def _fingerprint_worker(filename, limit=None, song_name=None):
# Pool.imap sends arguments as tuples so we have to unpack
# them ourself.
try:
filename, limit = filename
except ValueError:
pass
songname, extension = os.path.splitext(os.path.basename(filename))
song_name = song_name or songname
channels, Fs, file_hash = decoder.read(filename, limit)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
print("Fingerprinting channel %d/%d for %s" % (channeln + 1,
channel_amount,
filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
print("Finished channel %d/%d for %s" % (channeln + 1, channel_amount,
filename))
result |= set(hashes)
return song_name, result, file_hash
def _fingerprint_worker(filename, limit=None, song_name=None):
# Pool.imap sends arguments as tuples so we have to unpack
# them ourself.
try:
filename, limit = filename
except ValueError:
pass
songname, extension = os.path.splitext(os.path.basename(filename))
song_name = song_name or songname
channels, Fs, file_hash, audio_length = decoder.read(filename, limit)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
print("Fingerprinting channel %d/%d for %s" % (channeln + 1,
channel_amount,
filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
# Added by Sheng Zhao for debug issue https://github.com/worldveil/dejavu/issues/142
#print("Type of hashes in _fingerprint_worker: %s" % (type(hashes)))
#print(hashes)
print("Finished channel %d/%d for %s" % (channeln + 1, channel_amount,
filename))
result |= set(hashes)
return song_name, result, file_hash, audio_length
def _fingerprint_worker(filename, limit=None, song_name=None):
# Pool.imap sends arguments as tuples so we have to unpack
# them ourself.
try:
filename, limit = filename
except ValueError:
pass
songname, extension = os.path.splitext(os.path.basename(filename))
song_name = song_name or songname
channels, Fs, file_hash = decoder.read(filename, limit)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
# print("Fingerprinting channel %d/%d for %s" % (channeln + 1,
# channel_amount,
# filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
# print("Finished channel %d/%d for %s" % (channeln + 1, channel_amount,
# filename))
result |= set(hashes)
return song_name, result, file_hash
def _fingerprint_worker(filename, limit=None, song_name=None):
# Pool.imap sends arguments as tuples so we have to unpack
# them ourself.
try:
filename, limit = filename
except ValueError:
pass
songname, extension = os.path.splitext(os.path.basename(filename))
song_name = song_name or songname
channels, Fs, file_hash, audio_length = decoder.read(filename, limit)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
logging.getLogger('dejavu').info(
"Fingerprinting channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
logging.getLogger('dejavu').debug(
"Finished channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
result |= set(hashes)
return song_name, result, file_hash, audio_length
def _fingerprint_worker(filename, db, limit):
song_name, extension = os.path.splitext(os.path.basename(filename))
channels, Fs = decoder.read(filename, limit)
# insert song into database
sid = db.insert_song(song_name)
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
print("Fingerprinting channel %d/%d for %s" % (channeln + 1,
channel_amount,
filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
print("Finished channel %d/%d for %s" % (channeln + 1, channel_amount,
filename))
print("Inserting fingerprints for channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
db.insert_hashes(sid, hashes)
print("Finished inserting for channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
print("Marking %s finished" % (filename,))
db.set_song_fingerprinted(sid)
print("%s finished" % (filename,))
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
return match
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
logger.debug("Recognizig using file. Match: %s" % match)
return match
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
return match
def recognize_file(self, filename,start=None,limit=None):
#print limit
frames, self.Fs, file_hash = decoder.read(filename,start,limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
return match
def recognize_file(self, filename):
with Timer("decorder.read") as t:
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
with Timer() as t:
match = self._recognize(*frames)
if match:
match['match_time'] = t.ms
return match
def fingerprint_file(self, filepath, song_name=None):
channels, Fs = decoder.read(filepath)
if not song_name:
print "Song name: %s" % song_name
song_name = decoder.path_to_songname(filepath)
song_id = self.db.insert_song(song_name)
for data in channels:
hashes = fingerprint.fingerprint(data, Fs=Fs)
self.db.insert_hashes(song_id, hashes)
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
print "sample rate is %d \n" %self.Fs
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match_time = t
return match, match_time
def recognize_file(self, filename):
frames, self.Fs = decoder.read(filename, self.dejavu.limit)
t = time.time()
matches = self._recognize(*frames)
t = time.time() - t
if matches:
for match in matches:
match['match_time'] = t / len(matches)
return matches
def recognize_file(self, filename):
# 获得filename的每个通道的数据，是个list; 采样帧率; sha1唯一编码
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
return match
def recognize_file(self, filename):
frames, self.Fs = decoder.read(filename, self.dejavu.limit)
