def add_file_node(path):
if not path in log['step']['known_paths']:
return
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
label = log['step']['known_paths'][path]['label']
color = '#ffffff'
if log['step']['known_paths'][path]['type'] == 'fifo':
color = '#c4f099'
elif log['step']['known_paths'][path]['type'] == 'file':
color = '#8ae234'
elif log['step']['known_paths'][path]['type'] == 'step_file':
color = '#97b7c8'
if path in log['step']['known_paths']:
if 'size' in log['step']['known_paths'][path]:
label += "\\n%s" % misc.bytes_to_str(
log['step']['known_paths'][path]['size'])
hash['nodes'][misc.str_to_sha1(path)] = {
'label': label,
'fillcolor': color
}
def add_file_node(path):
if not path in log['step']['known_paths']:
return
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
label = os.path.basename(path)
color = '#ffffff'
if log['step']['known_paths'][path]['type'] in ['fifo', 'directory']:
color = '#c4f099'
elif log['step']['known_paths'][path]['type'] == 'file':
color = '#8ae234'
elif log['step']['known_paths'][path]['type'] == 'step_file':
color = '#97b7c8'
label = log['step']['known_paths'][path]['label']
if path in log['step']['known_paths']:
if 'size' in log['step']['known_paths'][path]:
label += "\\n%s" % misc.bytes_to_str(
log['step']['known_paths'][path]['size'])
pipe_hash['nodes'][misc.str_to_sha1(path)] = {
'label': label,
'fillcolor': color
}
def file_hash(path):
if path in log['step']['known_paths']:
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
return misc.str_to_sha1(path)
def pid_hash(pid, suffix = ''):
hashtag = "%s/%s/%d/%s" % (log['step']['name'],
log['run']['run_id'],
pid, suffix)
return misc.str_to_sha1(hashtag)
def render_pipeline_hash(log):
def pid_hash(pid, suffix = ''):
hashtag = "%s/%s/%d/%s" % (log['step']['name'],
log['run']['run_id'],
pid, suffix)
return misc.str_to_sha1(hashtag)
def file_hash(path):
if path in log['step']['known_paths']:
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
return misc.str_to_sha1(path)
hash = dict()
hash['nodes'] = dict()
hash['edges'] = dict()
hash['clusters'] = dict()
hash['graph_labels'] = dict()
def add_file_node(path):
if not path in log['step']['known_paths']:
return
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
label = log['step']['known_paths'][path]['label']
color = '#ffffff'
if log['step']['known_paths'][path]['type'] == 'fifo':
color = '#c4f099'
elif log['step']['known_paths'][path]['type'] == 'file':
color = '#8ae234'
elif log['step']['known_paths'][path]['type'] == 'step_file':
color = '#97b7c8'
if path in log['step']['known_paths']:
if 'size' in log['step']['known_paths'][path]:
label += "\\n%s" % misc.bytes_to_str(
log['step']['known_paths'][path]['size'])
hash['nodes'][misc.str_to_sha1(path)] = {
'label': label,
'fillcolor': color
}
for proc_info in copy.deepcopy(log['pipeline_log']['processes']):
pid = proc_info['pid']
label = "PID %d" % pid
name = '(unknown)'
if 'name' in proc_info:
name = proc_info['name']
label = "%s" % (proc_info['name'])
if 'writes' in proc_info['hints']:
for path in proc_info['hints']['writes']:
add_file_node(path)
if 'args' in proc_info:
stripped_args = []
for arg in copy.deepcopy(proc_info['args']):
if arg in log['step']['known_paths']:
add_file_node(arg)
if arg in log['step']['known_paths']:
if log['step']['known_paths'][arg]['type'] != 'step_file':
arg = log['step']['known_paths'][arg]['label']
else:
arg = os.path.basename(arg)
else:
if arg[0:4] != '/dev':
