def get_segmented_regions(app_context, base_dir):
try:
files = get_files(app_context.application_context)
output = []
for index, file in enumerate(files):
file = get_json(base_dir, file['file']['name'])
file_properties = File(file)
pages = file_properties.get_pages()
page_counts = len(pages)
start_time = time.time()
for page_index in range(page_counts):
print('processing for page : ', page_index)
page_lines = file_properties.get_lines(page_index)
page_regions = file_properties.get_regions(page_index)
page_words = file_properties.get_words(page_index)
#page_regions = region_unifier.region_unifier(page_lines,page_regions)
file_properties.set_regions(
page_index,
segment_regions(page_words, page_lines, page_regions))
output.append(file_properties.get_file())
output[index]['status'] = {'message': "block-segmenter successful"}
end_time = time.time()
extraction_time = (end_time - start_time) / page_counts
log_info(
'block segmentation per page completed in {}'.format(
extraction_time), app_context.application_context)
app_context.application_context["outputs"] = output
log_info("successfully completed block segmentation", None)
except Exception as e:
log_exception("Error occured during block segmentation ",
app_context.application_context, e)
return None
return app_context.application_context
def get_layout(app_context):
try:
files = get_files(app_context.application_context)
#files = get_json(app_context.application_context)
#files = get_files(json_data)
file_images = []
output = []
for index, file_new in enumerate(files):
file = get_json(file_new['file']['name'])[0]
file_properties = File(file)
page_paths = file_properties.get_pages()
start_time = time.time()
for idx, page_path in enumerate(page_paths):
page_lines = file_properties.get_lines(idx)
page_words = file_properties.get_words(idx)
line_coords = get_coord(page_lines)
#page_path = '/'.join(page_path.split('/')[-4:])
page_path = 'upload/' + page_path.split('upload/')[-1]
#masked_image, table_and_lines = extract_table_line_regions(page_path)
#cell_regions = cell_layout(table_and_lines,page_path)
if torch.cuda.is_available():
torch.cuda.device(0)
print("*******cuda available")
torch.cuda.empty_cache()
time.sleep(1)
regions = primalaynet.predict_primanet(page_path, line_coords)
#regions += cell_regions
file['pages'][idx]["regions"] = regions
file['file'] = file_new['file']
file['config'] = file_new['config']
output.append(file)
output[index]['status'] = {}
output[index]['status']['message'] = "layout-detector successful"
end_time = time.time()
extraction_time = (end_time - start_time) / len(page_paths)
log_info(
'Layout detection per page completed in {}'.format(
extraction_time), app_context.application_context)
app_context.application_context["outputs"] = output
log_info("successfully completed layout detection", None)
except Exception as e:
log_exception("Error occured during prima layout detection ",
app_context.application_context, e)
return None
return app_context.application_context
def get_layout(app_context) :
try:
files = get_files(app_context.application_context)
file_images = []
output = []
for index,file in enumerate(files):
file_properties = File(file)
page_paths = file_properties.get_pages()
for idx,page_path in enumerate(page_paths):
page_lines = file_properties.get_lines(idx)
page_words = file_properties.get_words(idx)
line_coords = get_coord(page_lines)
regions = primalaynet.predict_primanet(page_path, line_coords)
file['pages'][idx]["regions"]=regions
output.append(file)
output[index]['status']['message']="layout-detector successful"
app_context.application_context["outputs"] =output
log_info("successfully completed layout detection", None)
except Exception as e:
log_exception("Error occured during prima layout detection ", app_context.application_context, e)
return None
return app_context.application_context