t = time.time()
matches = self._recognize(*frames)
t = time.time() - t
if matches:
for match in matches:
match['match_time'] = t/len(matches)
return matches
def fp_printing(filename):
songname, extension = os.path.splitext(os.path.basename(filename))
channels, Fs, file_hash = decoder.read(filename, None)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
print("Fingerprinting channel %d/%d for %s" % (channeln + 1, channel_amount, filename))
hashes = fingerprint.fingerprint(channel, Fs=Fs)
print("Finished channel %d/%d for %s" % (channeln + 1, channel_amount, filename))
result |= set(hashes)
return song_name, result, file_hash
def recognize_file(self, filename, limit):
if not limit:
limit = self.dejavu.limit
frames, self.Fs = decoder.read(filename, limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
if match:
match['match_time'] = t
return match
def generate_hashes(mp3file):
channels, Fs, file_hash = decoder.read(mp3file, None)
chunk_sz = Fs * 2048
all_hashes = deque()
begin_chunk = 0
for chunk in chunker(channels[0], chunk_sz):
hashes = fingerprint.fingerprint(chunk, Fs=Fs, plot=False)
for hash, t_offset in hashes:
all_hashes.append((
hash,
t_offset + int(begin_chunk / 2048),
))
begin_chunk += chunk_sz
return all_hashes
def recognize_file(self, filename):
frames, self.Fs, file_hash, song_length = decoder.read(filename, self.dejavu.limit)
t = time.time()
hashes, match = self._recognize(*frames)
t = time.time() - t
if match:
if self.dejavu.multiple_match:
for track in match:
track['match_time'] = t/len(match)
else:
match['match_time'] = t
return hashes, match
def recognize_file(self, filename):
frames, self.Fs, file_hash, song_length = decoder.read(
filename, self.dejavu.limit)
t = time.time()
hashes, match = self._recognize(*frames)
t = time.time() - t
if match:
if self.dejavu.multiple_match:
for track in match:
track['match_time'] = t / len(match)
else:
match['match_time'] = t
return hashes, match
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
t = time.time()
match = self._recognize(*frames)
t = time.time() - t
msg = ("recognize_file match %s" % match)
msg = "[" + datetime.datetime.now().strftime(
"%Y-%m-%d %H:%M:%S") + "]" + msg
print msg
logging.info(msg)
if match:
match['match_time'] = t
return match
def get_fingerprints(file_path, limit=None, song_name=None):
#Fingerprints a file and send the number of fingerprints that found
try:
filename, limit = file_path
except ValueError:
pass
songname, extension = os.path.splitext(os.path.basename(file_path))
song_name = song_name or songname
channels, Fs = decoder.read(file_path, limit)
result = set()
channel_amount = len(channels)
for channeln, channel in enumerate(channels):
hashes = fingerprint.fingerprint(channel, Fs=Fs)
result |= set(hashes)
return result
def __init__(self, video_name):
"""
Takes the video name. Creates a Dejavu object.
Also obtains the duration and number of frames in the video.
"""
file_type = mimetypes.guess_type(video_name)[0]
if file_type[:3] != "vid":#The file is not a video file
print "No video file found"
raise Exception(INCORRECT_VIDEO_FILE_ERROR)
self.video_name = video_name
self.djv = Dejavu(CONFIG)
#We create the audio of the video given to us
ffmpeg.create_audio(self.video_name, TEMP_AUDIO)
self.frames, self.Fs, hash_val = decoder.read(TEMP_AUDIO)
self.frames = self.frames[0] #Since we always create single channel audio
self.duration = int(self.frames.shape[0] / (self.Fs*1.0)) #Duration of the entire video in seconds
def __init__(self, video_name):
"""
Takes the video name. Creates a Dejavu object.
Also obtains the duration and number of frames in the video.
"""
file_type = mimetypes.guess_type(video_name)[0]
if file_type[:3] != "vid":#The file is not a video file
print "No video file found"
raise Exception(INCORRECT_VIDEO_FILE_ERROR)
self.video_name = video_name
self.djv = Dejavu(CONFIG)
#We create the audio of the video given to us
ffmpeg.create_audio(self.video_name, TEMP_AUDIO)
self.frames, self.Fs, hash_val = decoder.read(TEMP_AUDIO)
self.frames = self.frames[0] #Since we always create single channel audio
self.duration = int(self.frames.shape[0] / (self.Fs*1.0)) #Duration of the entire video in seconds
def _fingerprint_worker(filename, limit=None, song_name=None):
'''对音频进行编码的总函数
Args:
filename: 将要fingerprint的音频(路劲/xxx.mp3)
limit: 限制每个音频编码多少秒
song_name: 音频的name
Returns:
song_name: 音频的name(不含后缀名)
result: 整个音频使用函数fingerprint()编码的结果，set()
file_hash: 整个音频使用sha1编码的结果
'''
# Pool.imap sends arguments as tuples so we have to unpack
# them ourself.
try:
filename, limit = filename
except ValueError:
pass
songname, extension = os.path.splitext(
os.path.basename(filename)) # 分离歌名和后缀名
song_name = song_name or songname # or 找到第一个非空/逻辑非的对象
# channels: list，每个元素是一个声道的数据
# Fs: 音频的采样频率
# file_hash: 该音频的sha1编码
channels, Fs, file_hash = decoder.read(filename, limit)
result = set()
channel_amount = len(channels) # 音频的声道数
for channeln, channel in enumerate(channels):
# TODO: Remove prints or change them into optional logging.
print("Fingerprinting channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
hashes = fingerprint(channel,
Fs=Fs) # 对音频进行编码，但是file_hash不是已经用sha1编码完成了吗
print("Finished channel %d/%d for %s" %
(channeln + 1, channel_amount, filename))
result |= set(hashes) # |= 类似与list的append
return song_name, result, file_hash
def recognize_file(self, filename):
frames, self.Fs, file_hash = decoder.read(filename, self.dejavu.limit)
matches = self._recognize(*frames)
return matches