arg = os.path.basename(arg)
if (len(arg) > 16) and re.match('^[A-Z]+$', arg):
arg = "%s[...]" % arg[:16]
stripped_args.append(arg.replace('\t', '\\t').replace(
'\\', '\\\\'))
tw = textwrap.TextWrapper(
width = 50,
break_long_words = False,
break_on_hyphens = False)
label = "%s" % ("\\n".join(tw.wrap(' '.join(stripped_args))))
if 'args' in proc_info:
output_fifo = False
for arg in proc_info['args']:
fifo_type = None
if name == 'cat4m' and arg == '-o':
output_fifo = True
elif name == 'dd' and arg.startswith('of='):
output_fifo = True
for known_path in log['step']['known_paths'].keys():
if known_path in arg:
add_file_node(known_path)
if name in ['cat4m', 'dd']:
if output_fifo:
fifo_type = 'output'
else:
fifo_type = 'input'
else:
# we can't know whether the fifo is for input or
# output, first look at the hints, then use the
# designation (if any was given)
if 'reads' in proc_info['hints'] and \
arg in proc_info['hints']['reads']:
fifo_type = 'input'
if 'writes' in proc_info['hints'] and \
arg in proc_info['hints']['writes']:
fifo_type = 'output'
if fifo_type is None:
fifo_type = log['step']['known_paths'][arg]\
['designation']
if fifo_type == 'input':
# add edge from file to proc
hash['edges'][(file_hash(arg), pid_hash(pid))] \
= dict()
elif fifo_type == 'output':
# add edge from proc to file
hash['edges'][(pid_hash(pid), file_hash(arg))] \
= dict()
if 'writes' in proc_info['hints']:
for path in proc_info['hints']['writes']:
hash['edges'][(pid_hash(pid), file_hash(path))] = dict()
# add proc
something_went_wrong = False
if 'signal' in proc_info:
something_went_wrong = True
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fce94f"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info:
label = "%s\\n(received %s%s)" % (
label,
'friendly ' \
if pid in log['pipeline_log']['ok_to_fail'] else '',
proc_info['signal_name'] if 'signal_name' in \
proc_info else 'signal %d' % proc_info['signal'])
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info['exit_code'])
else:
label = "%s\\n(no exit code)" % label
if 'max' in log['pipeline_log']['process_watcher']:
if pid in log['pipeline_log']['process_watcher']['max']:
label += "\\n%1.1f%% CPU, %s RAM (%1.1f%%)" % (
log['pipeline_log']['process_watcher']['max'][pid]\
['cpu_percent'],
misc.bytes_to_str(
log['pipeline_log']['process_watcher']['max'][pid]\
['rss']),
log['pipeline_log']['process_watcher']['max'][pid]\
['memory_percent'])
hash['nodes'][pid_hash(pid)] = {
'label': label,
'fillcolor': color
}
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
if ('exit_code' in proc_info[key]) and \
(proc_info[key]['exit_code'] == 0) and \
('length' in proc_info[key]) and \
(proc_info[key]['length'] == 0) and \
(not 'sink_full_path' in proc_info[key]):
# skip this stdout/stderr box if it leads to nothing
continue
size_label = '(empty)'
if ('length' in proc_info[key]) and \
(proc_info[key]['length'] > 0):
speed = float(proc_info[key]['length']) / (
proc_info[key]['end_time'] -
proc_info[key]['start_time']).total_seconds()
speed_label = "%s/s" % misc.bytes_to_str(speed)
size_label = "%s / %s lines (%s)" % (
misc.bytes_to_str(proc_info[key]['length']),
"{:,}".format(proc_info[key]['lines']),
speed_label)
label = "%s\\n%s" % (which, size_label)
something_went_wrong = False
if 'signal' in proc_info[key]:
something_went_wrong = True
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fdf3a7"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info[key]:
