def main():
# get params
params = get_params()
project = params['project']
# define file paths
INPUT_FILE = join(project, 'data', 'postprocessed', 'KMERPHENO.txt')
pim_file = join(project, 'data', 'preprocessed', 'pheno_int_map.pkl')
fsa_file = join(project, 'data', 'postprocessed', 'scored_kmers.fsa')
kim_file = join(project, 'data', 'preprocessed', 'kmer_int_map.pkl')
scored_kmers_file = join(project, 'data', 'postprocessed',
'scored_kmers.txt')
outdir = join(project, 'data', 'postprocessed')
# create output files if they do not exist
if file_exists(fsa_file):
fsa_file = None
if file_exists(scored_kmers_file):
scored_kmers_file = None
if fsa_file or scored_kmers_file:
lock = Manager().Lock()
pim = load_pickle(pim_file)
process_file(process,
INPUT_FILE,
lock=lock,
pim=pim,
uim_file=uim_file,
fsa_file=fsa_file,
scored_unitigs_file=scored_unitigs_file)
separate_phenos(scored_kmers_file, outdir, params['separate-phenos'],
params['no-consolidate'])
def process_file(self, name):
#try to get object from database by its base name
#need to use a .is_created function since whatever file is here will get processed and no longer exist
exists = MediaFile.objects.filter(name=name).exists()
if not exists:
istv = re.search(r"S([0-9]+)E([0-9]+)", name, re.I)
if istv:
if self.debug:
print "Found a tv show: %s for season: %s, episode %s" % (
name, istv.group(1), istv.group(2))
else:
if self.debug:
print "Found something that is not a tv show %s" % name
if not self.test:
f, created = MediaFile.objects.get_or_create(name=name)
if created:
#What is this?
from utility import process_file
if process_file(name):
self.file_count += 1
else:
logger.error("file:%s" % (name, ))
else:
self.ignore_count += 1
else:
print "This is where we would process if we weren't in test mode"
else:
if self.debug:
print "We have already encountered '%s' and are skipping it" % name
self.ignore_count += 1
def process_file( self, name ):
#try to get object from database by its base name
#need to use a .is_created function since whatever file is here will get processed and no longer exist
exists = MediaFile.objects.filter(name=name).exists()
if not exists:
istv = re.search(r"S([0-9]+)E([0-9]+)", name, re.I)
if istv:
if self.debug:
print "Found a tv show: %s for season: %s, episode %s" % (name, istv.group(1), istv.group(2))
else:
if self.debug:
print "Found something that is not a tv show %s" % name
if not self.test:
f, created = MediaFile.objects.get_or_create( name=name )
if created:
#What is this?
from utility import process_file
if process_file( name ):
self.file_count += 1
else:
logger.error("file:%s" % (name,) )
else:
self.ignore_count += 1
else:
print "This is where we would process if we weren't in test mode"
else:
if self.debug:
print "We have already encountered '%s' and are skipping it" % name
self.ignore_count += 1
def main():
# load params
params = get_params()
project = params['project']
k = params['k']
# define file paths
samples_file = join(project, 'data', 'raw', params['sample'])
outfile = join(project, 'data', 'preprocessed', 'unique_kmers.txt')
catted_samples = join(project, 'data', 'preprocessed', 'samples.fa')
# check if output file exists; if so, do nothing.
if file_exists(outfile):
exit(0)
# create catted samples file if it does not exist.
if not file_exists(catted_samples):
cat_samples(samples_file, catted_samples)
# multiprocessing queue for transferring data to the main thread
q = Manager().Queue()
# invoke process(...) on catted_samples files with kwargs, for each thread
process_file(process, catted_samples, q=q, k=k)
# consolidate all threads' counters into single counter holding all kmers
counter = Counter()
while not q.empty():
counter.update(q.get())
for kmer in counter.keys():
comp = complement(kmer)
if comp in counter:
comp_count = counter[comp]
counter[comp] = 0
counter[kmer] += comp_count
counter = +counter
printd('Finished consolidating counters.')
# write counter to file
write_dict(counter, outfile, sep='\t')
# remove catted samples file
if file_exists(catted_samples):
remove(catted_samples)
def process_item(self, item, base=None):
'''processes the given item, if base != none will prepend to
item if absolute path is needed, if base == none assumes item
is absolute path or we are in cwd relative to item so we wouldnt
need it :) '''
#get absolute path for item
if base != None:
name = os.path.join(base, item)
else:
name = item
#check what type of item this is and process accordingly
if os.path.isfile(name):
if os.path.islink(name):
#store a medialink for symlinks found
medialink, created = MediaLink.objects.get_or_create(
name=item, comments=name)
#@TODO find mediafile it points to or create it
filepath = os.readlink(name)
filelocation, linkpointsto = os.path.split(filepath)
f, created = MediaFile.objects.get_or_create(name=linkpointsto)
medialink.mediafile = f
medialink.save()
else:
#try to get object from database by its base name
f, created = MediaFile.objects.get_or_create(name=item)
if created:
f.classify()
f.process()
from utility import process_file
if process_file(name):
self.file_count = self.file_count + 1
else:
logger.error("file:%s" % (name, ))
else:
self.ignore_count = self.ignore_count + 1
elif os.path.isdir(name):
#need to link it up with its parent when created here...
