def test_generate(self):
dataset = Dataset("A fancy dataset")
parser = DiscogsSongParser(file_path=rel_path_to_file("../../files/releases_piece_small.xml", __file__),
dataset=dataset)
generator = get_clean_graph_generator_mongo_repos()
generator.generate_turtle_song_graph(
file_path=rel_path_to_file("../../files/out/large_discogs_release.ttl", __file__),
song_parser=parser)
def test_entity_detection(self):
parser = UsosSongParser(dataset=Dataset("some_dataset"),
source_file=rel_path_to_file("../../files/in/mini_uso.tsv", __file__))
counter = 0
for song in parser.parse_songs():
counter += 1
self.assertEquals(8, counter, "Unexpected number of songs. Expected 8 bur parsed " + str(counter))
def test_titles_detection(self):
parser = UsosSongParser(dataset=Dataset("some_dataset"),
source_file=rel_path_to_file("../../files/in/mini_uso.tsv", __file__))
expected_dict = {
"kentucky woman": False,
"why tell me why": False,
"acapella": False,
"human": False,
"carry out": False,
"don't stop the party": False,
"whistle": False,
"somebody that i used to know": False
}
unexpected = []
for song in parser.parse_songs():
if song.canonical.strip() in expected_dict:
expected_dict[song.canonical.strip()] = True
else:
unexpected.append(song.canonical)
self.assertEquals(0, len(unexpected), "Some unexpected songs where parsed: " + str(unexpected))
not_found = []
for a_expected in expected_dict:
if not expected_dict[a_expected]:
not_found.append(a_expected)
self.assertEquals(0, len(not_found), "Some expected songs were not found: " + str(not_found))
def test_writers_detection(self):
parser = UsosSongParser(dataset=Dataset("some_dataset"),
source_file=rel_path_to_file("../../files/in/mini_uso.tsv", __file__))
expected_dict = {
"Amy Heidemann": False,
"Martin Johnson": False,
"Nick Noonan": False,
"Sam Hollander": False,
"Brandon Flowers": False,
"Dave Keuning": False,
"Ronnie Vannucci": False,
"Jerome Harmon": False,
"Jim Beanz": False,
"Justin Timberlake": False,
"Timothy Clayton": False,
"Timothy Mosley": False,
"Backer": False,
"Wally De": False
}
unexpected = []
for song in parser.parse_songs():
for a_coll in song.collaborations:
if a_coll.collaborator.canonical.strip() in expected_dict:
expected_dict[a_coll.collaborator.canonical.strip()] = True
else:
unexpected.append(a_coll.collaborator.canonical)
self.assertEquals(0, len(unexpected), "Some unexpected writers where parsed: " + str(unexpected))
not_found = []
for a_expected in expected_dict:
if not expected_dict[a_expected]:
not_found.append(a_expected)
self.assertEquals(0, len(not_found), "Some expected writers were not found: " + str(not_found))
def test_artist_detection(self):
parser = UsosSongParser(dataset=Dataset("some_dataset"),
source_file=rel_path_to_file("../../files/in/mini_uso.tsv", __file__))
expected_dict = {
"Neil Diamond": False,
"Anita Meyer": False,
"Karmin": False,
"The Killers": False,
"Justin Timberlake": False,
"Timbaland": False,
"The Black Eyed Peas": False,
"Flo Rida": False,
"Gotye": False
}
unexpected = []
for song in parser.parse_songs():
for artist in song.artists:
if artist.canonical.strip() in expected_dict:
expected_dict[artist.canonical.strip()] = True
else:
unexpected.append(artist.canonical)
self.assertEquals(0, len(unexpected), "Some unexpected artists where parsed: " + str(unexpected))
not_found = []
for a_expected in expected_dict:
if not expected_dict[a_expected]:
not_found.append(a_expected)
self.assertEquals(0, len(not_found), "Some expected artist were not found: " + str(not_found))
def test_many_real_songs(self):
parser = DiscogsSongParser(file_path=rel_path_to_file("../../files/discogs_releases.xml", __file__),
dataset=Dataset(title="TestDataset"))
counter = 1
for a_song in parser.parse_songs():
counter += 1
if counter % 50000 == 0: # 50.000
break
def test_many_songs_parsed(self):
parser = UsosSongParser(dataset=Dataset("some_dataset"),
source_file=rel_path_to_file("../../files/in/bmat2heaven.tsv", __file__))
counter = 50000
for a_song in parser.parse_songs():
counter -= 1
if counter <= 0:
break
def get_executer_memory_repos_file_rdflib_graph(str_json_config):
print "loading graph"
#Graph from file
graph_path = rel_path_to_file("files/mini_usos/mini_usos_graph.ttl", __file__)
rdflib_graph = Graph()
rdflib_graph.load(graph_path, format="turtle")
mera_graph = MeraRdflibGraph(rdflib_graph)
print "loading artist ngrams"
#Memory artist repo from file
artist_ngram_path = rel_path_to_file("files/mini_usos/artist.json", __file__)
repo_artist = MemoryEntityNgrams(base_entity_uri=base_entities_URI,