label = "%s\\n(received %s%s)" % (
label,
"friendly " if pid in \
log['pipeline_log']['ok_to_fail'] else '',
proc_info[key]['signal_name'] if 'signal_name'\
in proc_info[key] else 'signal %d' %
proc_info[key]['signal'])
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info[key]['exit_code'])
else:
label = "%s\\n(no exit code)" % label
# add proc_which
hash['nodes'][pid_hash(pid, which)] = {
'label': label,
'fillcolor': color
}
if 'sink_full_path' in proc_info[key]:
path = proc_info[key]['sink_full_path']
add_file_node(path)
for proc_info in copy.deepcopy(log['pipeline_log']['processes']):
pid = proc_info['pid']
if 'use_stdin_of' in proc_info:
other_pid = proc_info['use_stdin_of']
hash['edges'][(pid_hash(other_pid, 'stdout'), pid_hash(pid))] \
= dict()
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
other_pid = proc_info[key]['pid']
hash['edges'][(pid_hash(pid), pid_hash(pid, which))] = dict()
if 'sink_full_path' in proc_info[key]:
hash['edges'][(
pid_hash(pid, which),
file_hash(proc_info[key]['sink_full_path']))] = dict()
# define nodes which go into subgraph
step_file_nodes = dict()
for path, path_info in log['step']['known_paths'].items():
if path_info['type'] == 'step_file':
step_file_nodes[file_hash(path)] = path_info['designation']
task_name = "%s/%s" % (log['step']['name'], log['run']['run_id'])
cluster_hash = misc.str_to_sha1(task_name)
hash['clusters'][cluster_hash] = dict()
hash['clusters'][cluster_hash]['task_name'] = task_name
hash['clusters'][cluster_hash]['group'] = list()
for node in hash['nodes'].keys():
if not node in step_file_nodes:
hash['clusters'][cluster_hash]['group'].append(node)
start_time = log['start_time']
end_time = log['end_time']
duration = end_time - start_time
text = "Task: %s\\lHost: %s\\lDuration: %s\\l" % (
task_name, socket.gethostname(),
misc.duration_to_str(duration, long = True)
)
hash['graph_labels'][task_name] = text
if 'max' in log['pipeline_log']['process_watcher']:
text = "CPU: %1.1f%%, %d CORES_Requested , RAM: %s (%1.1f%%)\\l" % (
log['pipeline_log']['process_watcher']['max']['sum']['cpu_percent'],
log['step']['cores'],
misc.bytes_to_str(log['pipeline_log']['process_watcher']['max']\
['sum']['rss']),
log['pipeline_log']['process_watcher']['max']['sum']['memory_percent'])
hash['graph_labels'][task_name] += text
if 'signal' in log:
hash['graph_labels'][task_name] += "Caught signal: %s\\l" % (
process_pool.ProcessPool.SIGNAL_NAMES[log['signal']])
hash['graph_labels'][task_name] += "\\l"
return hash
def create_hash_from_annotation(log):
def pid_hash(pid, suffix = ''):
hashtag = "%s/%s/%d/%s" % (log['step']['name'],
log['run']['run_id'],
pid, suffix)
return misc.str_to_sha1(hashtag)
def file_hash(path):
if path in log['step']['known_paths']:
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
return misc.str_to_sha1(path)
pipe_hash = dict()
pipe_hash['nodes'] = dict()
pipe_hash['edges'] = dict()
pipe_hash['clusters'] = dict()
pipe_hash['graph_labels'] = dict()
def add_file_node(path):
if not path in log['step']['known_paths']:
return
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
label = os.path.basename(path)
color = '#ffffff'
if log['step']['known_paths'][path]['type'] in ['fifo', 'directory']:
color = '#c4f099'
elif log['step']['known_paths'][path]['type'] == 'file':
color = '#8ae234'
elif log['step']['known_paths'][path]['type'] == 'step_file':
color = '#97b7c8'
label = log['step']['known_paths'][path]['label']
if path in log['step']['known_paths']:
if 'size' in log['step']['known_paths'][path]:
label += "\\n%s" % misc.bytes_to_str(
log['step']['known_paths'][path]['size'])
pipe_hash['nodes'][misc.str_to_sha1(path)] = {
'label': label,
'fillcolor': color
}
for proc_info in log['pipeline_log']['processes']:
pid = proc_info['pid']
# Set name and label variable
try:
name = proc_info['name']
label = "%s" % (proc_info['name'])
except KeyError:
name = '(unknown)'
label = "PID %d" % pid
try:
# Add file nodes for every file in hints
for path in proc_info['hints']['writes']:
add_file_node(path)
pipe_hash['edges'][(pid_hash(pid), file_hash(path))] = dict()
except KeyError:
pass
try:
# Add all the info for each process to pipe_hash
stripped_args = []
is_output_file = False
for arg in proc_info['args']:
# Try to investigate how fifos are integrated in data stream
# Hier muss die Entscheidung rein ob eine Datei für Input oder
# Output genutzt wird
io_type = None
if name == 'cat':
is_output_file = False
elif name == 'dd' and arg.startswith('of='):
is_output_file = True
elif name == 'dd' and arg.startswith('if='):
is_output_file = False
elif name in ['mkdir', 'mkfifo']:
is_output_file = True
for known_path in log['step']['known_paths'].keys():
# Check if arg contains a known path ...
if known_path in arg:
# ... if so add this file to the graph
add_file_node(known_path)
# Is the process able to in-/output files?
if name in ['cat', 'dd', 'mkdir', 'mkfifo']:
if is_output_file:
io_type = 'output'
else:
io_type = 'input'
elif name == 'fix_cutadapt.py':
if arg == proc_info['args'][-2]:
io_type = 'input'
elif arg == proc_info['args'][-1]:
io_type = 'output'
elif proc_info[ proc_info['args'].index(arg) - 1 ] \
== '--R2-in':
io_type = 'input'
elif proc_info[ proc_info['args'].index(arg) - 1 ] \
== '--R2-out':
io_type = 'output'
else:
# we can't know whether the fifo is for input or
# output, first look at the hints, then use the
# designation (if any was given)
if 'reads' in proc_info['hints'] and \
arg in proc_info['hints']['reads']:
io_type = 'input'
if 'writes' in proc_info['hints'] and \
arg in proc_info['hints']['writes']:
io_type = 'output'
if io_type is None:
io_type = log['step']['known_paths'][known_path]\
['designation']
if io_type is None:
io_type = 'input'
print('io_type: %s\nknown_path: %s' %
(io_type, known_path)
)
if io_type == 'input':
# add edge from file to proc
pipe_hash['edges']\
[(file_hash(known_path),pid_hash(pid))] = dict()
elif io_type == 'output':
# add edge from proc to file
pipe_hash['edges']\
[(pid_hash(pid), file_hash(known_path))] = dict()
basename = os.path.basename(known_path)
#if log['step']['known_paths'][known_path]['type'] != \
# 'step_file':
arg = arg.replace(known_path, basename)
# break
# else:
# arg = basename
# break
else:
# if arg[0:4] != '/dev':
# arg = os.path.basename(arg)
if (len(arg) > 16) and re.match('^[A-Z]+$', arg):
arg = "%s[...]" % arg[:16]
stripped_args.append(arg.replace('\t', '\\t').replace(
'\\', '\\\\'))
tw = textwrap.TextWrapper(
width = 50,
break_long_words = False,
break_on_hyphens = False)
label = "%s" % ("\\n".join(tw.wrap(' '.join(stripped_args))))
# If any key wasn't around let's go on
except KeyError:
pass
# add proc
something_went_wrong = False
if 'signal' in proc_info:
something_went_wrong = True
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fce94f"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info:
label = "%s\\n(received %s%s)" % (
label,
'friendly ' \
