def clean_job_for_quality(batch_job_dict, fallback=True):
"""
Removes jobs from the batch description that are not suitable for the master
video's quality. It applies only for encoding jobs!
@param batch_job_dict: the dict containing the batch description
@type batch_job_dict: dict
@param
@return: the cleaned dict
@rtype: dict
"""
survived_jobs = []
fallback_jobs = []
other_jobs = []
for job in batch_job_dict["jobs"]:
if job["mode"] == "encode":
if getval(job, "fallback") and fallback:
fallback_jobs.append(job)
if getval(job, "enforce"):
survived_jobs.append(job)
else:
profile = None
if getval(job, "profile"):
profile = get_encoding_profile(job["profile"])
if assure_quality(
input_file=batch_job_dict["input"],
aspect=chose2("aspect", job, profile),
target_width=chose2("width", job, profile),
target_height=chose2("height", job, profile),
target_bitrate=chose2("videobitrate", job, profile),
):
survived_jobs.append(job)
else:
other_jobs.append(job)
if survived_jobs:
survived_jobs.extend(other_jobs)
new_jobs = survived_jobs
else:
fallback_jobs.extend(other_jobs)
new_jobs = fallback_jobs
pprint(locals())
batch_job_dict["jobs"] = new_jobs
return batch_job_dict
def clean_job_for_quality(batch_job_dict, fallback=True):
"""
Removes jobs from the batch description that are not suitable for the master
video's quality. It applies only for encoding jobs!
@param batch_job_dict: the dict containing the batch description
@type batch_job_dict: dict
@param
@return: the cleaned dict
@rtype: dict
"""
survived_jobs = []
fallback_jobs = []
other_jobs = []
for job in batch_job_dict['jobs']:
if job['mode'] == 'encode':
if getval(job, 'fallback') and fallback:
fallback_jobs.append(job)
if getval(job, 'enforce'):
survived_jobs.append(job)
else:
profile = None
if getval(job, 'profile'):
profile = get_encoding_profile(job['profile'])
if assure_quality(input_file=batch_job_dict['input'],
aspect=chose2('aspect', job, profile),
target_width=chose2('width', job, profile),
target_height=chose2('height', job, profile),
target_bitrate=chose2(
'videobitrate', job, profile)):
survived_jobs.append(job)
else:
other_jobs.append(job)
if survived_jobs:
survived_jobs.extend(other_jobs)
new_jobs = survived_jobs
else:
fallback_jobs.extend(other_jobs)
new_jobs = fallback_jobs
pprint(locals())
batch_job_dict['jobs'] = new_jobs
return batch_job_dict
def clean_job_for_quality(batch_job_dict, fallback=True):
"""
Removes jobs from the batch description that are not suitable for the master
video's quality. It applies only for encoding jobs!
@param batch_job_dict: the dict containing the batch description
@type batch_job_dict: dict
@param
@return: the cleaned dict
@rtype: dict
"""
survived_jobs = []
fallback_jobs = []
other_jobs = []
for job in batch_job_dict['jobs']:
if job['mode'] == 'encode':
if getval(job, 'fallback') and fallback:
fallback_jobs.append(job)
if getval(job, 'enforce'):
survived_jobs.append(job)
else:
profile = None
if getval(job, 'profile'):
profile = get_encoding_profile(job['profile'])
if assure_quality(input_file=batch_job_dict['input'],
aspect=chose2('aspect', job, profile),
target_width=chose2('width', job, profile),
target_height=chose2('height', job, profile),
target_bitrate=chose2('videobitrate', job, profile)):
survived_jobs.append(job)
else:
other_jobs.append(job)
if survived_jobs:
survived_jobs.extend(other_jobs)
new_jobs = survived_jobs
else:
fallback_jobs.extend(other_jobs)
new_jobs = fallback_jobs
pprint(locals())
batch_job_dict['jobs'] = new_jobs
return batch_job_dict
def encode_video(input_file, output_file,
acodec=None, vcodec=None,
abitrate=None, vbitrate=None,
resolution=None,
passes=1,
special=None, specialfirst=None, specialsecond=None,
metadata=None,
width=None, height=None, aspect=None,
profile=None,
update_fnc=task_update_progress,
message_fnc=write_message
):
""" Starts an ffmpeg encoding process based on the given parameters.
The encoding is run as a subprocess. The progress of the subprocess is
continiously written to the given messaging functions. In a normale case,
these should be the ones of BibTask.
@param input_file: Path to the input video.
@type input_file: string
@param output_file: Path to the output file. If no other parameters are giv
than input and output files, FFmpeg tries to auto-discover the right codecs
for the given file extension. In this case, every other aspect like
resolution and bitrates will be the same as in the input video.
@type output_file: string
@param acodec: The audio codec to use. This must be an available codec of
libavcodec within FFmpeg.
@type acodec: string
@param vcodec: The video codec to use. This must be an available codec of
libavcodec within FFmpeg.
@type vcodec: string
@param abitrate: Bitrate of the audio stream. In bit/s.
@type abitrate: int
@param vbitrate: Bitrate of the video stream. In bit/s.
@type vbitrate: int
@param resolution: Fixed size of the frames in the transcoded video.