type_of_entity_collection=ARTIST_COLLECTION,
load_file=artist_ngram_path)
print "loading songs ngrams"
#Memory song repo from file
song_ngram_path = rel_path_to_file("files/mini_usos/song.json", __file__)
repo_song = MemoryEntityNgrams(base_entity_uri=base_entities_URI,
type_of_entity_collection=SONG_COLLECTION,
load_file=song_ngram_path)
print "loading repo counter"
#Memory counter repo from file
counter_path = rel_path_to_file("files/mini_usos/counter.json", __file__)
repo_counter = MemoryEntityCounter(load_file=counter_path)
print "Loading matcher"
#Matcher over those structures
matcher = MeraMatcher(graph=mera_graph,
artist_ngrams_repository=repo_artist,
song_ngrams_repository=repo_song,
entity_counter_repository=repo_counter,
match_config=translate_json_to_mera_match_config(json.loads(str_json_config)))
#Formater to json
formater = FormaterToJson()
print "loading executer"
#Executer over those structures
executer = QueryExecuter(matcher=matcher,
formater=formater)
return executer
def test_json_to_mera_match_config(self):
with open(rel_path_to_file(rel_path="../../files/usos/base_config.json",
base_file=__file__), "r") as json_source:
source_content = json.load(json_source)
config_result = translate_json_to_mera_match_config(source_content)
self.assertEquals(0.65, config_result.get_minimum_of_type("artist"))
self.assertEquals(0.65, config_result.get_minimum_of_type("song"))
self.assertEquals(40, config_result.top_k_blocking_function())
self.assertEquals(5, config_result.top_k_results())
self.assertEquals(1.60, config_result.get_command_threshold("find_song"))
self.assertEquals(0.80, config_result.get_command_relevance_of_a_type(command_name="find_song",
target_type="artist"))
def test_entity_detection(self):
parser = DiscogsSongParser(file_path=rel_path_to_file("../../files/releases_piece_big.xml", __file__),
dataset=Dataset(title="TestDataset"))
counter_songs = 0
counter_writers = 0
for a_song in parser.parse_songs():
counter_songs += 1
for a_coll in a_song.collaborations:
if a_coll.role == ROLE_WRITER:
counter_writers += 1
for an_alt in a_song.alternative_titles:
print a_song.canonical, an_alt
self.assertEqual(281, counter_songs, msg="Expected 281 songs, but parsed " + str(counter_songs))
self.assertEqual(427, counter_writers, msg="Expected 427 songs with writter, but parsed " + str(counter_writers))
__author__ = 'Dani'
from wmera.query_gen.query_generator_cwr import CWRQueryGenerator
from wmera.utils import rel_path_to_file
query_gen = CWRQueryGenerator(
queries_path=rel_path_to_file("../../files/in/cwr-json-to-mera-json/works_group_full.json",
__file__),
config_path="Doesntmatteryet")
query_gen.gen_srialized_mera_json(file_path="../../files/out/cwr-json-to-mera-json/posible_queries.json")
__author__ = 'Dani'
from wmera.mera_core.model.entities import Dataset
from wmera.utils import rel_path_to_file
from test.t_utils.t_factory import get_clean_graph_generator_memory_repos
from wmera.parsers.usos.usos_song_parser import UsosSongParser
parser = UsosSongParser(dataset=Dataset("Uso_bmat2heaven"),
source_file=rel_path_to_file(
"../../files/mini_usos/mini_bmat2heaven.tsv",
__file__))
generator = get_clean_graph_generator_memory_repos()
generator.generate_turtle_song_graph(file_path=rel_path_to_file("../../files/out/usos_graph.ttl",
__file__),
song_parser=parser,
isolated=True)
generator._repo_artists.save_content(rel_path_to_file("../../files/out/artist_ngrams_usos.json",
__file__))
generator._repo_songs.save_content(rel_path_to_file("../../files/out/song_ngrams_usos.json",
__file__))
generator._repo_counter.save_content(rel_path_to_file("../../files/out/counter_usos.json",
__file__))
def test_generate_song(self):
generator = get_clean_graph_generator_mongo_repos()
generator.generate_turtle_song_graph(file_path=rel_path_to_file("../../files/out/test_song_gen.ttl", __file__),
song_parser=FakeSongParser(dataset=Dataset("A_Dataset")))
__author__ = 'Dani'
from wmera.mera_core.model.entities import Dataset
from wmera.utils import rel_path_to_file
from test.t_utils.t_factory import get_clean_graph_generator_mongo_repos
from wmera.parsers.usos.usos_song_parser import UsosSongParser
parser = UsosSongParser(dataset=Dataset("Uso_bmat2heaven"),
source_file=rel_path_to_file(
"../../files/mini_usos/mini_bmat2heaven.tsv",
__file__))
generator = get_clean_graph_generator_mongo_repos()
generator.generate_turtle_song_graph(file_path=rel_path_to_file("../../files/out/usos_graph.ttl",
__file__),
song_parser=parser,
isolated=True)
