def translateString():
translateFile = open("translate.txt", "a+")
try:
for row_index in range(1, read_table.nrows):
error_reason_ch = read_table.cell(row_index, index_reason_ch).value
error_reason_en = translate(error_reason_ch)
translateFile.write("%s->%s\n" %
(error_reason_ch, error_reason_en))
finally:
translateFile.close()
def run(self):
with open('.\\config\\settin.json') as file:
data = json.load(file)
if not self.mode:
try:
result = translate(self.window, data)
self._signal.emit(result)
except Exception:
print_exc()
else:
data["sign"] += 1
with open('.\\config\\settin.json', 'w') as file:
json.dump(data, file)
try:
if data["sign"] % 2 == 0:
self.window.StartButton.setIcon(
qtawesome.icon('fa.pause', color='white'))
while True:
with open('.\\config\\settin.json') as file:
data = json.load(file)
if data["sign"] % 2 == 0:
try:
result = translate(self.window, data)
self._signal.emit(result)
sec = data["translateSpeed"] - 0.9
time.sleep(sec)
except Exception:
print_exc()
break
else:
self.window.StartButton.setIcon(
qtawesome.icon('fa.play', color='white'))
break
except Exception:
print_exc()
def translateData(src_ch, des_en, title=None):
"""
翻译错误原因, 写回表格
"""
try:
if title is not None:
print("Costume title:", title)
write_table.write(0, des_en, title)
else:
write_table.write(0, des_en, "翻译")
for row_index in range(1, read_table.nrows):
error_reason_ch = read_table.cell(row_index, src_ch).value
error_reason_en = translate(error_reason_ch)
write_table.write(row_index, des_en, error_reason_en)
finally:
write_data.save(fileName)
print("翻译 写入数据成功！")
def run_evaluation(model, source_vocab, target_vocabs, device, beam_size, filenames, ref_files, max_length):
'''
This method builds a model from scratch or using the encoder of a pre-trained model
model: the model being evaluated
source_vocabs: the source vocabulary for each file
target_vocabs: the target vocabulary for each file
beam_size: beam size during the translating
filenames: filenames of triples to process
ref_files: filenames with gold-standards for each process
max_length: max length of a sentence
'''
accuracies = []
for index, eval_name in enumerate(filenames):
eval_ref = ref_files[index]
eval_ref, corpus = '/'.join(eval_ref.split('/')[:-1]), eval_ref.split('/')[-1]
references = []
for i, fname in enumerate(sorted(os.listdir(eval_ref))):
if corpus in fname:
path = os.path.join(eval_ref, fname)
with open(path) as f:
doc = f.read().split('\n')
if i == 0:
references = [[w] for w in doc]
else:
for i, ref in enumerate(doc):
references[i].append(ref)
n = len(eval_name.split("/"))
name = eval_name.split("/")[n-1]
print(f'Reading {eval_name}')
with open(eval_name, "r") as f:
outputs = translate(model, index, f, source_vocab, target_vocabs[index], device,
beam_size=beam_size, max_length=max_length)
acc = 0.0
for j, output in enumerate(outputs):
if output.replace("<eos>","").strip().lower() in [w.lower() for w in references[j]]:
acc += 1
acc /= len(outputs)
accuracies.append(acc)
return accuracies
def parse(file_name, target_name):
fp = open(file_name, 'rb')
praser = PDFParser(fp)
doc = PDFDocument()
praser.set_document(doc)
doc.set_parser(praser)
doc.initialize()
if not doc.is_extractable:
raise PDFTextExtractionNotAllowed
else:
rsrcmgr = PDFResourceManager()
laparams = LAParams()
device = PDFPageAggregator(rsrcmgr, laparams=laparams)
interpreter = PDFPageInterpreter(rsrcmgr, device)
page_number = 1
for page in doc.get_pages():
print('page: ' + str(page_number))
interpreter.process_page(page)
layout = device.get_result()
# 这里layout是一个LTPage对象，里面存放着这个page解析出的各种对象
# 一般包括LTTextBox, LTFigure, LTImage, LTTextBoxHorizontal 等等
# 想要获取文本就获得对象的text属性
for x in layout:
if (isinstance(x, LTTextBoxHorizontal)):
with open(target_name, 'a') as f:
results = x.get_text()
translate_text = translate(results)
f.write(translate_text + '\n')
# if (isinstance(x, LTImage)):
# with open('patternColoring.txt', 'a') as f:
# results = x.get_image()
# f.write('###########\n' + results + '\n')
page_number += 1
def run_translate(model, source_vocab, target_vocabs, save_dir, device, beam_size, filenames, max_length):
'''