FFmpeg notation: 'WxH' or preset like 'vga'. See also 'width'
@param passes: Number of encoding passes. Either 1 or 2.
@type passes: int
@param special: Additional FFmpeg parameters.
@type special: string
@param specialfirst: Additional FFmpeg parameters for the first pass.
The 'special' parameter is ignored if this ist not 'None'
@type specialfirst: string
@param specialsecond: Additional FFmpeg parameters for the second pass.
The 'special' parameter is ignored if this is not 'None'
@type specialsecond: string
@param metadata: Metadata that should be added to the transcoded video.
This must be a dictionary. As with as metadata in FFmpeg, there is no
guarantee that the metadata specified in the dictionary will really be added
to the file, because it will largelydepend on the container format and its
supported fields.
@type metadata: dict
@param width: Instead of giving a fixed resolution, you can use width and
height as dimensional constrains. The algorithm will try to preserve the
original aspect and fit the new frame size into the given dimensions.
@type width: int
@param height: see 'width'
@type height: int
@param aspect: A float representing the aspect ratio of the video:
4:3 equals 1.33 and 16:9 equals 1.77.
This is a fallback in case the algorithm fails to determine the real aspect
ratio from the video. See also 'width'
@type aspect: float or "4:3" like string
@param profile: A profile to use. The priority is on the parameters
directly given to the function.
@type profile: string
@param update_fnc: A function called to display or log an the encoding
status. This function must accept a string.
@type update_fnc: function
@param message_fnc: A function to log important messages or errors.
This function must accept a string.
@type message_fnc: function
@return: True if the encoding was successful, False if not
@rtype: boolean
"""
def encode():
""" Subfunction to run the acutal encoding
"""
## Start process
process = subprocess.Popen(command,
stderr=log_file_handle,
close_fds=True)
## While the process is running
time.sleep(1)
while process.poll() is None:
# Update the status in bibsched
update_status()
time.sleep(4)
## If the process was terminated
if process.poll() == -15:
# Encoding was terminated by system
message_fnc("FFMPEG was terminated")
update_fnc(" FFMPEG was terminated")
return 0
## If there was an error during encoding
if process.poll() == 1:
update_fnc(" An FFMPEG error has appeared, see log")
message_fnc("An FFMPEG error has appeared encoding %s" % output_file)
message_fnc("Command was: %s" % ' '.join(command))
message_fnc("Last lines of the FFmpeg log:")
## open the logfile again an retrieve the size
log_file_handle2 = open(log_file_name, 'rb')
size = os.fstat(log_file_handle2.fileno())[6]
## Read the last lines
log_file_handle2.seek(-min(size, 10000), 2)
lastlines = log_file_handle2.read().splitlines()[-5:]
for line in lastlines:
message_fnc(line)
return 0
## If everything went fine
if process.poll() == 0:
message_fnc("Encoding of %s done" % output_file)
update_fnc("Encoding of %s done" % output_file)
return 1
def build_command(nofpass=1):
""" Builds the ffmpeg command according to the function params
"""
def insert(key, value):
""" Shortcut for inserting parameters into the arg list
"""
base_args.insert(-1, key)
base_args.insert(-1, value)
## Determine base command arguments from the pass to run
base_args = None
if passes == 1:
base_args = [CFG_PATH_FFMPEG, '-y', '-i', input_file, output_file]
elif passes == 2:
if nofpass == 1:
base_args = [CFG_PATH_FFMPEG, '-y', '-i', input_file,
'-pass', '1', '-passlogfile', pass_log_file,
'-an', '-f', 'rawvideo', '/dev/null']
elif nofpass == 2:
base_args = [CFG_PATH_FFMPEG, '-y', '-i', input_file,
'-pass', '2', '-passlogfile',
pass_log_file, output_file]
## Insert additional arguments
if acodec is not None:
insert('-acodec', acodec)
if vcodec is not None:
insert('-vcodec', vcodec)
if abitrate is not None:
insert('-b:a', str(abitrate))
if vbitrate is not None:
insert('-b:v', str(vbitrate))
## If a resolution is given
if resolution:
insert('-s', resolution)
## If not, you can give width and height and generate the resolution
else:
## Use our new function to get the size of the input
nresolution = determine_resolution_preserving_aspect(input_file,
width,
height,
aspect)
insert('-s', nresolution)
## Metadata additions
if type(metadata) is type(dict()):
## build metadata arguments for ffmpeg
for key, value in metadata.iteritems():
if value is not None:
meta_arg = (
CFG_BIBENCODE_FFMPEG_METADATA_ARGUMENT % (key, value)
)
insert("-metadata", meta_arg)
## Special argument additions
if passes == 1:
if passes == 1 and special is not None:
for val in special.split():
base_args.insert(-1, val)
elif passes == 2:
if nofpass == 1:
if specialfirst is not None:
for val in specialfirst.split():
base_args.insert(-1, val)
if nofpass == 2:
if specialsecond is not None:
for val in specialsecond.split():
base_args.insert(-1, val)
return base_args
def update_status():
""" Parses the encoding status and updates the task in bibsched
"""
def graphical(value):
""" Converts a percentage value to a nice graphical representation
"""
## If the given value is a valid precentage
if value >= 0 and value <= 100:
## This is to get nice, aligned output in bibsched
oval = str(value).zfill(3)
return (
"[" + "#"*(value/10) + " "*(10-(value/10)) +
"][%d/%d] %s%%" % (nofpass, passes, oval)
)
else:
## Sometimes the parsed values from FFMPEG are totaly off.
## Or maybe nneeded values are not avail. for the given video.
## In this case there is no estimate.
return "[ no est. ][%d/%d] " % (nofpass, passes)
## init variables
time_string = '0.0'
percentage_done = -1
## try to read the encoding log
try:
filehandle = open(log_file_name, 'rb')
except IOError:
message_fnc("Error opening %s" % log_file_name)
update_fnc("Could not open encoding log")
return
## Check the size of the file before reading from the end
size = os.path.getsize(log_file_name)
if not size:
return
## Go to the end of the log
filehandle.seek(-min(10000, size), 2)
chunk = filehandle.read()
lines = chunk.splitlines()
## try to parse the status
for line in reversed(lines):
if CFG_BIBENCODE_FFMPEG_ENCODE_TIME.match(line):
time_string = (
CFG_BIBENCODE_FFMPEG_ENCODE_TIME.match(line).groups()
)[0]
break
filehandle.close()
try:
percentage_done = int(timecode_to_seconds(time_string) / total_seconds * 100)
except:
precentage_done = -1
## Now update the bibsched progress
opath, ofile = os.path.split(output_file)
if len(opath) > 8:
opath = "..." + opath[-8:]
ohint = opath + '/' + ofile
update_fnc(graphical(percentage_done) + " > " + ohint)
#------------------#
# PROFILE HANDLING #
#------------------#
if profile:
profile = get_encoding_profile(profile)
acodec = chose(acodec, 'audiocodec', profile)
vcodec = chose(vcodec, 'videocodec', profile)
abitrate = chose(abitrate, 'audiobitrate', profile)
vbitrate = chose(vbitrate, 'videobitrate', profile)
resolution = chose(resolution, 'resolution', profile)
passes = getval(profile, 'passes', 1)
special = chose(special, 'special', profile)
specialfirst = chose(specialfirst, 'special_firstpass', profile)
specialsecond = chose(specialsecond, 'special_secondpass', profile)
metadata = chose(metadata, 'metadata', profile)
width = chose(width, 'width', profile)
height = chose(height, 'height', profile)
aspect = chose(aspect, 'aspect', profile)
#----------#
# ENCODING #
#----------#
## Mark Task as stoppable
# task_sleep_now_if_required()
tech_metadata = ffprobe_metadata(input_file)
try:
total_seconds = float(tech_metadata['format']['duration'])
except:
total_seconds = 0.0
## Run the encoding
pass_log_file = CFG_BIBENCODE_FFMPEG_PASSLOGFILE_PREFIX % (
os.path.splitext(os.path.split(input_file)[1])[0],
str(uuid.uuid4()))
no_error = True
## For every encoding pass to do
for apass in range(0, passes):
nofpass = apass + 1
if no_error:
## Create Logfiles
log_file_name = _filename_log(output_file, nofpass)
try:
log_file_handle = open(log_file_name, 'w')
except IOError:
message_fnc("Error creating %s" % log_file_name)
update_fnc("Error creating logfile")
return 0
## Build command for FFMPEG
command = build_command(nofpass)
## Start encoding, result will define to continue or not to
no_error = encode()
## !!! Status Update
return no_error
def process_batch_job(batch_job_file):
""" Processes a batch job description dictionary
@param batch_job_file: a fullpath to a batch job file
@type batch_job_file: string
@return: 1 if the process was successfull, 0 if not
@rtype; int
"""
def upload_marcxml_file(marcxml):
""" Creates a temporary marcxml file and sends it to bibupload
"""
xml_filename = 'bibencode_'+ str(batch_job['recid']) + '_' + str(uuid.uuid4()) + '.xml'
xml_filename = os.path.join(invenio.config.CFG_TMPSHAREDDIR, xml_filename)
xml_file = file(xml_filename, 'w')
xml_file.write(marcxml)
xml_file.close()
targs = ['-c', xml_filename]
task_low_level_submission('bibupload', 'bibencode', *targs)
#---------#
# GENERAL #
#---------#
_task_write_message("----------- Handling Master -----------")
## Check the validity of the batch file here
batch_job = json_decode_file(batch_job_file)
## Sanitise batch description and raise errrors
batch_job = sanitise_batch_job(batch_job)
## Check if the record exists
if record_exists(batch_job['recid']) < 1:
raise Exception("Record not found")
recdoc = BibRecDocs(batch_job['recid'])
#--------------------#
# UPDATE FROM MASTER #
#--------------------#
## We want to add new stuff to the video's record, using the master as input
if getval(batch_job, 'update_from_master'):
found_master = False
bibdocs = recdoc.list_bibdocs()
for bibdoc in bibdocs:
bibdocfiles = bibdoc.list_all_files()
for bibdocfile in bibdocfiles:
comment = bibdocfile.get_comment()
description = bibdocfile.get_description()
subformat = bibdocfile.get_subformat()
m_comment = getval(batch_job, 'bibdoc_master_comment', comment)
m_description = getval(batch_job, 'bibdoc_master_description', description)
m_subformat = getval(batch_job, 'bibdoc_master_subformat', subformat)
if (comment == m_comment and
description == m_description and
subformat == m_subformat):
found_master = True
batch_job['input'] = bibdocfile.get_full_path()
## Get the aspect of the from the record
try:
## Assumes pbcore metadata mapping
batch_job['aspect'] = get_fieldvalues(124, CFG_BIBENCODE_ASPECT_RATIO_MARC_FIELD)[0]
except IndexError:
pass
break
if found_master:
break
if not found_master:
_task_write_message("Video master for record %d not found"
% batch_job['recid'])
task_update_progress("Video master for record %d not found"
% batch_job['recid'])