if pid in log['pipeline_log']['ok_to_fail'] else '',
proc_info['signal_name'] if 'signal_name' in \
proc_info else 'signal %d' % proc_info['signal'])
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info['exit_code'])
else:
label = "%s\\n(no exit code)" % label
if 'max' in log['pipeline_log']['process_watcher']:
if pid in log['pipeline_log']['process_watcher']['max']:
label += "\\n%1.1f%% CPU, %s RAM (%1.1f%%)" % (
log['pipeline_log']['process_watcher']['max'][pid]\
['cpu_percent'],
misc.bytes_to_str(
log['pipeline_log']['process_watcher']['max'][pid]\
['rss']),
log['pipeline_log']['process_watcher']['max'][pid]\
['memory_percent'])
pipe_hash['nodes'][pid_hash(pid)] = {
'label': label,
'fillcolor': color,
'start_time': proc_info['start_time']
}
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
if ('exit_code' in proc_info[key]) and \
(proc_info[key]['exit_code'] == 0) and \
('length' in proc_info[key]) and \
(proc_info[key]['length'] == 0) and \
(not 'sink_full_path' in proc_info[key]):
# skip this stdout/stderr box if it leads to nothing
continue
size_label = '(empty)'
if ('length' in proc_info[key]) and \
(proc_info[key]['length'] > 0):
speed = float(proc_info[key]['length']) / (
proc_info[key]['end_time'] -
proc_info[key]['start_time']).total_seconds()
speed_label = "%s/s" % misc.bytes_to_str(speed)
size_label = "%s / %s lines (%s)" % (
misc.bytes_to_str(proc_info[key]['length']),
"{:,}".format(proc_info[key]['lines']),
speed_label)
label = "%s\\n%s" % (which, size_label)
something_went_wrong = False
if 'signal' in proc_info[key]:
something_went_wrong = True
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fdf3a7"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info[key]:
label = "%s\\n(received %s%s)" % (
label,
"friendly " if pid in \
log['pipeline_log']['ok_to_fail'] else '',
proc_info[key]['signal_name'] if 'signal_name'\
in proc_info[key] else 'signal %d' %
proc_info[key]['signal'])
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info[key]['exit_code'])
else:
label = "%s\\n(no exit code)" % label
# add proc_which
pipe_hash['nodes'][pid_hash(pid, which)] = {
'label': label,
'fillcolor': color
}
if 'sink_full_path' in proc_info[key]:
path = proc_info[key]['sink_full_path']
add_file_node(path)
for proc_info in copy.deepcopy(log['pipeline_log']['processes']):
pid = proc_info['pid']
if 'use_stdin_of' in proc_info:
other_pid = proc_info['use_stdin_of']
pipe_hash['edges'][(pid_hash(other_pid, 'stdout'), pid_hash(pid))] \
= dict()
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
other_pid = proc_info[key]['pid']
pipe_hash['edges'][(pid_hash(pid), pid_hash(pid, which))] = dict()
if 'sink_full_path' in proc_info[key]:
pipe_hash['edges'][(
pid_hash(pid, which),
file_hash(proc_info[key]['sink_full_path']))] = dict()
# define nodes which go into subgraph
step_file_nodes = dict()
for path, path_info in log['step']['known_paths'].items():
if path_info['type'] == 'step_file':
step_file_nodes[file_hash(path)] = path_info['designation']
task_name = "%s/%s" % (log['step']['name'], log['run']['run_id'])
cluster_hash = misc.str_to_sha1(task_name)
pipe_hash['clusters'][cluster_hash] = dict()
pipe_hash['clusters'][cluster_hash]['task_name'] = task_name
pipe_hash['clusters'][cluster_hash]['group'] = list()
for node in pipe_hash['nodes'].keys():
if not node in step_file_nodes:
pipe_hash['clusters'][cluster_hash]['group'].append(node)
start_time = log['start_time']
end_time = log['end_time']
