#!/usr/bin/env python3
import time
from mongo import db
from housepy import util, strings
for session in db.sessions.find():
print("%s\t%s\t%s" % (session['_id'], util.dt(session['t'], tz="America/New_York"), strings.format_time(session['duration']) if 'duration' in session else ""))
import time, json, threading, subprocess, queue, platform, os, sys
import numpy as np
from housepy import log, config, strings, net, s3, util, process, drawing
from scipy.io import wavfile
DURATION = 10
AUDIO_TMP = os.path.abspath(os.path.join(os.path.dirname(__file__), "audio_tmp"))
t = sys.argv[1]
filename = "%s/%s.wav" % (AUDIO_TMP, t)
sample_rate, signal = wavfile.read(filename)
log.debug("samples %s" % len(signal))
log.debug("sample_rate %s" % sample_rate)
duration = float(len(signal)) / sample_rate
log.debug("duration %ss" % strings.format_time(duration))
signal = (np.array(signal).astype('float') / (2**16 * 0.5)) # assuming 16-bit PCM, -1 - 1
signal = abs(signal) # magnitude
ctx = drawing.Context()
ctx.plot(signal)
ctx.line(0, config['noise_threshold'], 1, config['noise_threshold'], stroke=(255, 0, 0))
ctx.output("screenshots")
log.debug("noise threshold is %s" % config['noise_threshold'])
log.debug("found magnitude")
content_samples = 0
for sample in signal:
if sample > config['noise_threshold']:
content_samples += 1
total_content_time = float(content_samples) / sample_rate
#!/usr/bin/env python3
import time
from mongo import db
from housepy import util, strings
for session in db.sessions.find():
print("%s\t%s\t%s" %
(session['_id'], util.dt(session['t'], tz="America/New_York"),
strings.format_time(session['duration'])
if 'duration' in session else ""))
query = {
'site': site,
't_utc': {
'$gt': timeutil.timestamp(timeutil.string_to_dt(start,
tz=config['tz'])),
'$lt': timeutil.timestamp(timeutil.string_to_dt(stop, tz=config['tz']))
}
}
log.info(query)
results = db.entries.find(query)
results = list(results)
print(json.dumps(results[0], indent=4, default=lambda d: str(d)))
ts = [d['t_utc'] for d in results]
duration = ts[-1] - ts[0]
print("DURATION %s %s" % (duration, strings.format_time(duration)))
signals = []
rates = []
labels = list(config['labels'].values())
labels.sort()
for i, label in enumerate(labels):
log.info(label)
try:
values = [d[label] if label in d else None for d in results]
values = sp.remove_shots(values, nones=True) # repair missing values
signal = sp.resample(ts, values)
num_samples = len(signal)
sample_rate = num_samples / duration
rates.append(sample_rate)
signal = sp.normalize(signal)
signal = sp.smooth(signal, 15)
# start = "2016-08-01"
# stop = "2016-08-07"
start = "2017-02-27"
stop = "2017-03-01"
print(config['sites'][site])
query = {'site': site, 't_utc': {'$gt': timeutil.timestamp(timeutil.string_to_dt(start, tz=config['tz'])), '$lt': timeutil.timestamp(timeutil.string_to_dt(stop, tz=config['tz']))}}
log.info(query)
results = db.entries.find(query)
results = list(results)
print(json.dumps(results[0], indent=4, default=lambda d: str(d)))
ts = [d['t_utc'] for d in results]
duration = ts[-1] - ts[0]
print("DURATION %s %s" % (duration, strings.format_time(duration)))
signals = []
rates = []
labels = list(config['labels'].values())
labels.sort()
for i, label in enumerate(labels):
log.info(label)
try:
values = [d[label] if label in d else None for d in results]
values = sp.remove_shots(values, nones=True) # repair missing values
signal = sp.resample(ts, values)
num_samples = len(signal)
sample_rate = num_samples / duration
rates.append(sample_rate)
signal = sp.normalize(signal)
signal = sp.smooth(signal, 15)
except Exception as e:
# log.warning(e)
continue
hrs.append(hr)
ts.append(t)
log.info("--> done")
log.info("Processing...")
total_time = ts[-1] - ts[0]
total_samples = int(total_time / 60) # once per minute
log.debug("last_t %s" % ts[-1])
log.debug("total_time %s" % total_time)
log.debug("total_time_f %s" % strings.format_time(total_time))
log.debug("total_samples %s" % total_samples)
sample_length = total_time / total_samples
log.debug("sample_length %s" % sample_length)
signal = sp.resample(ts, hrs, total_samples)
signal = sp.normalize(signal)
signal = signal - sp.smooth(signal, size=100) # flatten it out a bit
threshold = np.average(signal) + (2 * np.std(signal)) # threshold is average plus 2 std deviation
smoothed_signal = sp.smooth(signal, size=10)
peaks, valleys = sp.detect_peaks(smoothed_signal, lookahead=10, delta=.001)
max_peak = max(peaks, key=lambda p: p[1])
log.info("max_peak %s" % max_peak)
peaks = [peak for peak in peaks if peak[1] > threshold]
def draw():