## Maybe send an email?
return 1
## Clean the job to do no upscaling etc
if getval(batch_job, 'assure_quality'):
batch_job = clean_job_for_quality(batch_job)
global _BATCH_STEPS
_BATCH_STEPS = len(batch_job['jobs'])
## Generate the docname from the input filename's name or given name
bibdoc_video_docname, bibdoc_video_extension = decompose_file(batch_job['input'])[1:]
if not bibdoc_video_extension or getval(batch_job, 'bibdoc_master_extension'):
bibdoc_video_extension = getval(batch_job, 'bibdoc_master_extension')
if getval(batch_job, 'bibdoc_master_docname'):
bibdoc_video_docname = getval(batch_job, 'bibdoc_master_docname')
write_message("Creating BibDoc for %s" % bibdoc_video_docname)
## If the bibdoc exists, receive it
if bibdoc_video_docname in recdoc.get_bibdoc_names():
bibdoc_video = recdoc.get_bibdoc(bibdoc_video_docname)
## Create a new bibdoc if it does not exist
else:
bibdoc_video = recdoc.add_bibdoc(docname=bibdoc_video_docname)
## Get the directory auf the newly created bibdoc to copy stuff there
bibdoc_video_directory = bibdoc_video.get_base_dir()
#--------#
# MASTER #
#--------#
if not getval(batch_job, 'update_from_master'):
if getval(batch_job, 'add_master'):
## Generate the right name for the master
## The master should be hidden first an then renamed
## when it is really available
## !!! FIX !!!
_task_write_message("Adding %s master to the BibDoc"
% bibdoc_video_docname)
master_format = compose_format(
bibdoc_video_extension,
getval(batch_job, 'bibdoc_master_subformat', 'master')
)
## If a file of the same format is there, something is wrong, remove it!
## it might be caused by a previous corrupted submission etc.
if bibdoc_video.format_already_exists_p(master_format):
bibdoc_video.delete_file(master_format, 1)
bibdoc_video.add_file_new_format(
batch_job['input'],
version=1,
description=getval(batch_job, 'bibdoc_master_description'),
comment=getval(batch_job, 'bibdoc_master_comment'),
docformat=master_format
)
#-----------#
# JOBS LOOP #
#-----------#
return_code = 1
global _BATCH_STEP
for job in batch_job['jobs']:
_task_write_message("----------- Job %s of %s -----------"
% (_BATCH_STEP, _BATCH_STEPS))
## Try to substitute docname with master docname
if getval(job, 'bibdoc_docname'):
job['bibdoc_docname'] = Template(job['bibdoc_docname']).safe_substitute({'bibdoc_master_docname': bibdoc_video_docname})
#-------------#
# TRANSCODING #
#-------------#
if job['mode'] == 'encode':
## Skip the job if assure_quality is not set and marked as fallback
if not getval(batch_job, 'assure_quality') and getval(job, 'fallback'):
continue
if getval(job, 'profile'):
profile = get_encoding_profile(job['profile'])
else:
profile = None
## We need an extension defined fot the video container
bibdoc_video_extension = getval(job, 'extension',
getval(profile, 'extension'))
if not bibdoc_video_extension:
raise Exception("No container/extension defined")
## Get the docname and subformat
bibdoc_video_subformat = getval(job, 'bibdoc_subformat')
bibdoc_slave_video_docname = getval(job, 'bibdoc_docname', bibdoc_video_docname)
## The subformat is incompatible with ffmpegs name convention
## We do the encoding without and rename it afterwards
bibdoc_video_fullpath = compose_file(
bibdoc_video_directory,
bibdoc_slave_video_docname,
bibdoc_video_extension
)
_task_write_message("Transcoding %s to %s;%s" % (bibdoc_slave_video_docname,
bibdoc_video_extension,
bibdoc_video_subformat))
## We encode now directly into the bibdocs directory
encoding_result = encode_video(
input_file=batch_job['input'],
output_file=bibdoc_video_fullpath,
acodec=getval(job, 'audiocodec'),
vcodec=getval(job, 'videocodec'),
abitrate=getval(job, 'videobitrate'),
vbitrate=getval(job, 'audiobitrate'),
resolution=getval(job, 'resolution'),
passes=getval(job, 'passes', 1),
special=getval(job, 'special'),
specialfirst=getval(job, 'specialfirst'),
specialsecond=getval(job, 'specialsecond'),
metadata=getval(job, 'metadata'),
width=getval(job, 'width'),
height=getval(job, 'height'),
aspect=getval(batch_job, 'aspect'), # Aspect for every job
profile=getval(job, 'profile'),
update_fnc=_task_update_overall_status,
message_fnc=_task_write_message
)
return_code &= encoding_result
## only on success
if encoding_result:
## Rename it, adding the subformat
os.rename(bibdoc_video_fullpath,
compose_file(bibdoc_video_directory,
bibdoc_video_extension,
bibdoc_video_subformat,
1,
bibdoc_slave_video_docname)
)
#bibdoc_video._build_file_list()
bibdoc_video.touch()
bibdoc_video._sync_to_db()
bibdoc_video_format = compose_format(bibdoc_video_extension,
bibdoc_video_subformat)
if getval(job, 'bibdoc_comment'):
bibdoc_video.set_comment(getval(job, 'bibdoc_comment'),
bibdoc_video_format)
if getval(job, 'bibdoc_description'):
bibdoc_video.set_description(getval(job, 'bibdoc_description'),
bibdoc_video_format)