duration = end_time - start_time
text = "Task: %s\\lHost: %s\\lDuration: %s\\l" % (
task_name, socket.gethostname(),
misc.duration_to_str(duration, long = True)
)
pipe_hash['graph_labels'][task_name] = text
if 'max' in log['pipeline_log']['process_watcher']:
text = "CPU: %1.1f%%, %d CORES_Requested , RAM: %s (%1.1f%%)\\l" % (
log['pipeline_log']['process_watcher']['max']['sum']['cpu_percent'],
log['step']['cores'],
misc.bytes_to_str(log['pipeline_log']['process_watcher']['max']\
['sum']['rss']),
log['pipeline_log']['process_watcher']['max']['sum']['memory_percent'])
pipe_hash['graph_labels'][task_name] += text
if 'signal' in log:
pipe_hash['graph_labels'][task_name] += "Caught signal: %s\\l" % (
process_pool.ProcessPool.SIGNAL_NAMES[log['signal']])
pipe_hash['graph_labels'][task_name] += "\\l"
return pipe_hash
def file_hash(path):
if path in log['step']['known_paths']:
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
return misc.str_to_sha1(path)
def pid_hash(pid, suffix=''):
hashtag = "%s/%s/%d/%s" % (log['step']['name'], log['run']['run_id'],
pid, suffix)
return misc.str_to_sha1(hashtag)
def create_hash_from_annotation(log):
def pid_hash(pid, suffix=''):
hashtag = "%s/%s/%d/%s" % (log['step']['name'], log['run']['run_id'],
pid, suffix)
return misc.str_to_sha1(hashtag)
def file_hash(path):
if path in log['step']['known_paths']:
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
return misc.str_to_sha1(path)
pipe_hash = dict()
pipe_hash['nodes'] = dict()
pipe_hash['edges'] = dict()
pipe_hash['clusters'] = dict()
pipe_hash['graph_labels'] = dict()
def add_file_node(path):
if not path in log['step']['known_paths']:
return
if 'real_path' in log['step']['known_paths'][path]:
path = log['step']['known_paths'][path]['real_path']
label = os.path.basename(path)
color = '#ffffff'
if log['step']['known_paths'][path]['type'] in ['fifo', 'directory']:
color = '#c4f099'
elif log['step']['known_paths'][path]['type'] == 'file':
color = '#8ae234'
elif log['step']['known_paths'][path]['type'] == 'step_file':
color = '#97b7c8'
label = log['step']['known_paths'][path]['label']
if path in log['step']['known_paths']:
if 'size' in log['step']['known_paths'][path]:
label += "\\n%s" % misc.bytes_to_str(
log['step']['known_paths'][path]['size'])
pipe_hash['nodes'][misc.str_to_sha1(path)] = {
'label': label,
'fillcolor': color
}
for proc_info in log['pipeline_log']['processes']:
pid = proc_info['pid']
# Set name and label variable
try:
name = proc_info['name']
label = "%s" % (proc_info['name'])
except KeyError:
name = '(unknown)'
label = "PID %d" % pid
try:
# Add file nodes for every file in hints
for path in proc_info['hints']['writes']:
add_file_node(path)
pipe_hash['edges'][(pid_hash(pid), file_hash(path))] = dict()
except KeyError:
pass
try:
# Add all the info for each process to pipe_hash
stripped_args = []
is_output_file = False
for arg in proc_info['args']:
# Try to investigate how fifos are integrated in data stream
# Hier muss die Entscheidung rein ob eine Datei für Input oder
# Output genutzt wird
io_type = None
if name == 'cat':
is_output_file = False
elif name == 'dd' and arg.startswith('of='):
is_output_file = True
elif name == 'dd' and arg.startswith('if='):
is_output_file = False
elif name in ['mkdir', 'mkfifo']:
is_output_file = True
for known_path in log['step']['known_paths'].keys():
# Check if arg contains a known path ...
if known_path in arg:
# ... if so add this file to the graph
add_file_node(known_path)
