def ccopen(source, *args, **kwargs):
"""Guess the identity of a particular log file and return an instance of it.
Inputs:
source - a single logfile, a list of logfiles (for a single job),
an input stream, or an URL pointing to a log file.
*args, **kwargs - arguments and keyword arguments passed to filetype
Returns:
one of ADF, DALTON, GAMESS, GAMESS UK, Gaussian, Jaguar,
Molpro, MOPAC, NWChem, ORCA, Psi3, Psi/Psi4, QChem, CJSON or None
(if it cannot figure it out or the file does not exist).
"""
inputfile = None
is_stream = False
# Check if source is a link or contains links. Retrieve their content.
# Try to open the logfile(s), using openlogfile, if the source is a string (filename)
# or list of filenames. If it can be read, assume it is an open file object/stream.
if isinstance(source, pathlib.PurePath):
source = str(source)
if isinstance(source, pathlib.PurePath)\
and all([isinstance(s, pathlib.PurePath) for s in source]):
source = [str(item) for item in source]
is_string = isinstance(source, str)
is_url = True if is_string and URL_PATTERN.match(source) else False
is_listofstrings = isinstance(source, list) and all([isinstance(s, str) for s in source])
if is_string or is_listofstrings:
# Process links from list (download contents into temporary location)
if is_listofstrings:
filelist = []
for filename in source:
if not URL_PATTERN.match(filename):
filelist.append(filename)
else:
try:
response = urlopen(filename)
tfile = NamedTemporaryFile(delete=False)
tfile.write(response.read())
# Close the file because Windows won't let open it second time
tfile.close()
filelist.append(tfile.name)
# Delete temporary file when the program finishes
atexit.register(os.remove, tfile.name)
except (ValueError, URLError) as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
source = filelist
if not is_url:
try:
inputfile = logfileparser.openlogfile(source)
except IOError as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
else:
try:
response = urlopen(source)
is_stream = True
# Retrieve filename from URL if possible
filename = re.findall(r"\w+\.\w+", source.split('/')[-1])
filename = filename[0] if filename else ""
inputfile = logfileparser.openlogfile(filename, object=response.read())
except (ValueError, URLError) as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
elif hasattr(source, "read"):
inputfile = source
is_stream = True
# Streams are tricky since they don't have seek methods or seek won't work
# by design even if it is present. We solve this now by reading in the
# entire stream and using a StringIO buffer for parsing. This might be
# problematic for very large streams. Slow streams might also be an issue if
# the parsing is not instantaneous, but we'll deal with such edge cases
# as they arise. Ideally, in the future we'll create a class dedicated to
# dealing with these issues, supporting both files and streams.
if is_stream:
try:
inputfile.seek(0, 0)
except (AttributeError, IOError):
contents = inputfile.read()
try:
inputfile = io.StringIO(contents)
except:
inputfile = io.StringIO(unicode(contents))
inputfile.seek(0, 0)
# Proceed to return an instance of the logfile parser only if the filetype
# could be guessed. Need to make sure the input file is closed before creating
# an instance, because parsers will handle opening/closing on their own.
filetype = guess_filetype(inputfile)
# If the input file isn't a standard compchem log file, try one of
# the readers, falling back to Open Babel.
if not filetype:
if kwargs.get("cjson"):
filetype = readerclasses['cjson']
elif source and not is_stream:
ext = os.path.splitext(source)[1][1:].lower()
for extension in readerclasses:
if ext == extension:
filetype = readerclasses[extension]
# Proceed to return an instance of the logfile parser only if the filetype
# could be guessed. Need to make sure the input file is closed before creating
# an instance, because parsers will handle opening/closing on their own.
if filetype:
# We're going to close and reopen below anyway, so this is just to avoid
# the missing seek method for fileinput.FileInput. In the long run
# we need to refactor to support for various input types in a more
# centralized fashion.
if is_listofstrings:
pass
else:
inputfile.seek(0, 0)
if not is_stream:
if is_listofstrings:
if filetype == Turbomole:
source = sort_turbomole_outputs(source)
inputfile.close()
return filetype(source, *args, **kwargs)
return filetype(inputfile, *args, **kwargs)
def ccopen(source, *args, **kwargs):
"""Guess the identity of a particular log file and return an instance of it.
