def _fill(csv_file_list):
"""
Fill the database given the list of data CSVs.
:param csv_file_list: List of CSV file paths where the data is.
:return: Database filled.
"""
song_added = 0
for csv_file_name in tqdm(csv_file_list, total=len(csv_file_list)):
with open(csv_file_name, 'r') as csv_file:
rows = [row for row in csv.reader(csv_file) if row][1:]
for row in rows:
try:
if len(row) == len(SCRIPT_ROW):
song_name = row[SCRIPT_ROW.index('SONG_NAME')]
artist_name = row[SCRIPT_ROW.index('ARTIST_NAME')]
if not song_service.get_song_by_name_and_artist(
song_name, artist_name):
lyrics = row[SCRIPT_ROW.index('LYRICS')]
artist_url = row[SCRIPT_ROW.index('ARTIST_URL')]
song_url = row[SCRIPT_ROW.index('SONG_URL')]
song_id = song_service.add_song(
artist_name, song_name, lyrics, artist_url,
song_url)
song_added += int(bool(song_id))
except Exception as e:
log.warn(f'Skipping row due to [{e}]')
log.warn(f'Row: {row}')
log.info(f'Songs added: [{song_added}]')
def extract_features_from_content(content):
"""
Extract features from a content representing song lyrics.
:param content: Song lyrics.
:return: Features vector representing the given content.
"""
content = word2vec.clean_content(content)
if content:
return _extract(content)
else:
log.warn(f'Content empty after cleaning it: [{content}]')
return None
def extract_features_from_song(song_id):
"""
Extract feature from a song given its database identifier.
:param song_id: Song identifier.
:return: Features vector representing the given song.
"""
song = song_service.get_song(song_id)
if song:
return _extract(song.lyrics)
else:
log.warn(f'Not song found with id: [{song_id}]')
return None
def _delete_output_folder(unzipping_output_folder):
"""
Delete all the extracted files from the input zip file.
:param unzipping_output_folder: Folder to delete.
:return: Folder removed.
"""
if unzipping_output_folder:
try:
shutil.rmtree(unzipping_output_folder)
except Exception as e:
log.warn(
f'Could not delete [{unzipping_output_folder}] due to [{e}].')
def read_maximum_distance():
"""
Read the maximum distance from the file.
:return: Maximum distance if it exists or it was generated, None otherwise.
"""
if os.path.isfile(FILE_NAME_MAXIMUM_DISTANCE):
try:
with open(FILE_NAME_MAXIMUM_DISTANCE, 'r') as file:
maximum_distance = float(str(file.read()))
return maximum_distance
except Exception as e:
log.warn(f'Error reading maximum distance: [{e}]')
return None
else:
return None
def set_index_id(song_id, index_id):
"""
Update the song instance given its identifier setting a new NMSLIB index identifier.
:param song_id: Song identifier.
:param index_id: NMSLIB index identifier.
:return: True if the update was successful, False otherwise.
"""
song = db_session().query(Song).filter_by(id=song_id).first()
if song:
song.index_id = index_id
commit_session()
return True
else:
log.warn(f'Not song found with id [{song_id}]')
return False
def _extract(lyrics):
"""
Internal function for extracting features of a given bunch of words.
:param lyrics: String contains some words representing the lyrics of a song.
:return: Features vector representing the given lyrics, None if it was not possible.
"""
lyrics = word2vec.clean_lyrics(lyrics)
lyrics = ' '.join(lyrics)
w2v_instance = word2vec.load_w2v_instance(FILE_NAME_W2V)
lyrics = word2vec.normalize(lyrics, w2v_instance)
if lyrics is not None:
lyrics = lyrics.reshape((1, NUM_FEATURES))
return lyrics
else:
log.warn('Empty lyrics after normalizing it.')
return None
def search(features, amount_results=API_SONG_SIMILARITY_LIMIT, song_id=None):
"""
Query to the index given a features vector.
:param features: Features vector needed for querying the index.
:param amount_results: Maximum amount of results to return.
:param song_id: Song identifier that represents the features vector for avoiding to return itself as a result.
:return: Query results.
"""
results = []
index_id = -1
if song_id:
song = song_service.get_song(song_id)
if song:
index_id = song.index_id
else:
log.warn(f'Not song found with id: [{song_id}]')
nmslib_index = Nmslib()
nmslib_index.load(FILE_NAME_INDEX)
query_results = nmslib_index.batch_query(features, NEIGHBOURHOOD_AMOUNT)
closest, distances = query_results[0]
maximum_distance = read_maximum_distance()
for i, dist in zip(closest, distances):
i = int(i)
dist = float(dist)
if i != index_id:
song = song_service.get_song_by_index_id(i)
if song:
result = song.serialize()
if maximum_distance:
dist = 100.0 - min(100.0,
(dist * 100.0) / maximum_distance)
dist = float(f'{dist:.2f}')
result['percentage'] = dist
results.append(result)
if len(results) >= amount_results:
break
return results
def search():
"""
Controller for searching songs from the database.
:return: JSON response.
"""
try:
# Parameters retrieving
query = request.args.get('query')
if not query:
return response.make(
error=True, message='`query` missed as a query parameter.')
query = query.strip()
if len(query) <= 2:
log.warn(f'Query is too short: [{query}]')
return response.make(error=False, response=dict(results=[]))
# Cache processing
method = search.__name__
key = '{}'.format(query)
results_cached = cache.get(method, key)
if results_cached is not None:
return response.make(response=results_cached, cached=True)
# Searching
results = song_service.get_song_by_query(query)
results = [{
'id': q.id,
'name': f'{q.artist_name} - {q.song_name}'
} for q in results]
results = sorted(results, key=lambda q: q['name'])
# Return results and refresh cache
return response.make(error=False,
response=dict(results=results),
method=method,
key=key)
except Exception as e:
log.error(f'Unexpected error: [{e}]')
log.exception(e)
return response.make(error=True, message='Unexpected error.')
