def run_post_p_wt(
output_folder='/home/alex/LabGit/ECG_random_walk/experiments/record_test/result0607/post_p/'
):
import glob, json
from mcmc.post_p import post_p_wt
def fix_P_annots(raw_sig, annots):
P_annots = filter(lambda x: x[1][0] == 'P', annots)
annots = filter(lambda x: x[1][0] != 'P', annots)
P_annots = post_p_wt(raw_sig, P_annots, 500.0)
annots.extend(P_annots)
return annots
files = glob.glob(
'/home/alex/LabGit/ECG_random_walk/experiments/record_test/result0607/*.json'
)
avg_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/avg/'
cloader = ECGLoader(2, 1)
for result_file_name in files:
cID = os.path.split(result_file_name)[-1]
cID = cID.split('.')[0]
print 'Ploting cID:', cID
with open(result_file_name, 'r') as fin:
annots = json.load(fin)
# with open(avg_result_folder + '%s.json' % cID, 'r') as fin:
# avg_annots = json.load(fin)
raw_sig = cloader.loadID(cID)
annots = fix_P_annots(raw_sig, annots)
# with open(output_folder + '%s.json' % cID, 'w') as fout:
# json.dump(annots, fout)
# plot_result(raw_sig, annots)
pdb.set_trace()
def debug_plot_result():
import glob, json
files = glob.glob(
'/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/avg/*.json'
)
avg_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/'
cloader = ECGLoader(2, 1)
for result_file_name in files:
cID = os.path.split(result_file_name)[-1]
cID = cID.split('.')[0]
print 'Ploting cID:', cID
with open(result_file_name, 'r') as fin:
hiking_annots = json.load(fin)
with open(avg_result_folder + '%s.json' % cID, 'r') as fin:
avg_annots = json.load(fin)
raw_sig = cloader.loadID(cID)
# plot_result(raw_sig, annots)
fig, ax = plt.subplots(1, 1, figsize=(25, 6))
plt.plot(raw_sig)
plt.grid(True)
hiking_annots = filter(lambda x: x[0] >= 0 and x[0] < len(raw_sig),
hiking_annots)
plotExpertLabels(ax,
raw_sig,
hiking_annots,
color_in='r',
label_prefix='Normal ')
plotExpertLabels(ax, raw_sig, avg_annots, label_prefix='Hiking ')
plt.title(u'长庚 ' + cID)
plt.xlim((750, len(raw_sig) - 750))
# plt.savefig('/home/alex/图片/work_pngs/RandomWalkPath/hiking/hiking_%s.png' % cID)
plt.show(block=False)
pdb.set_trace()
def debug_plot_result_post_p():
import glob, json
from mcmc.post_p import post_p_mcmc
def fix_P_annots(raw_sig, annots):
P_annots = filter(lambda x: x[1][0] == 'P', annots)
annots = filter(lambda x: x[1][0] != 'P', annots)
P_annots = post_p_mcmc(raw_sig, P_annots, 500.0)
annots.extend(P_annots)
return annots
files = glob.glob(
'/home/alex/LabGit/ECG_random_walk/experiments/record_test/result0607/*.json'
)
postp_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/result0607/post_p/'
cloader = ECGLoader(2, 1)
fig, ax = plt.subplots(1, 1, figsize=(25, 6))
for result_file_name in files:
cID = os.path.split(result_file_name)[-1]
cID = cID.split('.')[0]
print 'Ploting cID:', cID
with open(result_file_name, 'r') as fin:
annots = json.load(fin)
if os.path.exists(postp_result_folder + '%s.json' % cID) == False:
continue
with open(postp_result_folder + '%s.json' % cID, 'r') as fin:
postp_annots = json.load(fin)
raw_sig = cloader.loadID(cID)
ax.cla()
plt.plot(raw_sig)
plotExpertLabels(ax, raw_sig, annots, label_prefix='Normal')
postp_annots = filter(lambda x: x[1][0] == 'P', postp_annots)
plotExpertLabels(ax,
raw_sig,
postp_annots,
color_in='y',
label_prefix='post_p ')
plt.title(u'长庚 ' + cID)
plt.xlim((750, len(raw_sig) - 750))
plt.grid(True)
# plt.savefig('/home/alex/图片/work_pngs/RandomWalkPath/hiking/hiking_%s.png' % cID)
plt.show(block=False)
pdb.set_trace()
def debug_plot_result():
import glob, json
files = glob.glob(
'/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/*.json'
)
avg_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/avg/'
cloader = ECGLoader(2, 1)
for result_file_name in files:
cID = os.path.split(result_file_name)[-1]
cID = cID.split('.')[0]
print 'Ploting cID:', cID
with open(result_file_name, 'r') as fin:
annots = json.load(fin)
with open(avg_result_folder + '%s.json' % cID, 'r') as fin:
avg_annots = json.load(fin)
raw_sig = cloader.loadID(cID)
test_tools.plot_result(raw_sig, annots)
def TestChanggeng(debug=False):
import json
from randomwalk.test_api import Testing
with open(
'/home/alex/LabGit/ECG_random_walk/tools/annotations/inputs/IDlist.json',
'r') as fin:
# with open('/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/normal/IDlist.json', 'r') as fin:
IDlist = json.load(fin)
avg_save_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/result0607/'
hiking_save_result_folder = '/home/alex/LabGit/ECG_random_walk/experiments/record_test/hiking/QRS_bias/hiking/'