d, created = MediaLocation.objects.get_or_create(name=item)
if created:
d.classify()
d.process()
from utility import process_dir
if process_dir(name):
self.dir_count = self.dir_count + 1
else:
self.ignore_count = self.ignore_count + 1
self.dir_count = self.dir_count + 1
else:
self.ignore_count = self.ignore_count + 1
def process_item( self, item, base=None ):
'''processes the given item, if base != none will prepend to
item if absolute path is needed, if base == none assumes item
is absolute path or we are in cwd relative to item so we wouldnt
need it :) '''
#get absolute path for item
if base != None:
name = os.path.join(base, item )
else:
name = item
#check what type of item this is and process accordingly
if os.path.isfile( name ):
if os.path.islink( name ):
#store a medialink for symlinks found
medialink, created = MediaLink.objects.get_or_create(name=item, comments=name)
#@TODO find mediafile it points to or create it
filepath = os.readlink(name)
filelocation, linkpointsto = os.path.split(filepath)
f, created = MediaFile.objects.get_or_create(name=linkpointsto)
medialink.mediafile = f
medialink.save()
else:
#try to get object from database by its base name
f, created = MediaFile.objects.get_or_create( name=item )
if created:
f.classify()
f.process()
from utility import process_file
if process_file( name ):
self.file_count = self.file_count+1
else:
logger.error("file:%s" % (name,) )
else:
self.ignore_count = self.ignore_count + 1
elif os.path.isdir( name ):
#need to link it up with its parent when created here...
d, created = MediaLocation.objects.get_or_create( name=item )
if created:
d.classify()
d.process()
from utility import process_dir
if process_dir( name ):
self.dir_count = self.dir_count+1
else:
self.ignore_count = self.ignore_count+1
self.dir_count = self.dir_count+1
else:
self.ignore_count = self.ignore_count+1
def test():
print("Loading test data...")
start_time = time.time()
x_test, y_test = process_file(test_dir, word_to_id, config.seq_length)
session = tf.Session()
session.run(tf.global_variables_initializer())
saver = tf.train.Saver()
saver.restore(sess=session, save_path=save_path) # 读取保存的模型
print('Testing...')
loss_test, acc_test = evaluate(session, x_test, y_test)
msg = 'Test Loss: {0:>6.2}, Test Acc: {1:>7.2%}'
print(msg.format(loss_test, acc_test))
batch_size = 128
data_len = len(x_test)
num_batch = int((data_len - 1) / batch_size) + 1
# 载入标签
one_hot = read_obj('resources/one_hot_encoder.pkl')
categories = one_hot.classes_
y_pred = np.zeros(shape=(len(x_test), len(categories)),
dtype=np.int32) # 保存预测结果
for i in range(num_batch): # 逐批次处理
start_id = i * batch_size
end_id = min((i + 1) * batch_size, data_len)
feed_dict = {
model.input_x: x_test[start_id:end_id],
model.keep_prob: 1.0
}
y_pred[start_id:end_id] = session.run(tf.round(
tf.nn.sigmoid(model.logits)),
feed_dict=feed_dict)
# 评估
print("Precision, Recall and F1-Score...")