Inputs:
source - a single logfile, a list of logfiles (for a single job),
an input stream, or an URL pointing to a log file.
*args, **kwargs - arguments and keyword arguments passed to filetype
Returns:
one of ADF, DALTON, GAMESS, GAMESS UK, Gaussian, Jaguar,
Molpro, MOPAC, NWChem, ORCA, Psi3, Psi/Psi4, QChem, CJSON or None
(if it cannot figure it out or the file does not exist).
"""
inputfile = None
is_stream = False
# Check if source is a link or contains links. Retrieve their content.
# Try to open the logfile(s), using openlogfile, if the source is a string (filename)
# or list of filenames. If it can be read, assume it is an open file object/stream.
is_string = isinstance(source, str)
is_url = True if is_string and URL_PATTERN.match(source) else False
is_listofstrings = isinstance(source, list) and all([isinstance(s, str) for s in source])
if is_string or is_listofstrings:
# Process links from list (download contents into temporary location)
if is_listofstrings:
filelist = []
for filename in source:
if not URL_PATTERN.match(filename):
filelist.append(filename)
else:
try:
response = urlopen(filename)
tfile = NamedTemporaryFile(delete=False)
tfile.write(response.read())
# Close the file because Windows won't let open it second time
tfile.close()
filelist.append(tfile.name)
# Delete temporary file when the program finishes
atexit.register(os.remove, tfile.name)
except (ValueError, URLError) as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
source = filelist
if not is_url:
try:
inputfile = logfileparser.openlogfile(source)
except IOError as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
else:
try:
response = urlopen(source)
is_stream = True
# Retrieve filename from URL if possible
filename = re.findall("\w+\.\w+", source.split('/')[-1])
filename = filename[0] if filename else ""
inputfile = logfileparser.openlogfile(filename, object=response.read())
except (ValueError, URLError) as error:
if not kwargs.get('quiet', False):
(errno, strerror) = error.args
return None
elif hasattr(source, "read"):
inputfile = source
is_stream = True
# Streams are tricky since they don't have seek methods or seek won't work
# by design even if it is present. We solve this now by reading in the
# entire stream and using a StringIO buffer for parsing. This might be
# problematic for very large streams. Slow streams might also be an issue if
# the parsing is not instantaneous, but we'll deal with such edge cases
# as they arise. Ideally, in the future we'll create a class dedicated to
# dealing with these issues, supporting both files and streams.
if is_stream:
try:
inputfile.seek(0, 0)
except (AttributeError, IOError):
contents = inputfile.read()
try:
inputfile = io.StringIO(contents)
except:
inputfile = io.StringIO(unicode(contents))
inputfile.seek(0, 0)
# Proceed to return an instance of the logfile parser only if the filetype
# could be guessed. Need to make sure the input file is closed before creating
# an instance, because parsers will handle opening/closing on their own.
filetype = guess_filetype(inputfile)
# If the input file isn't a standard compchem log file, try one of
# the readers, falling back to Open Babel.
if not filetype:
if kwargs.get("cjson"):
filetype = readerclasses['cjson']
elif source and not is_stream:
ext = os.path.splitext(source)[1][1:].lower()
for extension in readerclasses:
if ext == extension:
filetype = readerclasses[extension]
# Proceed to return an instance of the logfile parser only if the filetype
# could be guessed. Need to make sure the input file is closed before creating
# an instance, because parsers will handle opening/closing on their own.
if filetype:
# We're going to clase and reopen below anyway, so this is just to avoid
# the missing seek method for fileinput.FileInput. In the long run
# we need to refactor to support for various input types in a more
# centralized fashion.
if is_listofstrings:
pass
else:
inputfile.seek(0, 0)
if not is_stream:
if is_listofstrings:
if filetype == Turbomole:
source = sort_turbomole_outputs(source)
inputfile.close()
return filetype(source, *args, **kwargs)
return filetype(inputfile, *args, **kwargs)