def analyze(video_id):
print "Beginning audio analysis...\n"
file_name = audio_scraper.get_wav_from_vid(video_id)
decision_tree = tree.generate()
audio_file = AudioFile.open(file_name)
frames = audio_file.frames(16384)
save_spectrum_image(audio_file)
analysis = []
i = 0
for frame in frames:
#print "Processing [audio frame {}]...".format(i)
frequencies, levels = get_power_spectrum(frame)
data = training_data.generate(frequencies, levels, file_name)
try:
analysis.append(decision_tree.predict(data))
except ValueError as e:
print "Encountered error: {0}".format(e)
i += 1
print "Audio analysis complete.\n"
return analysis, frames
def generate_ai():
ais = []
prices, results = generate()
for n in range(50):
indexes = []
for m in range(3):
indexes.append(random.randint(0, 7))
ais.append(
AI({
'AAPL': [],
'AMD': [],
'AMZN': [],
"INTC": [],
"MSFT": [],
"CSCO": [],
"GPRO": [],
"NVDA": []
}))
for each in ais:
each.train(
[prices[indexes[0]], prices[indexes[1]], prices[indexes[2]]],
[results[indexes[0]], results[indexes[1]], results[indexes[2]]],
1000 * random.randint(4, 10))
print("Generation finished!")
return ais
def analyze(video_id):
print "Beginning audio analysis...\n"
file_name = audio_scraper.get_wav_from_vid(video_id)
decision_tree = tree.generate()
audio_file = AudioFile.open(file_name)
frames = audio_file.frames(16384)
save_spectrum_image(audio_file)
analysis = []
i = 0
for frame in frames:
# print "Processing [audio frame {}]...".format(i)
frequencies, levels = get_power_spectrum(frame)
data = training_data.generate(frequencies, levels, file_name)
try:
analysis.append(decision_tree.predict(data))
except ValueError as e:
print "Encountered error: {0}".format(e)
i += 1
print "Audio analysis complete.\n"
return analysis, frames
def build_dataset():
dataset = []
class_labels = []
audacity_files = glob.glob(os.getcwd() + '/audio/audio_data/*.txt')
for file_name in audacity_files:
if "SKIP" in file_name:
continue
frequencies = []
levels = []
with open(file_name, 'r') as file:
category = file.readline()
class_labels.append(category)
lines = file.readlines()
for line in lines:
line = line.split(",")
frequencies.append(float(line[0]))
levels.append(float(line[1]))
data = training_data.generate(frequencies, levels, file_name)
dataset.append(data)
return dataset, class_labels