#------------#
# EXTRACTION #
#------------#
# if there are multiple extraction jobs, all the produced files
# with the same name will be in the same bibdoc! Make sure that
# you use different subformats or docname templates to avoid
# conflicts.
if job['mode'] == 'extract':
if getval(job, 'profile'):
profile = get_extract_profile(job['profile'])
else:
profile = {}
bibdoc_frame_subformat = getval(job, 'bibdoc_subformat')
_task_write_message("Extracting frames to temporary directory")
tmpdir = invenio.config.CFG_TMPDIR + "/" + str(uuid.uuid4())
os.mkdir(tmpdir)
#Move this to the batch description
bibdoc_frame_docname = getval(job, 'bibdoc_docname', bibdoc_video_docname)
tmpfname = (tmpdir + "/" + bibdoc_frame_docname + '.'
+ getval(profile, 'extension',
getval(job, 'extension', 'jpg')))
extraction_result = extract_frames(input_file=batch_job['input'],
output_file=tmpfname,
size=getval(job, 'size'),
positions=getval(job, 'positions'),
numberof=getval(job, 'numberof'),
width=getval(job, 'width'),
height=getval(job, 'height'),
aspect=getval(batch_job, 'aspect'),
profile=getval(job, 'profile'),
update_fnc=_task_update_overall_status,
)
return_code &= extraction_result
## only on success:
if extraction_result:
## for every filename in the directorys, create a bibdoc that contains
## all sizes of the frame from the two directories
files = os.listdir(tmpdir)
for filename in files:
## The docname was altered by BibEncode extract through substitution
## Retrieve it from the filename again
bibdoc_frame_docname, bibdoc_frame_extension = os.path.splitext(filename)
_task_write_message("Creating new bibdoc for %s" % bibdoc_frame_docname)
## If the bibdoc exists, receive it
if bibdoc_frame_docname in recdoc.get_bibdoc_names():
bibdoc_frame = recdoc.get_bibdoc(bibdoc_frame_docname)
## Create a new bibdoc if it does not exist
else:
bibdoc_frame = recdoc.add_bibdoc(docname=bibdoc_frame_docname)
## The filename including path from tmpdir
fname = os.path.join(tmpdir, filename)
bibdoc_frame_format = compose_format(bibdoc_frame_extension, bibdoc_frame_subformat)
## Same as with the master, if the format allready exists,
## override it, because something went wrong before
if bibdoc_frame.format_already_exists_p(bibdoc_frame_format):
bibdoc_frame.delete_file(bibdoc_frame_format, 1)
_task_write_message("Adding %s jpg;%s to BibDoc"
% (bibdoc_frame_docname,
getval(job, 'bibdoc_subformat')))
bibdoc_frame.add_file_new_format(
fname,
version=1,
description=getval(job, 'bibdoc_description'),
comment=getval(job, 'bibdoc_comment'),
docformat=bibdoc_frame_format)
## Remove the temporary folders
_task_write_message("Removing temporary directory")
shutil.rmtree(tmpdir)
_BATCH_STEP = _BATCH_STEP + 1
#-----------------#
# FIX BIBDOC/MARC #
#-----------------#
_task_write_message("----------- Handling MARCXML -----------")
## Fix the BibDoc for all the videos previously created
_task_write_message("Updating BibDoc of %s" % bibdoc_video_docname)
bibdoc_video._build_file_list()
## Fix the MARC
_task_write_message("Fixing MARC")
cli_fix_marc({}, [batch_job['recid']], False)
if getval(batch_job, 'collection'):
## Make the record visible by moving in from the collection
marcxml = ("<record><controlfield tag=\"001\">%d</controlfield>"
"<datafield tag=\"980\" ind1=\" \" ind2=\" \">"
"<subfield code=\"a\">%s</subfield></datafield></record>"
) % (batch_job['recid'], batch_job['collection'])
upload_marcxml_file(marcxml)
#---------------------#
# ADD MASTER METADATA #
#---------------------#
if getval(batch_job, 'add_master_metadata'):
_task_write_message("Adding master metadata")
pbcore = pbcore_metadata(input_file = getval(batch_job, 'input'),
pbcoreIdentifier = batch_job['recid'],
aspect_override = getval(batch_job, 'aspect'))
marcxml = format(pbcore, CFG_BIBENCODE_PBCORE_MARC_XSLT)
upload_marcxml_file(marcxml)
#------------------#
# ADD MARC SNIPPET #
#------------------#
if getval(batch_job, 'marc_snippet'):
marc_snippet = open(getval(batch_job, 'marc_snippet'))
marcxml = marc_snippet.read()
marc_snippet.close()
upload_marcxml_file(marcxml)
#--------------#
# DELETE INPUT #
#--------------#
if getval(batch_job, 'delete_input'):
_task_write_message("Deleting input file")
# only if successfull
if not return_code:
# only if input matches pattern
if getval(batch_job, 'delete_input_pattern', '') in getval(batch_job, 'input'):
try:
os.remove(getval(batch_job, 'input'))
except OSError:
pass
#--------------#
