"""
Module with simple functions for the complete inspection procedure.
"""
import os
import webbrowser
import cPickle as pickle
import tempfile
import traceback
import warnings
import mdp
from mdp import hinet
from mdp import numx
from bimdp import BiFlow
from bihinet_translator import BIHINET_STYLE
from trace_inspection import (
_trace_biflow_training, TraceDebugException,
INSPECT_TRACE_STYLE, SLIDE_CSS_FILENAME, PICKLE_EXT,
TraceHTMLTranslator, TraceHTMLInspector,
prepare_training_inspection, remove_inspection_residues
)
from trace_slideshow import (
INSPECT_SLIDESHOW_STYLE, TrainHTMLSlideShow, ExecuteHTMLSlideShow,
SectExecuteHTMLSlideShow)
from utils import robust_write_file,robust_pickle,first_iterable_elem
# style for slides, used when the slides are not viewed in a slideshow
SLIDE_STYLE = (hinet.HINET_STYLE + BIHINET_STYLE +
INSPECT_TRACE_STYLE)
# style for slideshow, can be used when embedding the slideshow
INSPECTION_STYLE = (hinet.HINET_STYLE + BIHINET_STYLE +
INSPECT_TRACE_STYLE + INSPECT_SLIDESHOW_STYLE)
def _open_custom_brower(open_browser, url):
"""Helper function to support opening a custom browser."""
if isinstance(open_browser, str):
try:
custom_browser = webbrowser.get(open_browser)
custom_browser.open(url)
except webbrowser.Error:
err = ("Could not open browser '%s', using default." %
open_browser)
warnings.warn(err)
webbrowser.open(url)
else:
webbrowser.open(url)
def inspect_training(snapshot_path, x_samples, msg_samples=None,
stop_messages=None, inspection_path=None,
trace_inspector=None, debug=False,
slide_style=SLIDE_STYLE, show_size=False,
verbose=True, **kwargs):
"""Return the HTML code for an inspection slideshow of the training.
This function must be used after the training was completed. Before the
training prepare_training_inspection must have been called to create
snapshots. After training one should call remove_inspection_residues.
Note that the file into which the returned slideshow HTML is inserted must
be in the snapshot_path.
snapshot_path -- Path were the flow training snapshots are stored.
css_filename -- Filename of the CSS file for the slides.
x_samples, msg_samples -- Lists with the input data for the training trace.
stop_messages -- The stop msg for the training trace.
inspection_path -- Path were the slides will be stored. If None (default
value) then the snapshot_path is used.
trace_inspector -- Instance of HTMLTraceInspector, can be None for
default class.
debug -- If True (default is False) then any exception will be
caught and the gathered data up to that point is returned in the
normal way. This is useful for bimdp debugging.
slide_style -- CSS code for the individual slides (when they are
viewed as single HTML files), has no effect on the slideshow appearance.
show_size -- Show the approximate memory footprint of all nodes.
verbose -- If True (default value) a status message is printed for each
loaded snapshot.
**kwargs -- Additional arguments for flow.train can be specified
as keyword arguments.
"""
if not inspection_path:
inspection_path = snapshot_path
## create CSS file for the slides
css_filename = SLIDE_CSS_FILENAME
robust_write_file(path=inspection_path, filename=css_filename,
content=slide_style)
del slide_style
## create slides
try:
# combine all the arguments to be used as kwargs
all_kwargs = kwargs
all_kwargs.update(vars())
del all_kwargs["kwargs"]
del all_kwargs["all_kwargs"]
slide_filenames, slide_node_ids, index_table = \
_trace_biflow_training(**all_kwargs)
except TraceDebugException, debug_exception:
slide_filenames, slide_node_ids, index_table = debug_exception.result
if index_table is None:
return None # no snapshots were found
# create slideshow
slideshow = TrainHTMLSlideShow(filenames=slide_filenames,
node_ids=slide_node_ids,
index_table=index_table,
delay=500, delay_delta=100, loop=False)
return str(slideshow)
def show_training(flow, data_iterables, msg_iterables=None, stop_messages=None,
path=None, trace_inspector=None, debug=False,
show_size=False, open_browser=True, **kwargs):
"""Perform both the flow training and the training inspection.
The return value is the filename of the slideshow HTML file.