print(
metrics.classification_report(y_test, y_pred, target_names=categories))
time_dif = get_time_dif(start_time)
print("Time usage:", time_dif)
def main():
# get params
params = get_params()
project = params['project']
# define file paths
unique_kmers_file = join(project, 'data', 'preprocessed', 'unique_kmers.txt')
phenos_file = join(project, 'data', 'raw', params['pheno'])
samples_file = join(project, 'data', 'raw', params['sample'])
similarities_tsv = join(project, 'data', 'preprocessed', 'sample_similarities.tsv')
hist_orig_file = join(project, 'data', 'preprocessed', 'hist_orig.png')
hist_sim_scaled_file = join(project, 'data', 'preprocessed', 'hist_sim_scaled.png')
hist_dissim_scaled_file = join(project, 'data', 'preprocessed', 'hist_dissim_scaled.png')
similar_sample_file = join(project, 'data', 'preprocessed', 'similarSample_obs.txt')
dissimilar_sample_file = join(project, 'data', 'preprocessed', 'dissimilarSample_obs.txt')
kmer_sample_file = join(project, 'data', 'preprocessed', 'kmer_sample_map.txt')
kmer_pheno_file = join(project, 'data', 'preprocessed', 'kmer_pheno_map.txt')
sim_file = join(project, 'data', 'preprocessed', 'sample_int_map.pkl')
pim_file = join(project, 'data', 'preprocessed', 'pheno_int_map.pkl')
uim_file = join(project, 'data', 'preprocessed', 'kmer_int_map.pkl')
# create and load sample and pheno int maps
if not file_exists(sim_file):
int_maps.create_sample_int_map(samples_file, phenos_file, sim_file)
if not file_exists(pim_file):
int_maps.create_pheno_int_map(phenos_file, pim_file)
sim = load_pickle(sim_file)
# only do processing if output files do not exist
if (not file_exists(kmer_sample_file) or not file_exists(kmer_pheno_file)
or ((not file_exists(similar_sample_file) or not file_exists(dissimilar_sample_file))
and not file_exists(similarities_tsv))):
# dfs holding samples that display vs not display pheno
dfdisp, dfnodisp = create_disp_nodisp_dfs(phenos_file, sim)
# read in all sequences in input into python object
seqs = parse_input(samples_file)
# number of samples
n_samples = int(len(sim) / 2)
# upper and lower bounds for frequency of samples to filter kmers by
upper = int(params['maxkf'] * n_samples)
lower = int(params['minkf'] * n_samples)
# multiprocessing queue for transferring data to the main thread
m = Manager()
q = m.Queue()
# multiprocessing lock for locking file before writing to it
lock = m.Lock()
# kmers file name reference for subprocesses to write to
kmer_sample_file_ref = kmer_sample_file # because the int map uses it
if file_exists(kmer_sample_file):
kmer_sample_file_ref = None
if file_exists(kmer_pheno_file):
kmer_pheno_file = None
kwargs = dict(raw=seqs, k=params['k'], thresh=params['correlation-thresh'],
upper=upper, lower=lower, dfdisp=dfdisp, dfnodisp=dfnodisp,
sim=sim, n=n_samples,
kmer_sample_file=kmer_sample_file_ref,
kmer_pheno_file=kmer_pheno_file)
process_file(create_kmer_sample_map, unique_kmers_file, q=q, lock=lock, **kwargs)
sample_matrix = np.zeros((n_samples, n_samples))
num_kmers = 0
# write all chunks to output files sequentially
while not q.empty():
q_num_kmers, q_sample_matrix = q.get()
num_kmers += q_num_kmers
sample_matrix += q_sample_matrix
# create sample similarity file if the similarities tsv does not exist
if not file_exists(similar_sample_file) or not file_exists(dissimilar_sample_file):
similar_sample(sample_matrix, num_kmers, similarities_tsv,
hist_orig_file, hist_sim_scaled_file, hist_dissim_scaled_file,
similar_sample_file, dissimilar_sample_file)
if (not file_exists(similar_sample_file) or not file_exists(dissimilar_sample_file)) and file_exists(similarities_tsv):
similar_sample(None, None, similarities_tsv, hist_orig_file,
hist_sim_scaled_file, hist_dissim_scaled_file,
similar_sample_file, dissimilar_sample_file)
# create kmer int map
if not file_exists(uim_file):
int_maps.create_kmer_int_map(kmer_sample_file, uim_file)
def train():
print("Configuring TensorBoard and Saver...")
# 配置 Tensorboard，重新训练时，请将tensorboard文件夹删除，不然图会覆盖
tensorboard_dir = 'tensorboard/textrnn'
if not os.path.exists(tensorboard_dir):
os.makedirs(tensorboard_dir)
tf.summary.scalar("loss", model.loss)
tf.summary.scalar("accuracy", model.acc)
merged_summary = tf.summary.merge_all()
writer = tf.summary.FileWriter(tensorboard_dir)
# 配置 Saver
saver = tf.train.Saver()
if not os.path.exists(save_dir):
os.makedirs(save_dir)
print("Loading training and validation data...")
# 载入训练集与验证集
start_time = time.time()
x_train, y_train = process_file(train_dir, word_to_id, config.seq_length)
x_val, y_val = process_file(val_dir, word_to_id, config.seq_length)
time_dif = get_time_dif(start_time)
print("Time usage:", time_dif)
# 创建session
session = tf.Session()
session.run(tf.global_variables_initializer())
writer.add_graph(session.graph)
print('Training and evaluating...')
