def run(self):
if not self.job.image_path:
stderr = "No image path could be determined"
return self.results(stdout=None, stderr=stderr, exitcode=1)
if not os.path.isfile(self.job.image_path):
stderr = "Could not find page image %s" % self.job.image_path
return self.results(stdout=None, stderr=stderr, exitcode=1)
# Create output parent directory if it doesn't exist
if not os.path.isdir(self.output_parent_dir):
mkdirs_exists_ok(self.output_parent_dir)
cmd = ["tesseract", self.job.image_path, self.output_filename, "-l", self.job.font.name, self.cfg]
proc = exec_cmd(cmd)
if proc.exitcode != 0:
return self.results(stdout=proc.stdout, stderr=proc.stderr, exitcode=proc.exitcode)
# Rename hOCR file to XML
if os.path.isfile(self.job.hocr_file) and not os.path.isfile(self.job.xml_file):
logger.debug("Renaming %s to %s" % (self.job.hocr_file, self.job.xml_file))
os.rename(self.job.hocr_file, self.job.xml_file)
self.job.page_result.ocr_text_path = self.job.txt_file
self.job.page_result.ocr_xml_path = self.job.xml_file
return self.results(stdout=None, stderr=None, exitcode=0)
def run(self):
self.generate_input_doc_list()
if not self.input_font_path:
stderr = "No input font path could be determined"
return self.results(stdout=None, stderr=stderr, exitcode=1)
if not os.path.isfile(self.input_font_path):
stderr = "Could not find input font path %s" % self.input_font_path
return self.results(stdout=None, stderr=stderr, exitcode=1)
# Create output parent directory if it doesn't exist
if not os.path.isdir(self.output_path):
mkdirs_exists_ok(self.output_path)
cmd = [
"java", self.java_max_heap,
"-Done-jar.main.class=edu.berkeley.cs.nlp.ocular.main.Transcribe",
"-jar", self.jar,
"-outputPath", self.output_path,
"-inputDocListPath", self.input_doc_list_path,
"-inputFontPath", self.input_font_path,
"-inputLmPath", self.input_lm_path,
"-inputGsmPath", self.input_gsm_path,
"-allowGlyphSubstitution", "true",
"-skipAlreadyTranscribedDocs", 'true',
"-emissionEngine", self.job.settings.ocular_emission_engine,
]
if self.extra_command_parameters:
cmd = cmd + self.extra_command_parameters
proc = exec_cmd(cmd, realtime=True)
if proc.exitcode != 0:
#logger.info("OcularTranscribe STDOUT: %s", proc.stdout)
return self.results(stdout=proc.stdout, stderr=proc.stderr, exitcode=proc.exitcode)
# Loop over each of this job's pages and build transcribed output paths
# These paths are added as results if the file is found
for j in self.job.jobs:
_image_basename = os.path.basename(j.image_path)
_image_name = os.path.splitext(_image_basename)[0]
_txt_name = "%s%s" % (_image_name, self.ocr_text_suffix)
_alto_name = "%s.alto.xml" % _image_name
_txt_path = os.path.join(self.transcribed_output_path, _txt_name)
_alto_path = os.path.join(self.transcribed_output_path, _alto_name)
if os.path.isfile(_txt_path):
j.page_result.ocr_text_path = _txt_path
if os.path.isfile(_alto_path):
j.page_result.ocr_xml_path = _alto_path
# Add extra transfers
if os.path.isdir(self.transcription_dir):
self.job.extra_transfers.append(self.transcription_dir)
return self.results(stdout=None, stderr=None, exitcode=0)
def run(self):
self.generate_input_doc_list()
if not self.input_font_path:
stderr = "No input font path could be determined"
return self.results(stdout=None, stderr=stderr, exitcode=1)
if not os.path.isfile(self.input_font_path):
stderr = "Could not find input font path %s" % self.input_font_path
return self.results(stdout=None, stderr=stderr, exitcode=1)
# Create output parent directory if it doesn't exist
if not os.path.isdir(self.output_path):
mkdirs_exists_ok(self.output_path)
cmd = [
"java", self.java_max_heap,
"-Done-jar.main.class=edu.berkeley.cs.nlp.ocular.main.TrainFont",
"-jar", self.jar,
"-outputPath", self.output_path,
"-inputDocListPath", self.input_doc_list_path,
"-inputFontPath", self.input_font_path,
"-inputLmPath", self.input_lm_path,
"-inputGsmPath", self.input_gsm_path,
#"-numDocs", str((len(self.images))),
"-outputFontPath", self.output_font_path,
"-outputLmPath", self.output_lm_path,
"-outputGsmPath", self.output_gsm_path,
"-continueFromLastCompleteIteration", "true",
"-allowGlyphSubstitution", "true",
"-updateLM", "true",
"-updateGsm", "true",
"-emissionEngine", self.job.settings.ocular_emission_engine,
]
if self.extra_command_parameters:
cmd = cmd + self.extra_command_parameters
proc = exec_cmd(cmd, realtime=True)
if proc.exitcode != 0:
#logger.info("OcularFontTraining STDOUT: %s", proc.stdout)
return self.results(stdout=proc.stdout, stderr=proc.stderr, exitcode=proc.exitcode)
# Only set font_training_result on one page (job) since this is a per-work result
if os.path.isfile(self.output_font_path):
self.job.font_training_result.font_path = self.output_font_path
if os.path.isfile(self.output_lm_path):
self.job.font_training_result.language_model_path = self.output_lm_path
if os.path.isfile(self.output_gsm_path):
self.job.font_training_result.glyph_substitution_model_path = self.output_gsm_path
return self.results(stdout=None, stderr=None, exitcode=0)
def save(self, data, dirname, filename, overwrite=False):
if not os.path.isdir(dirname):
logger.debug("Creating payload directory %s" % dirname)
mkdirs_exists_ok(dirname)
if not overwrite and os.path.exists(filename):
logger.error("payload file %s already exists" % filename)
return None
if overwrite:
logger.debug("Overwriting payload file at %s" % filename)
else:
logger.debug("Saving payload to %s" % filename)
with open(filename, 'w') as outfile:
json.dump(data, outfile)
return True