cloader = ECGLoader(1, 1)
for cID in IDlist:
print 'Testing cID:', cID
raw_sig = cloader.loadID(cID)
model_folder = '/home/alex/LabGit/ECG_random_walk/randomwalk/data/annots0605/db2/'
annots = randomwalk.TestingWithModelFolder(raw_sig, 500.0,
model_folder)
with open(avg_save_result_folder + '%s.json' % cID, 'w') as fout:
json.dump(annots, fout, indent=4)
def testing(
changgengID='8790',
leadname='II',
model_folder='/home/alex/LabGit/ECG_random_walk/randomwalk/data/annots0605/db2/',
test_method='normal'):
cloader = ECGLoader(1, 1)
sig = cloader.loadID(changgengID, leadname=leadname)
fs = 500.0
sig = scipy.signal.resample(sig, int(len(sig) / float(fs) * 250.0))
fs_inner = 250.0
walker = RandomWalker(target_label='T',
random_pattern_file_name=model_folder +
'random_pattern.json')
walker.load_model(model_folder + 'T.mdl')
feature_extractor = walker.GetFeatureExtractor(sig)
def test_seed(seed_position, label='T', plot_path=True):
path = walker.testing_walk_extractor(feature_extractor,
seed_position,
iterations=1000)
path_len = len(path)
pos_list = [x[0] for x in path]
output_position = np.mean(pos_list[path_len / 2:])
if plot_path:
plt.figure(1)
plt.clf()
plt.subplot(2, 1, 1)
plt.plot(sig)
amp_list = np.linspace(0, -2, len(pos_list))
plt.plot(pos_list, amp_list, 'r', alpha=0.5, lw=2)
plt.xlim((min(pos_list) - 200, max(pos_list) + 200))
plt.plot(output_position,
sig[output_position],
'ro',
markersize=12)
# Path and confidence
# plt.figure(2)
# plt.clf()
plt.subplot(2, 1, 2)
plt.plot(pos_list, label='Path')
m1 = min(pos_list)
m2 = max(pos_list)
confidence_list = [x[1] * (m2 - m1) + m1 for x in path]
plt.plot(
confidence_list,
label='Confidence',
alpha=0.9,
)
plt.show(block=False)
pdb.set_trace()
# plt.savefig('/home/alex/图片/work_pngs/RandomWalkPath/confidence-%s.png' % seed_position)
return (output_position, label)
from hiking_test import hiking
def hiking_test(seed_position, label='T'):
path = walker.testing_walk_extractor(feature_extractor,
seed_position,
iterations=1000)
pos = hiking(path)
return (pos, label)
def avg_test(seed_position, label='T'):
path = walker.testing_walk_extractor(feature_extractor,
seed_position,
iterations=1000)
pos_list = [x[0] for x in path]
pos = np.mean(pos_list[len(pos_list) / 2:])
return (pos, label)
dpi = DPI(debug_info=dict())
r_list = dpi.QRS_Detection(sig, fs=fs_inner)
# annots = list()
annots = zip(r_list, [
'R',
] * len(r_list))
# for seed in xrange(100, len(sig), 200):
for rpos in r_list:
seed = rpos + fs_inner * 0.26
if seed >= len(sig):
continue
# print 'testing seed:', seed
if test_method == 'hiking':
annots.append(hiking_test(seed))
elif test_method == 'normal':
annots.append(avg_test(seed))
elif test_method == 'test_seed':
annots.append(test_seed(seed))
# plot_result(sig, annots)
return annots
def show_result_ECG(cID='8790'):
loader = ECGLoader(1, 1)
sig = loader.loadID(cID)
plot_result(sig, annots)
def testing(
changgengID='8790',
leadname='II',
model_folder='/home/alex/LabGit/ECG_random_walk/randomwalk/data/annots0605/db2/'
):
cloader = ECGLoader(1, 1)
sig = cloader.loadID(changgengID, leadname=leadname)
fs = 500.0
# sig = scipy.signal.resample(sig, int(len(sig) / float(fs) * 250.0))
walker = RandomWalker(target_label='T',
random_pattern_file_name=model_folder +
'random_pattern.json')
walker.load_model(model_folder + 'T.mdl')
feature_extractor = walker.GetFeatureExtractor(sig)
def test_seed(seed_position, label='T', plot_path=True):
path = walker.testing_walk_extractor(feature_extractor,
seed_position,
iterations=1000)
path_len = len(path)
pos_list = [x[0] for x in path]
output_position = np.mean(pos_list[path_len / 2:])
if plot_path:
plt.figure(1)
plt.clf()
plt.subplot(2, 1, 1)
plt.plot(sig)
amp_list = np.linspace(0, -2, len(pos_list))
plt.plot(pos_list, amp_list, 'r', alpha=0.5, lw=2)
plt.xlim((min(pos_list) - 200, max(pos_list) + 200))
plt.plot(output_position,
sig[output_position],
'ro',
markersize=12)
plt.plot(hiking(path),
sig[output_position],
'sm',
markersize=13,
alpha=0.5)
# Path and confidence
# plt.figure(2)
# plt.clf()
plt.subplot(2, 1, 2)
plt.plot(pos_list, label='Path')
m1 = min(pos_list)
m2 = max(pos_list)
confidence_list = [x[1] * (m2 - m1) + m1 for x in path]
plt.plot(
confidence_list,
label='Confidence',
alpha=0.9,
)
plt.figure(2)
plt.clf()
plt.hist(pos_list, bins=30)
plt.show(block=False)
pdb.set_trace()
return (output_position, label)
annots = list()
for seed in xrange(100, len(sig), 250):
print 'testing seed:', seed
annots.append(test_seed(seed))
test_tools.plot_result(sig, annots)