print len(art1), len(art2_3), len(art4_plus)
with open("songs_1_artist.tsv", "w") as target_file:
for song in art1:
target_file.write(serialize_song_tsv_artist(song) + "\n")
with open("songs_2_3_artist.tsv", "w") as target_file:
for song in art2_3:
target_file.write(serialize_song_tsv_artist(song) + "\n")
with open("songs_4_plus_artist.tsv", "w") as target_file:
for song in art4_plus:
target_file.write(serialize_song_tsv_artist(song) + "\n")
num_songs = 45458287
# num_songs = 100
# desired_songs = 3500
extract_random_tsv_songs(file_path=rel_path_to_file("../../files/discogs_releases.xml", __file__),
total_songs=num_songs,
desired_songs=3500)
return result
def name_included_in_min_levenshtein_distance(name, target_et):
for elem in target_et:
if levenshtein(elem, name) <= 1:
return True
return False
##################### ------------ Program ----------- #########################
print "Eh que soy yo"
clean_artist = set_of_words_from_file_separator(rel_path_to_file("files/clean_artist.txt",
__file__))
noisy_artist = set_of_words_from_file_separator(rel_path_to_file("files/noisy_artist.txt",
__file__))
clean_songs = set_of_words_from_file_separator(rel_path_to_file("files/clean_song.txt",
__file__))
noisy_songs = set_of_words_from_file_separator(rel_path_to_file("files/noisy_song.txt",
__file__))
artist_parsed = clean_artist.union(noisy_artist)
songs_parsed = clean_songs.union(noisy_songs)
noisy_queries = list_of_lists_of_words_from_file_lines(
rel_path_to_file("files/noisy-musical-queries.txt",
from wmera.utils import rel_path_to_file
__author__ = 'Dani'
from test.t_utils.t_factory import get_clean_repo_counter_memory, get_clean_repo_song_memory, \
get_clean_repo_artist_memory, get_clean_repo_artist_mongo, get_clean_repo_counter_mongo, \
get_clean_repo_songs_mongo
from wmera.adapters.in_memory_to_mogno import dump_in_memory_ngrams_into_mongo_ngrams, \
dump_in_memory_counter_into_mongo_counter
# Artist ngrmas
mongo_artist = get_clean_repo_artist_mongo()
memory_artist = get_clean_repo_artist_memory()
memory_artist.load_content(rel_path_to_file("../../files/out/artist_ngrams_usos.json",
__file__))
dump_in_memory_ngrams_into_mongo_ngrams(in_memory_repo=memory_artist,
mongo_repo=mongo_artist)
memory_artist = None # Free memory
# Song ngrams
mongo_song = get_clean_repo_songs_mongo()
memory_song = get_clean_repo_song_memory()
memory_song.load_content(rel_path_to_file("../../files/out/song_ngrams_usos.json",
__file__))
dump_in_memory_ngrams_into_mongo_ngrams(in_memory_repo=memory_song,
mongo_repo=mongo_song)
memory_song = None # Free memory
from wmera.mera_core.model.entities import Dataset
class FakeFormater(MeraFormaterInterface):
def format_mera_results(self, list_of_dicts_with_list_of_base_results):
for a_stuff in list_of_dicts_with_list_of_base_results:
print a_stuff
executer = QueryExecuter(matcher=get_mera_matcher_with_data(graph_path="../../files/out/usos_graph.ttl",
ngram_song_path="../../files/out/song_ngrams_usos.json",
ngram_artist_path="../../files/out/artist_ngrams_usos.json",
counter_path="../../files/out/counter_usos.json"),
formater=FormaterToJson())
# executer = QueryExecuter(matcher=get_empty_mera_matcher(),
# formater=FormaterToJson())
res = executer.execute_queries_from_file(
file_path=rel_path_to_file("../../files/out/cwr-json-to-mera-json/posible_queries.json",
__file__))
print res
with open("../../files/in/mera_results_mini_usos.json", "r") as file_io:
json_matches = file_io.read()
executer.introduce_json_matches_in_graph(json_matches_str=json_matches,
dataset_obj=Dataset(title="MiDatasetCWR"),
serialization_path="../../files/out/usos_graph_ENRICHED.ttl")
def test_graph_repeated_artist(self):
file_path = rel_path_to_file("../../files/out/repeated_artists_mini_graph.ttl", __file__)
artist_parser = FakeRepeatedArtistsParser(Dataset("MyTest", date="2000-feb-15"))
generator = get_clean_graph_generator_mongo_repos()
generator.generate_turtle_artist_graph(file_path, artist_parser)
result = {}
with open("../files/consultas_aol/consultas-AOL.txt") as in_file:
tmp_line = ""
counter = 0
for line in in_file:
counter += 1
if counter % 100000 == 0:
print counter
tmp_line = line.split("\t")
if len(tmp_line) == 2:
tmp_line = tmp_line[1].strip().lower()
for a_feat_var in self._feat_vars:
if a_feat_var in tmp_line:
# print tmp_line
if tmp_line in result:
result[tmp_line] += 1
else:
result[tmp_line] = 1
break
return result
##############
extractor = AolNoisyQueryExtractor(rel_path_to_file("../files/consultas_aol/", __file__))
extractor.run()
