row[row.find(';') + 1:len(row) - 1]
	# The term's parent (if any) will be whatever has the address one level up the hierarchy - i.e. A01.101.202's parent would be A01.101 - so we'll parse this out
	parent_index = term_index[:term_index.rfind('.')]
	# If a parent term has already been parsed, put that at the front of the 'term' to provide a hierarchy
	if parent_index in tree_array:
		term = ','.join([tree_array[parent_index], row[:row.find(';')]])
	# Otherwise, just take whatever precedes the semicolon as the 'term'
	else:
		term = row[:row.find(';')]
	# Add this term to our array (for future parent searches)
	tree_array[term_index] = term
	# Finally, write this term to our text file
	print(term, file=tree_outfile)