This function is more convenient than inspect_training since it includes
all required steps, but it is also less customizable. After everything
is complete the inspection slideshow is opened in the browser.
flow -- The untrained Flow or BiFlow. After this function has been called
the flow will be fully trained.
data_iterables, msg_iterables, stop_messages -- Same as for calling train
on a flow.
path -- Path were both the training snapshots and the inspection slides
will be stored. If None (default value) a temporary directory will be
used.
trace_inspector -- Instance of HTMLTraceInspector, can be None for
default class.
debug -- Ignore exception during training and try to complete the slideshow
(default value is False).
show_size -- Show the approximate memory footprint of all nodes.
open_browser -- If True (default value) then the slideshow file is
automatically opened in a webbrowser. One can also use string value
with the browser name (for webbrowser.get) to request a specific
browser.
**kwargs -- Additional arguments for flow.train can be specified
as keyword arguments.
"""
if path is None:
path = tempfile.mkdtemp(prefix='MDP_')
# get first part of data iterators as sample data for inspection
# if data_iterables is an array, wrap it up in a list
if isinstance(data_iterables, numx.ndarray):
data_iterables = [[data_iterables]] * len(flow)
x_samples = []
for i, data_iterable in enumerate(data_iterables):
if data_iterable is None:
x_sample, new_data_iterable = None, None
else:
x_sample, new_data_iterable = first_iterable_elem(data_iterable)
x_samples.append(x_sample)
data_iterables[i] = new_data_iterable
del x_sample
if msg_iterables:
msg_samples = []
for i, msg_iterable in enumerate(msg_iterables):
if msg_iterable is None:
msg_sample, new_msg_iterable = None, None
else:
msg_sample, new_msg_iterable = first_iterable_elem(msg_iterable)
msg_samples.append(msg_sample)
msg_iterables[i] = new_msg_iterable
del msg_sample
else:
msg_samples = None
# store the data to disk to disk to save memory and safeguard against
# any change made to the data during the training
robust_pickle(path, "training_data_samples.pckl",
(x_samples, msg_samples, stop_messages))
del x_samples
del msg_samples
# perform the training and gather snapshots
prepare_training_inspection(flow=flow, path=path)
try:
if isinstance(flow, BiFlow):
flow.train(data_iterables, msg_iterables, stop_messages, **kwargs)
else:
flow.train(data_iterables, **kwargs)
except Exception:
if debug:
traceback.print_exc()
print ("exception during training, " +
"inspecting up to failure point...")
# create the last snapshot manually
try:
# if a normal mdp.Flow instance was given then this fails
flow._bi_reset()
except Exception:
pass
filename = (flow._snapshot_name_ + "_%d" % flow._snapshot_counter_
+ PICKLE_EXT)
robust_pickle(flow._snapshot_path_, filename, flow)
else:
raise
remove_inspection_residues(flow)
# reload data samples
sample_file = open(os.path.join(path, "training_data_samples.pckl"), "rb")
x_samples, msg_samples, stop_messages = pickle.load(sample_file)
sample_file.close()
# create slideshow
slideshow = inspect_training(snapshot_path=path,
inspection_path=path,
x_samples=x_samples,
msg_samples=msg_samples,
stop_messages=stop_messages,
debug=debug, show_size=show_size,
verbose=False)
filename = os.path.join(path, "training_inspection.html")
title = "Training Inspection"
html_file = open(filename, 'w')
html_file.write('<html>\n<head>\n<title>%s</title>\n' % title)
html_file.write('<style type="text/css" media="screen">')
html_file.write(INSPECTION_STYLE)
html_file.write(mdp.utils.BASIC_STYLE)
html_file.write('</style>\n</head>\n<body>\n')
html_file.write('<h3>%s</h3>\n' % title)
html_file.write(slideshow)
html_file.write('</body>\n</html>')
html_file.close()
if open_browser:
_open_custom_brower(open_browser, os.path.abspath(filename))
return filename
def inspect_execution(flow, x, msg=None, target=None, path=None, name=None,
trace_inspector=None, debug=False,
slide_style=SLIDE_STYLE, show_size=False,
**kwargs):
"""Return the HTML code for an inspection slideshow of the execution
and the return value of the execution (in a tuple).
