def ocr_box_file(self, box_fname):
# Set up the names of output files
replace = lambda s: box_fname.replace('.box', s)
asis_fname = replace('.ml.txt')
nogram_out_fname = replace('.nogram.txt')
ngram_out_fname = replace('.gram.txt')
log_fname = replace('.{}.log'.format(self.loglevelname))
logging.basicConfig(filename=log_fname,
level=self.loglevel,
filemode="w")
# Read Bantries & get Most likely output
bf = BantryFile(box_fname)
with open(asis_fname, 'w', encoding='utf-8') as f:
f.write(post_process(bf.text))
# Process using ngrams
ngrammed_lines, notgrammed_lines = [], []
for linenum in range(bf.num_lines):
print("Line ", linenum)
line_bantries = bf.get_line_bantires(linenum)
gramgraph = GramGraph(line_bantries)
gramgraph.process_tree()
notgrammed_lines.append(gramgraph.get_best_apriori_str())
ngrammed_lines.append(gramgraph.get_best_str())
nogram_out = post_process("\n".join(notgrammed_lines))
with open(nogram_out_fname, 'w', encoding='utf-8') as out_file:
out_file.write(nogram_out)
ngram_out = post_process("\n".join(ngrammed_lines))
with open(ngram_out_fname, 'w', encoding='utf-8') as out_file:
out_file.write(ngram_out)
print("Input : ", box_fname)
print("As is output : ", asis_fname)
print("Without ngram : ", nogram_out_fname)
print("With ngram : ", ngram_out_fname)
print("Log : ", log_fname)
def ocr_box_file(self, box_fname):
# Set up the names of output files
replace = lambda s: box_fname.replace('.box', s)
asis_fname = replace('.ml.txt')
nogram_out_fname = replace('.nogram.txt')
ngram_out_fname = replace('.gram.txt')
log_fname = replace('.{}.log'.format(self.loglevelname))
logging.basicConfig(filename=log_fname,
level=self.loglevel,
filemode="w")
# Read Bantries & get Most likely output
bf = BantryFile(box_fname)
with open(asis_fname, 'w', encoding='utf-8') as f:
f.write(post_process(bf.text))
# Process using ngrams
ngrammed_lines, notgrammed_lines = [], []
for linenum in range(bf.num_lines):
print("Line ", linenum)
line_bantries = bf.get_line_bantires(linenum)
gramgraph = GramGraph(line_bantries)
gramgraph.process_tree()
notgrammed_lines.append(gramgraph.get_best_apriori_str())
ngrammed_lines.append(gramgraph.get_best_str())
nogram_out = post_process("\n".join(notgrammed_lines))
with open(nogram_out_fname, 'w', encoding='utf-8') as out_file:
out_file.write(nogram_out)
ngram_out = post_process("\n".join(ngrammed_lines))
with open(ngram_out_fname, 'w', encoding='utf-8') as out_file:
out_file.write(ngram_out)
print("Input : ", box_fname)
print("As is output : ", asis_fname)
print("Without ngram : ", nogram_out_fname)
print("With ngram : ", ngram_out_fname)
print("Log : ", log_fname)
if __name__ == "__main__":
import sys
from scaler import ScalerFactory
from bantry import Bantry, BantryFile
from classifier import Classifier
from ngram import Ngram
nnet_file = sys.argv[1] if len(sys.argv) > 1 else "library/nn.pkl"
banti_file_name = sys.argv[2] if len(sys.argv) > 2 else "sample_images/praasa.box"
scaler_prms_file = sys.argv[3] if len(sys.argv) > 3 else "scalings/relative48.scl"
labellings_file = sys.argv[4] if len(sys.argv) > 4 else "labellings/alphacodes.lbl"
ngram_file = "library/mega.123.pkl"
Bantry.scaler = ScalerFactory(scaler_prms_file)
Bantry.classifier = Classifier(nnet_file, labellings_file, logbase=1)
bf = BantryFile(banti_file_name)
ngram = Ngram(ngram_file)
GramGraph.set_ngram(ngram)
for linenum in range(bf.num_lines):
print('*' * 80)
bantires = bf.get_line_bantires(linenum)
gramgraph = GramGraph(bantires)
gramgraph.process_tree()
gramgraph.find_top_ngram_paths()
for node, children in enumerate(gramgraph.lchildren):
print(gramgraph.top_pathnodes_at(node, 1))
print(gramgraph.get_best_str('|'))
print(gramgraph.get_best_apriori_str('|'))
log_fname = replace('.{}.log'.format(logging._levelToName[loglevel]).lower())
asis_fname = replace('.ml.txt')
nogram_out_fname = replace('.nogram.txt')
ngram_out_fname = replace('.gram.txt')
logging.basicConfig(filename=log_fname, level=loglevel, filemode="w")
############################## Set-up scaler, classifier, ngram etc.