This method builds a model from scratch or using the encoder of a pre-trained model
model: the model being evaluated
source_vocabs: the source vocabulary for each file
target_vocabs: the target vocabulary for each file
save_dir: path where the outpus will be saved
beam_size: beam size during the translating
filenames: filenames of triples to process
max_length: max length of a sentence
'''
for index, eval_name in enumerate(filenames):
n = len(eval_name.split("/"))
name = eval_name.split("/")[n-1]
print(f'Reading {eval_name}')
fout = open(save_dir + name + "." + str(index) + ".out", "w")
with open(eval_name, "r") as f:
outputs = translate(model, index, f, source_vocab, target_vocabs[index], device,
beam_size=beam_size, max_length=max_length)
for output in outputs:
fout.write(output.replace("<eos>","").strip() + "\n")
fout.close()
def single_file():
file = "FastaFiles/MT019529.txt"
genome = make_genome_from_txt(file)
sequence = genome['sequence'] #29899
# https://www.ncbi.nlm.nih.gov/nuccore/MT019529
# 5'UTR
# https://www.ncbi.nlm.nih.gov/nuccore/MT019529.1?from=1&to=265
# 1..265
#
_5UTR = sequence[:265]
# Its link is: the same + ".1?location=266:13468,13468:21555"
# Basically encach section has a "gene" and a "CDS" section
# gene: gives us the overall start finish and gene name
# LINK: source, gene, translation, and genome
# CDS: this gives us all the real information that we need. it is basically
# the gene section but with more info. This does have a protein id link
# in it that you can use.
# LINK: source, gene, translation, genome
# PROTEINIDLINK: source, gene, translation
#
# KEY TAKEAWAY
# the links themselves aren't that helpful bc everything in the links are
# already shown in the main link's info
# We get: start, stop, gene name, product name, translation, protein id, etc
# join(266..13468,13468..21555)
# gene = orf1ab
# ribosomal_slippage What does this mean
# note= pp1ab; translated by -1 ribosomal frameshift Correlation?
# product = orf1ab polyprotein look what that is
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293612
# translation = ...
orf1ab = sequence[265:21555] # -> gets translated to the translation
# TODO: probably gonna wanna get the translation at some point
# GAP OF STUFF FROM 21555 -> 21563 (13)
# 21563..25384
# gene = s
# note = structural protein
# product = surface glycoprotein wtf is that
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293613
# translation = ...
s = sequence[21562:25384]
# GAP 25384 -> 25393 (9)
# 25393..26220
# gene = orf3a
# product = orf3a protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293614
# translation = ...
orf3a = sequence[25392:26220]
# Gap 26220 -> 26245 (25)
# 26245..26472
# gene = e
# product = envelope protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293615
# translation = ..
e = sequence[26244:26472]
# Gap 26472 -> 26523 (51)
# 26523..27191
# gene = m
# note = structual protein
# product = membrane glycoprotein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293616
# translation = ...
m = sequence[26522:27191]
# Gap 27191 -> 27202 (11)
# 27202..27387
# gene = orf6
# product = orf 6 protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293617
# translation = ...
orf6 = sequence[27201:27387]
# Gap 27387 -> 27394 (7)
# 27394..27759
# gene = orf7a
# product = orf7a protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293618
# translation = ....
orf7a = sequence[27393:27759]
# Gap 27759 -> 27894 (135)
# 27894..28259
# gene = orf8
# product = orf8 protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293619
# translation = ...
orf8 = sequence[27893:28259]
# Gap 28259 -> 28274 (15)
# 28274..29533
# gene = N
# note = structual protein
# product = nucleocapsid phosphoprotein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293620
# translation = ...
n = sequence[28273:29533]
# Gap 29533 -> 29558 (25)
# 29558..29674
# gene = orf10
# product = orf10 protein
# protein id = https://www.ncbi.nlm.nih.gov/protein/1805293621
# translation = ...
orf10 = sequence[29557:29674]
# 3'UTR
# 29675..29899
_3_utr = sequence[29674:29899]
print(translate(orf7a))