# Is the process able to in-/output files?
if name in ['cat', 'dd', 'mkdir', 'mkfifo']:
if is_output_file:
io_type = 'output'
else:
io_type = 'input'
elif name == 'fix_cutadapt.py':
if arg == proc_info['args'][-2]:
io_type = 'input'
elif arg == proc_info['args'][-1]:
io_type = 'output'
elif proc_info[ proc_info['args'].index(arg) - 1 ] \
== '--R2-in':
io_type = 'input'
elif proc_info[ proc_info['args'].index(arg) - 1 ] \
== '--R2-out':
io_type = 'output'
else:
# we can't know whether the fifo is for input or
# output, first look at the hints, then use the
# designation (if any was given)
if 'reads' in proc_info['hints'] and \
arg in proc_info['hints']['reads']:
io_type = 'input'
if 'writes' in proc_info['hints'] and \
arg in proc_info['hints']['writes']:
io_type = 'output'
if io_type is None:
io_type = log['step']['known_paths'][known_path]\
['designation']
if io_type is None:
io_type = 'input'
print('io_type: %s\nknown_path: %s' %
(io_type, known_path))
if io_type == 'input':
# add edge from file to proc
pipe_hash['edges']\
[(file_hash(known_path),pid_hash(pid))] = dict()
elif io_type == 'output':
# add edge from proc to file
pipe_hash['edges']\
[(pid_hash(pid), file_hash(known_path))] = dict()
basename = os.path.basename(known_path)
#if log['step']['known_paths'][known_path]['type'] != \
# 'step_file':
arg = arg.replace(known_path, basename)
# break
# else:
# arg = basename
# break
else:
# if arg[0:4] != '/dev':
# arg = os.path.basename(arg)
if (len(arg) > 16) and re.match('^[A-Z]+$', arg):
arg = "%s[...]" % arg[:16]
stripped_args.append(
arg.replace('\t', '\\t').replace('\\', '\\\\'))
tw = textwrap.TextWrapper(width=50,
break_long_words=False,
break_on_hyphens=False)
label = "%s" % ("\\n".join(tw.wrap(' '.join(stripped_args))))
# If any key wasn't around let's go on
except KeyError:
pass
# add proc
something_went_wrong = False
if 'signal' in proc_info:
something_went_wrong = True
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fce94f"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info:
label = "%s\\n(received %s%s)" % (
label,
'friendly ' \
if pid in log['pipeline_log']['ok_to_fail'] else '',
proc_info['signal_name'] if 'signal_name' in \
proc_info else 'signal %d' % proc_info['signal'])
elif 'exit_code' in proc_info:
if proc_info['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info['exit_code'])
else:
label = "%s\\n(no exit code)" % label
if 'max' in log['pipeline_log']['process_watcher']:
if pid in log['pipeline_log']['process_watcher']['max']:
label += "\\n%1.1f%% CPU, %s RAM (%1.1f%%)" % (
log['pipeline_log']['process_watcher']['max'][pid]\
['cpu_percent'],
misc.bytes_to_str(
log['pipeline_log']['process_watcher']['max'][pid]\
['rss']),
log['pipeline_log']['process_watcher']['max'][pid]\
['memory_percent'])
pipe_hash['nodes'][pid_hash(pid)] = {
'label': label,
'fillcolor': color,
'start_time': proc_info['start_time']
}
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
if ('exit_code' in proc_info[key]) and \
(proc_info[key]['exit_code'] == 0) and \
('length' in proc_info[key]) and \
(proc_info[key]['length'] == 0) and \
(not 'sink_full_path' in proc_info[key]):
# skip this stdout/stderr box if it leads to nothing
continue
size_label = '(empty)'
if ('length' in proc_info[key]) and \
(proc_info[key]['length'] > 0):
speed = float(proc_info[key]['length']) / (
