tools/transform.py (b438983) - fs-words.git

transform.py

import sys
import json

def sqlformat(x):
	if (x is None):
		return "NULL"
	elif (type(x) == str):
		return ("'%s'" % x.replace("'", "\\'"))
	else:
		return str(x)

def lastid(tablename):
	return ("(SELECT id FROM %s ORDER BY id DESC LIMIT 1)" % (tablename))

def main():
	data = json.loads(sys.stdin.read())
	print(len(data))
	return
	for entry in data[324:]:
		query = ""
		query += ("INSERT INTO\n\tconcepts(type,description)\nVALUES\n\t(%s,%s)\n;\n" % (sqlformat(entry["type"]), sqlformat(entry["description"])))
		concept_id = lastid("concepts")
		for (iso_639_2,lemmata) in list(entry["translations"].items()):
			for lemma in lemmata:
				query += ("INSERT INTO\n\twords(language_id,lemma,pronunciation)\nVALUES\n\t((SELECT id FROM languages WHERE iso_639_2=%s),%s,%s)\n;\n" % (sqlformat(iso_639_2), sqlformat(lemma), sqlformat(None)))
				word_id = lastid("words")
				strength = None
				query += ("INSERT INTO\n\tcorrelations(concept_id,word_id,strength)\nVALUES\n\t(%s,%s,%s)\n;\n" % (concept_id, word_id, sqlformat(strength)))
		for tag in entry["tags"]:
			query += ("INSERT INTO tags(concept_id,tag) VALUES (%s,%s);\n" % (concept_id, sqlformat(tag)))
		query += "--\n"
		sys.stdout.write(query)

main()

fs-words.git

transition to sqlite-db