start_time = time.time()
total_batch = 0 # 总批次
best_acc_val = 0.0 # 最佳验证集准确率
last_improved = 0 # 记录上一次提升批次
require_improvement = 1000 # 如果超过1000轮未提升，提前结束训练
flag = False
for epoch in range(config.num_epochs):
print('Epoch:', epoch + 1)
batch_train = batch_iter(x_train, y_train, config.batch_size)
for x_batch, y_batch in batch_train:
feed_dict = feed_data(x_batch, y_batch, config.dropout_keep_prob)
if total_batch % config.save_per_batch == 0:
# 每多少轮次将训练结果写入tensorboard scalar
s = session.run(merged_summary, feed_dict=feed_dict)
writer.add_summary(s, total_batch)
if total_batch % config.print_per_batch == 0:
# 每多少轮次输出在训练集和验证集上的性能
feed_dict[model.keep_prob] = 1.0
loss_train, acc_train = session.run([model.loss, model.acc],
feed_dict=feed_dict)
loss_val, acc_val = evaluate(session, x_val, y_val)
if acc_val > best_acc_val:
# 保存最好结果
best_acc_val = acc_val
last_improved = total_batch
saver.save(sess=session, save_path=save_path)
improved_str = '*'
else:
improved_str = ''
time_dif = get_time_dif(start_time)
msg = 'Iter: {0:>6}, Train Loss: {1:>6.2}, Train Acc: {2:>7.2%},' \
+ ' Val Loss: {3:>6.2}, Val Acc: {4:>7.2%}, Time: {5} {6}'
print(
msg.format(total_batch, loss_train, acc_train, loss_val,
acc_val, time_dif, improved_str))
session.run(model.optim, feed_dict=feed_dict) # 运行优化
total_batch += 1
if total_batch - last_improved > require_improvement:
# 验证集正确率长期不提升，提前结束训练
print("No optimization for a long time, auto-stopping...")
flag = True
break # 跳出循环
if flag: # 同上
break
def main():
# get params
params = get_params()
project = params['project']
# define data paths
sim_file = join(project, 'data', 'preprocessed', 'sample_int_map.pkl')
pim_file = join(project, 'data', 'preprocessed', 'pheno_int_map.pkl')
kim_file = join(project, 'data', 'preprocessed', 'kmer_int_map.pkl')
kmer_sample_map_file = join(project, 'data', 'preprocessed',
'kmer_sample_map.txt')
kmer_pheno_map_file = join(project, 'data', 'preprocessed',
'kmer_pheno_map.txt')
phenos_file = join(project, 'data', 'raw', params['pheno'])
contains_sample_kmer_file = join(project, 'data', 'preprocessed',
'contains_obs.txt')
value_sample_pheno_file = join(project, 'data', 'preprocessed',
'samplePheno_obs.txt')
value_kmer_pheno_file = join(project, 'data', 'preprocessed',
'kmerPheno_target.txt')
similar_pheno_pheno_file = join(project, 'data', 'preprocessed',
'similarPheno_obs.txt')
sim = load_pickle(sim_file)
pim = load_pickle(pim_file)
# incorporate truth data
if params.get('truth'):
truths_infile = join(project, 'data', 'raw', params['truth'])
truths_dict = create_truths_dict(truths_infile, pim)
truth_kmer_pheno_file = join(project, 'data', 'preprocessed',
'kmerPheno_truth.txt')
else:
truths_dict = None
truth_kmer_pheno_file = None
# incorporate baseline data
if params.get('baseline'):
baseline_infile = join(project, 'data', 'raw', params['baseline'])
baseline_dict = create_truths_dict(baseline_infile, pim)
baseline_kmer_pheno_file = join(project, 'data', 'preprocessed',
'baseline_obs.txt')
else:
baseline_dict = None
baseline_kmer_pheno_file = None
# create smaller psl input files that can be efficiently done w 1 thread
if not file_exists(value_sample_pheno_file):
sample_pheno(phenos_file, sim, pim, value_sample_pheno_file)
if not file_exists(similar_pheno_pheno_file):
similar_pheno(phenos_file, pim, similar_pheno_pheno_file)
contains_exists = file_exists(contains_sample_kmer_file)
value_exists = file_exists(value_kmer_pheno_file)
truths_exists = file_exists(truth_kmer_pheno_file) if params.get(
'truth') else True
baseline_exists = file_exists(baseline_kmer_pheno_file) if params.get(
'baseline') else True
lock = Manager().Lock()
if not contains_exists:
process_file(kmer_sample_db,
kmer_sample_map_file,
kim_file=kim_file,
lock=lock,
truths=truths_dict,
contains_sample_kmer_file=contains_sample_kmer_file)
if not value_exists or not truths_exists or not baseline_exists:
if value_exists:
value_kmer_pheno_file = None
if truths_exists:
truth_kmer_pheno_file = None
if baseline_exists:
baseline_kmer_pheno_file = None
process_file(kmer_pheno_db,
kmer_pheno_map_file,
kim_file=kim_file,
value_kmer_pheno_file=value_kmer_pheno_file,
truth_kmer_pheno_file=truth_kmer_pheno_file,
lock=lock,
truths=truths_dict,
baseline=baseline_dict,
baseline_kmer_pheno_file=baseline_kmer_pheno_file)