# NOTIFICATION #
#--------------#
## Send Notification emails on errors
if not return_code:
if getval(batch_job, 'notify_user'):
_notify_error_user(getval(batch_job, 'notify_user'),
getval(batch_job, 'submission_filename', batch_job['input']),
getval(batch_job, 'recid'),
getval(batch_job, 'submission_title', ""))
_task_write_message("Notify user because of an error")
if getval(batch_job, 'notify_admin'):
_task_write_message("Notify admin because of an error")
if type(getval(batch_job, 'notify_admin') == type(str()) ):
_notify_error_admin(batch_job,
getval(batch_job, 'notify_admin'))
else:
_notify_error_admin(batch_job)
else:
if getval(batch_job, 'notify_user'):
_task_write_message("Notify user because of success")
_notify_success_user(getval(batch_job, 'notify_user'),
getval(batch_job, 'submission_filename', batch_job['input']),
getval(batch_job, 'recid'),
getval(batch_job, 'submission_title', ""))
return 1
def process_batch_job(batch_job_file):
""" Processes a batch job description dictionary
@param batch_job_file: a fullpath to a batch job file
@type batch_job_file: string
@return: 1 if the process was successfull, 0 if not
@rtype; int
"""
def upload_marcxml_file(marcxml):
""" Creates a temporary marcxml file and sends it to bibupload
"""
xml_filename = 'bibencode_' + str(batch_job['recid']) + '_' + str(
uuid.uuid4()) + '.xml'
xml_filename = os.path.join(invenio.config.CFG_TMPSHAREDDIR,
xml_filename)
xml_file = file(xml_filename, 'w')
xml_file.write(marcxml)
xml_file.close()
targs = ['-c', xml_filename]
task_low_level_submission('bibupload', 'bibencode', *targs)
#---------#
# GENERAL #
#---------#
_task_write_message("----------- Handling Master -----------")
## Check the validity of the batch file here
batch_job = json_decode_file(batch_job_file)
## Sanitise batch description and raise errrors
batch_job = sanitise_batch_job(batch_job)
## Check if the record exists
if record_exists(batch_job['recid']) < 1:
raise Exception("Record not found")
recdoc = BibRecDocs(batch_job['recid'])
#--------------------#
# UPDATE FROM MASTER #
#--------------------#
## We want to add new stuff to the video's record, using the master as input
if getval(batch_job, 'update_from_master'):
found_master = False
bibdocs = recdoc.list_bibdocs()
for bibdoc in bibdocs:
bibdocfiles = bibdoc.list_all_files()
for bibdocfile in bibdocfiles:
comment = bibdocfile.get_comment()
description = bibdocfile.get_description()
subformat = bibdocfile.get_subformat()
m_comment = getval(batch_job, 'bibdoc_master_comment', comment)
m_description = getval(batch_job, 'bibdoc_master_description',
description)
m_subformat = getval(batch_job, 'bibdoc_master_subformat',
subformat)
if (comment == m_comment and description == m_description
and subformat == m_subformat):
found_master = True
batch_job['input'] = bibdocfile.get_full_path()
## Get the aspect of the from the record
try:
## Assumes pbcore metadata mapping
batch_job['aspect'] = get_fieldvalues(
124, CFG_BIBENCODE_ASPECT_RATIO_MARC_FIELD)[0]
except IndexError:
pass
break
if found_master:
break
if not found_master:
_task_write_message("Video master for record %d not found" %
batch_job['recid'])
task_update_progress("Video master for record %d not found" %
batch_job['recid'])
## Maybe send an email?
return 1
## Clean the job to do no upscaling etc
if getval(batch_job, 'assure_quality'):
batch_job = clean_job_for_quality(batch_job)
global _BATCH_STEPS
_BATCH_STEPS = len(batch_job['jobs'])
## Generate the docname from the input filename's name or given name
bibdoc_video_docname, bibdoc_video_extension = decompose_file(
batch_job['input'])[1:]
if not bibdoc_video_extension or getval(batch_job,
'bibdoc_master_extension'):
bibdoc_video_extension = getval(batch_job, 'bibdoc_master_extension')
if getval(batch_job, 'bibdoc_master_docname'):
bibdoc_video_docname = getval(batch_job, 'bibdoc_master_docname')
write_message("Creating BibDoc for %s" % bibdoc_video_docname)
## If the bibdoc exists, receive it
if bibdoc_video_docname in recdoc.get_bibdoc_names():
bibdoc_video = recdoc.get_bibdoc(bibdoc_video_docname)
## Create a new bibdoc if it does not exist
else:
bibdoc_video = recdoc.add_bibdoc(docname=bibdoc_video_docname)
## Get the directory auf the newly created bibdoc to copy stuff there
bibdoc_video_directory = bibdoc_video.get_base_dir()
#--------#
# MASTER #
#--------#
if not getval(batch_job, 'update_from_master'):
if getval(batch_job, 'add_master'):
