class BookParserTest(unittest.TestCase):
def setUp(self):
self.book_parser = BookParser()
def test_load_word_data(self):
expected_dict = {
'hello': [
{'word': 'hello',
'sentence': 'he said hello',
'book': 'Treasure Island',
'chapter': '10 - Ahoy',
'filename': 'file1.html'}
]
}
self.book_parser.load_word_data(
'hello', 'file1.html', 'he said hello', 'Treasure Island', '10 - Ahoy')
self.assertItemsEqual(expected_dict, self.book_parser.get_full_dict())
def create_database():
"""
Uses BookParser to process the html book files into a data structure in memory
then iterates through and stores in a sqlite file.
"""
conn = sqlite3.connect('test.db')
cur = conn.cursor()
create_table = """
CREATE TABLE IF NOT EXISTS words
(word TEXT PRIMARY KEY, serialized_context_list TEXT)
"""
cur.execute(create_table)
insert_row = """
INSERT OR REPLACE INTO words VALUES (?, ?)
"""
book_parser = BookParser()
book_parser.initiate()
for word, context_list in book_parser.get_full_dict().iteritems():
cur.execute(insert_row, (word.decode('utf-8'), json.dumps(context_list)))
conn.commit()
conn.close()