proc_info[key]['end_time'] -
proc_info[key]['start_time']).total_seconds()
speed_label = "%s/s" % misc.bytes_to_str(speed)
size_label = "%s / %s lines (%s)" % (misc.bytes_to_str(
proc_info[key]['length']), "{:,}".format(
proc_info[key]['lines']), speed_label)
label = "%s\\n%s" % (which, size_label)
something_went_wrong = False
if 'signal' in proc_info[key]:
something_went_wrong = True
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
something_went_wrong = True
else:
something_went_wrong = True
color = "#fdf3a7"
if something_went_wrong:
if not pid in log['pipeline_log']['ok_to_fail']:
color = "#d5291a"
if 'signal' in proc_info[key]:
label = "%s\\n(received %s%s)" % (
label,
"friendly " if pid in \
log['pipeline_log']['ok_to_fail'] else '',
proc_info[key]['signal_name'] if 'signal_name'\
in proc_info[key] else 'signal %d' %
proc_info[key]['signal'])
elif 'exit_code' in proc_info[key]:
if proc_info[key]['exit_code'] != 0:
label = "%s\\n(failed with exit code %d)" % (
label, proc_info[key]['exit_code'])
else:
label = "%s\\n(no exit code)" % label
# add proc_which
pipe_hash['nodes'][pid_hash(pid, which)] = {
'label': label,
'fillcolor': color
}
if 'sink_full_path' in proc_info[key]:
path = proc_info[key]['sink_full_path']
add_file_node(path)
for proc_info in copy.deepcopy(log['pipeline_log']['processes']):
pid = proc_info['pid']
if 'use_stdin_of' in proc_info:
other_pid = proc_info['use_stdin_of']
pipe_hash['edges'][(pid_hash(other_pid, 'stdout'), pid_hash(pid))] \
= dict()
for which in ['stdout', 'stderr']:
key = "%s_copy" % which
if key in proc_info:
other_pid = proc_info[key]['pid']
pipe_hash['edges'][(pid_hash(pid), pid_hash(pid,
which))] = dict()
if 'sink_full_path' in proc_info[key]:
pipe_hash['edges'][(pid_hash(
pid, which), file_hash(
proc_info[key]['sink_full_path']))] = dict()
# define nodes which go into subgraph
step_file_nodes = dict()
for path, path_info in log['step']['known_paths'].items():
if path_info['type'] == 'step_file':
step_file_nodes[file_hash(path)] = path_info['designation']
task_name = "%s/%s" % (log['step']['name'], log['run']['run_id'])
cluster_hash = misc.str_to_sha1(task_name)
pipe_hash['clusters'][cluster_hash] = dict()
pipe_hash['clusters'][cluster_hash]['task_name'] = task_name
pipe_hash['clusters'][cluster_hash]['group'] = list()
for node in pipe_hash['nodes'].keys():
if not node in step_file_nodes:
pipe_hash['clusters'][cluster_hash]['group'].append(node)
start_time = log['start_time']
end_time = log['end_time']
duration = end_time - start_time
text = "Task: %s\\lHost: %s\\lDuration: %s\\l" % (
task_name, socket.gethostname(),
misc.duration_to_str(duration, long=True))
pipe_hash['graph_labels'][task_name] = text
if 'max' in log['pipeline_log']['process_watcher']:
text = "CPU: %1.1f%%, %d CORES_Requested , RAM: %s (%1.1f%%)\\l" % (
log['pipeline_log']['process_watcher']['max']['sum']['cpu_percent'],
log['step']['cores'],
misc.bytes_to_str(log['pipeline_log']['process_watcher']['max']\
['sum']['rss']),
log['pipeline_log']['process_watcher']['max']['sum']['memory_percent'])
pipe_hash['graph_labels'][task_name] += text
if 'signal' in log:
pipe_hash['graph_labels'][task_name] += "Caught signal: %s\\l" % (
process_pool.ProcessPool.SIGNAL_NAMES[log['signal']])
pipe_hash['graph_labels'][task_name] += "\\l"
return pipe_hash