Bantry.scaler = ScalerFactory(scaler_fname)
Bantry.classifier = Classifier(nnet_fname, labels_fname, logbase=1)
ng = Ngram(ngram_fname)
Bantry.ngram = ng
GramGraph.set_ngram(ng)
############################## Read Bantries & get Most likely output
bf = BantryFile(banti_fname)
with open(asis_fname, 'w', encoding='utf-8') as f:
f.write(post_process(bf.text))
############################## Process using ngrams
ngrammed_lines, notgrammed_lines = [], []
for linenum in range(bf.num_lines):
print("Line ", linenum)
line_bantries = bf.get_line_bantires(linenum)
gramgraph = GramGraph(line_bantries)
gramgraph.process_tree()
notgrammed_lines.append(gramgraph.get_best_apriori_str())
ngrammed_lines.append(gramgraph.get_best_str())
nogram_out = post_process("\n".join(notgrammed_lines))
with open(scaler_prms_file, 'r') as sfp:
scaler_prms = ast.literal_eval(sfp.read())
with open(nnet_prms_file_name, 'rb') as nnet_prms_file:
nnet_prms = pickle.load(nnet_prms_file)
with open(labelings_file_name, encoding='utf-8') as labels_fp:
labellings = ast.literal_eval(labels_fp.read())
# print(labellings)
chars = LabelToUnicodeConverter(labellings).onecode
############################################# Init Network
Bantry.scaler = ScalerFactory(scaler_prms)
bf = BantryFile(banti_file_name)
nnet_prms['training_params']['BATCH_SZ'] = 1
ntwk = NeuralNet(**nnet_prms)
tester = ntwk.get_data_test_model(go_nuts=True)
############################################# Image saver
dir_name = os.path.basename(nnet_prms_file_name)[:-7] + '/'
if not os.path.exists(dir_name):
os.makedirs(dir_name)
namer = (dir_name + '{:03d}_{}_{:02d}.png').format
print("Look for me in :", dir_name)
def saver(outs, ch, debug=True):
saver.index += 1
for i, out in enumerate(outs):
from bantry import Bantry, BantryFile
from classifier import Classifier
from ngram import Ngram
nnet_file = sys.argv[1] if len(sys.argv) > 1 else "library/nn.pkl"
banti_file_name = sys.argv[2] if len(
sys.argv) > 2 else "sample_images/praasa.box"
scaler_prms_file = sys.argv[3] if len(
sys.argv) > 3 else "scalings/relative48.scl"
labellings_file = sys.argv[4] if len(
sys.argv) > 4 else "labellings/alphacodes.lbl"
ngram_file = "library/mega.123.pkl"
Bantry.scaler = ScalerFactory(scaler_prms_file)
Bantry.classifier = Classifier(nnet_file, labellings_file, logbase=1)
bf = BantryFile(banti_file_name)
ngram = Ngram(ngram_file)
GramGraph.set_ngram(ngram)
for linenum in range(bf.num_lines):
print('*' * 80)
bantires = bf.get_line_bantires(linenum)
gramgraph = GramGraph(bantires)
gramgraph.process_tree()
gramgraph.find_top_ngram_paths()
for node, children in enumerate(gramgraph.lchildren):
print(gramgraph.top_pathnodes_at(node, 1))
print(gramgraph.get_best_str('|'))
print(gramgraph.get_best_apriori_str('|'))