## Generate the right name for the master
## The master should be hidden first an then renamed
## when it is really available
## !!! FIX !!!
_task_write_message("Adding %s master to the BibDoc" %
bibdoc_video_docname)
master_format = compose_format(
bibdoc_video_extension,
getval(batch_job, 'bibdoc_master_subformat', 'master'))
## If a file of the same format is there, something is wrong, remove it!
## it might be caused by a previous corrupted submission etc.
if bibdoc_video.format_already_exists_p(master_format):
bibdoc_video.delete_file(master_format, 1)
bibdoc_video.add_file_new_format(
batch_job['input'],
version=1,
description=getval(batch_job, 'bibdoc_master_description'),
comment=getval(batch_job, 'bibdoc_master_comment'),
docformat=master_format)
#-----------#
# JOBS LOOP #
#-----------#
return_code = 1
global _BATCH_STEP
for job in batch_job['jobs']:
_task_write_message("----------- Job %s of %s -----------" %
(_BATCH_STEP, _BATCH_STEPS))
## Try to substitute docname with master docname
if getval(job, 'bibdoc_docname'):
job['bibdoc_docname'] = Template(
job['bibdoc_docname']).safe_substitute(
{'bibdoc_master_docname': bibdoc_video_docname})
#-------------#
# TRANSCODING #
#-------------#
if job['mode'] == 'encode':
## Skip the job if assure_quality is not set and marked as fallback
if not getval(batch_job, 'assure_quality') and getval(
job, 'fallback'):
continue
if getval(job, 'profile'):
profile = get_encoding_profile(job['profile'])
else:
profile = None
## We need an extension defined fot the video container
bibdoc_video_extension = getval(job, 'extension',
getval(profile, 'extension'))
if not bibdoc_video_extension:
raise Exception("No container/extension defined")
## Get the docname and subformat
bibdoc_video_subformat = getval(job, 'bibdoc_subformat')
bibdoc_slave_video_docname = getval(job, 'bibdoc_docname',
bibdoc_video_docname)
## The subformat is incompatible with ffmpegs name convention
## We do the encoding without and rename it afterwards
bibdoc_video_fullpath = compose_file(bibdoc_video_directory,
bibdoc_slave_video_docname,
bibdoc_video_extension)
_task_write_message(
"Transcoding %s to %s;%s" %
(bibdoc_slave_video_docname, bibdoc_video_extension,
bibdoc_video_subformat))
## We encode now directly into the bibdocs directory
encoding_result = encode_video(
input_file=batch_job['input'],
output_file=bibdoc_video_fullpath,
acodec=getval(job, 'audiocodec'),
vcodec=getval(job, 'videocodec'),
abitrate=getval(job, 'videobitrate'),
vbitrate=getval(job, 'audiobitrate'),
resolution=getval(job, 'resolution'),
passes=getval(job, 'passes', 1),
special=getval(job, 'special'),
specialfirst=getval(job, 'specialfirst'),
specialsecond=getval(job, 'specialsecond'),
metadata=getval(job, 'metadata'),
width=getval(job, 'width'),
height=getval(job, 'height'),
aspect=getval(batch_job, 'aspect'), # Aspect for every job
profile=getval(job, 'profile'),
update_fnc=_task_update_overall_status,
message_fnc=_task_write_message)
return_code &= encoding_result
## only on success
if encoding_result:
## Rename it, adding the subformat
os.rename(
bibdoc_video_fullpath,
compose_file(bibdoc_video_directory,
bibdoc_video_extension,
bibdoc_video_subformat, 1,
bibdoc_slave_video_docname))
#bibdoc_video._build_file_list()
bibdoc_video.touch()
bibdoc_video._sync_to_db()
bibdoc_video_format = compose_format(bibdoc_video_extension,
bibdoc_video_subformat)
if getval(job, 'bibdoc_comment'):
bibdoc_video.set_comment(getval(job, 'bibdoc_comment'),
bibdoc_video_format)
if getval(job, 'bibdoc_description'):
bibdoc_video.set_description(
getval(job, 'bibdoc_description'), bibdoc_video_format)