Note that the file into which the slideshow HTML is inserted must be in the
snapshot_path.
flow -- The flow for the execution.
x, msg, target -- Data for the execution, msg and target can only be used
for a BiFlow (default value is None).
path -- Path were the slideshow will be stored, if None (default value)
a temporary directory will be used.
name -- Name string to be used for the slide files.
trace_inspector -- Optionally provide a custom HTMLTraceInspector instance.
debug -- If True (default is False) then any exception will be
caught and the gathered data up to that point is returned in the
normal way. This is useful for bimdp debugging.
slide_style -- CSS code for the individual slides (when they are
viewed as single HTML files), has no effect on the slideshow appearance.
show_size -- Show the approximate memory footprint of all nodes.
**kwargs -- Additional arguments for flow.execute can be specified
as keyword arguments.
"""
if path is None:
path = tempfile.mkdtemp(prefix='MDP_')
if not name:
name = "execution_inspection"
# create CSS file for the slides
css_filename = SLIDE_CSS_FILENAME
robust_write_file(path=path, filename=css_filename,
content=slide_style)
del slide_style
if not trace_inspector:
trace_translator = TraceHTMLTranslator(show_size=show_size)
trace_inspector = TraceHTMLInspector(
trace_translator=trace_translator,
css_filename=css_filename)
# create slides
try:
slide_filenames, slide_node_ids, section_ids, result = \
trace_inspector.trace_execution(path=path,
trace_name=name,
flow=flow,
x=x, msg=msg, target=target,
debug=debug,
**kwargs)
except TraceDebugException, debug_exception:
if not debug_exception.result:
return None
traceback.print_exc()
print ("exception during excecution, " +
"create inspection up to failure point...")
slide_filenames, slide_node_ids, section_ids = debug_exception.result
result = None
# create slideshow file
if not section_ids:
slideshow = ExecuteHTMLSlideShow(filenames=slide_filenames,
node_ids=slide_node_ids,
delay=500, delay_delta=100,
loop=False)
else:
# after an exception the last section_id entry can be missing
if len(section_ids) < len(slide_filenames):
section_ids.append(section_ids[-1])
slideshow = SectExecuteHTMLSlideShow(filenames=slide_filenames,
node_ids=slide_node_ids,
section_ids=section_ids,
delay=500, delay_delta=100,
loop=False)
return str(slideshow), result
def show_execution(flow, x, msg=None, target=None, path=None, name=None,
trace_inspector=None, debug=False, show_size=False,
open_browser=True, **kwargs):
"""Write the inspection slideshow into an HTML file and open it in the
browser.
The return value is a tuple with the slideshow filename and the return
value of the execution.
flow -- The flow for the execution.
x, msg, target -- Data for the execution, msg and target can only be used
for a BiFlow (default value is None).
path -- Path were the slideshow will be stored, if None (default value)
a temporary directory will be used.
name -- A name for the slideshow.
trace_inspector -- Optionally provide a custom HTMLTraceInspector instance.
debug -- If True (default is False) then any exception will be
caught and the gathered data up to that point is returned in the
normal way. This is useful for bimdp debugging.
show_size -- Show the approximate memory footprint of all nodes.
open_browser -- If True (default value) then the slideshow file is
automatically opened in a webbrowser. One can also use string value
with the browser name (for webbrowser.get) to request a specific
browser.
**kwargs -- Additional arguments for flow.execute can be specified
as keyword arguments.
"""
if path is None:
path = tempfile.mkdtemp(prefix='MDP_')
if not name:
name = "execution_inspection"
title = "Execution Inspection"
else:
title = "Execution Inspection: " + name
filename = os.path.join(path, name + ".html")
slideshow, result = inspect_execution(
flow=flow,
path=path,
x=x, msg=msg, target=target,
name=name,
trace_inspector=trace_inspector,
debug=debug,
show_size=show_size,
**kwargs)
# inspect execution created the path if required, so no need to check here
html_file = open(filename, 'w')
html_file.write('<html>\n<head>\n<title>%s</title>\n' % title)
html_file.write('<style type="text/css" media="screen">')
html_file.write(INSPECTION_STYLE)
html_file.write(mdp.utils.BASIC_STYLE)
html_file.write('</style>\n</head>\n<body>\n')
html_file.write('<h3>%s</h3>\n' % title)
html_file.write(slideshow)
html_file.write('</body>\n</html>')
html_file.close()
if open_browser:
_open_custom_brower(open_browser, os.path.abspath(filename))
return filename, result
|