#------------#
# EXTRACTION #
#------------#
# if there are multiple extraction jobs, all the produced files
# with the same name will be in the same bibdoc! Make sure that
# you use different subformats or docname templates to avoid
# conflicts.
if job['mode'] == 'extract':
if getval(job, 'profile'):
profile = get_extract_profile(job['profile'])
else:
profile = {}
bibdoc_frame_subformat = getval(job, 'bibdoc_subformat')
_task_write_message("Extracting frames to temporary directory")
tmpdir = invenio.config.CFG_TMPDIR + "/" + str(uuid.uuid4())
os.mkdir(tmpdir)
#Move this to the batch description
bibdoc_frame_docname = getval(job, 'bibdoc_docname',
bibdoc_video_docname)
tmpfname = (
tmpdir + "/" + bibdoc_frame_docname + '.' +
getval(profile, 'extension', getval(job, 'extension', 'jpg')))
extraction_result = extract_frames(
input_file=batch_job['input'],
output_file=tmpfname,
size=getval(job, 'size'),
positions=getval(job, 'positions'),
numberof=getval(job, 'numberof'),
width=getval(job, 'width'),
height=getval(job, 'height'),
aspect=getval(batch_job, 'aspect'),
profile=getval(job, 'profile'),
update_fnc=_task_update_overall_status,
)
return_code &= extraction_result
## only on success:
if extraction_result:
## for every filename in the directorys, create a bibdoc that contains
## all sizes of the frame from the two directories
files = os.listdir(tmpdir)
for filename in files:
## The docname was altered by BibEncode extract through substitution
## Retrieve it from the filename again
bibdoc_frame_docname, bibdoc_frame_extension = os.path.splitext(
filename)
_task_write_message("Creating new bibdoc for %s" %
bibdoc_frame_docname)
## If the bibdoc exists, receive it
if bibdoc_frame_docname in recdoc.get_bibdoc_names():
bibdoc_frame = recdoc.get_bibdoc(bibdoc_frame_docname)
## Create a new bibdoc if it does not exist
else:
bibdoc_frame = recdoc.add_bibdoc(
docname=bibdoc_frame_docname)
## The filename including path from tmpdir
fname = os.path.join(tmpdir, filename)
bibdoc_frame_format = compose_format(
bibdoc_frame_extension, bibdoc_frame_subformat)
## Same as with the master, if the format allready exists,
## override it, because something went wrong before
if bibdoc_frame.format_already_exists_p(
bibdoc_frame_format):
bibdoc_frame.delete_file(bibdoc_frame_format, 1)
_task_write_message("Adding %s jpg;%s to BibDoc" %
(bibdoc_frame_docname,
getval(job, 'bibdoc_subformat')))
bibdoc_frame.add_file_new_format(
fname,
version=1,
description=getval(job, 'bibdoc_description'),
comment=getval(job, 'bibdoc_comment'),
docformat=bibdoc_frame_format)
## Remove the temporary folders
_task_write_message("Removing temporary directory")
shutil.rmtree(tmpdir)
_BATCH_STEP = _BATCH_STEP + 1
#-----------------#
# FIX BIBDOC/MARC #
#-----------------#
_task_write_message("----------- Handling MARCXML -----------")
## Fix the BibDoc for all the videos previously created
_task_write_message("Updating BibDoc of %s" % bibdoc_video_docname)
bibdoc_video._build_file_list()
## Fix the MARC
_task_write_message("Fixing MARC")
cli_fix_marc({}, [batch_job['recid']], False)
if getval(batch_job, 'collection'):
## Make the record visible by moving in from the collection
marcxml = ("<record><controlfield tag=\"001\">%d</controlfield>"
"<datafield tag=\"980\" ind1=\" \" ind2=\" \">"
"<subfield code=\"a\">%s</subfield></datafield></record>"
) % (batch_job['recid'], batch_job['collection'])
upload_marcxml_file(marcxml)
#---------------------#
# ADD MASTER METADATA #
#---------------------#
if getval(batch_job, 'add_master_metadata'):
_task_write_message("Adding master metadata")
pbcore = pbcore_metadata(input_file=getval(batch_job, 'input'),
pbcoreIdentifier=batch_job['recid'],
aspect_override=getval(batch_job, 'aspect'))
marcxml = format(pbcore, CFG_BIBENCODE_PBCORE_MARC_XSLT)
upload_marcxml_file(marcxml)
#------------------#
# ADD MARC SNIPPET #
#------------------#
if getval(batch_job, 'marc_snippet'):
marc_snippet = open(getval(batch_job, 'marc_snippet'))
marcxml = marc_snippet.read()
marc_snippet.close()
upload_marcxml_file(marcxml)
#--------------#
# DELETE INPUT #
#--------------#
if getval(batch_job, 'delete_input'):
_task_write_message("Deleting input file")
# only if successfull
if not return_code:
# only if input matches pattern
if getval(batch_job, 'delete_input_pattern',
'') in getval(batch_job, 'input'):
try:
os.remove(getval(batch_job, 'input'))
except OSError:
pass
#--------------#
# NOTIFICATION #
#--------------#
## Send Notification emails on errors
if not return_code:
if getval(batch_job, 'notify_user'):
_notify_error_user(
getval(batch_job, 'notify_user'),
getval(batch_job, 'submission_filename', batch_job['input']),
getval(batch_job, 'recid'),
getval(batch_job, 'submission_title', ""))
_task_write_message("Notify user because of an error")
if getval(batch_job, 'notify_admin'):
_task_write_message("Notify admin because of an error")
if type(getval(batch_job, 'notify_admin') == type(str())):
_notify_error_admin(batch_job, getval(batch_job,
'notify_admin'))
else:
_notify_error_admin(batch_job)
else:
if getval(batch_job, 'notify_user'):
_task_write_message("Notify user because of success")
_notify_success_user(
getval(batch_job, 'notify_user'),
getval(batch_job, 'submission_filename', batch_job['input']),
getval(batch_job, 'recid'),
getval(batch_job, 'submission_title', ""))
return 1