Commit latest
This commit is contained in:
parent
e1e13bbe4d
commit
f956dbd6da
2 changed files with 95 additions and 50 deletions
|
|
@ -41,10 +41,10 @@ LEAVE=False
|
|||
|
||||
# Flags
|
||||
CODE401 = True
|
||||
CODE405 = False
|
||||
CODE102 = True
|
||||
CODE405 = True
|
||||
CODE102 = False
|
||||
CODE122 = False
|
||||
CODE101 = True
|
||||
CODE101 = False
|
||||
CODE355655 = False
|
||||
CODE357 = False
|
||||
CODE657 = False
|
||||
|
|
@ -221,7 +221,7 @@ def translateNote(event, regex):
|
|||
jaString = re.sub(r'\n', ' ', oldJAString)
|
||||
|
||||
# Translate
|
||||
response = translateGPT(jaString, 'Reply with the English translation of the NPC name.', True)
|
||||
response = translateGPT(jaString, 'Reply with the English translation of the description.', True)
|
||||
translatedText = response[0]
|
||||
|
||||
# Textwrap
|
||||
|
|
@ -477,7 +477,9 @@ def searchCodes(page, pbar):
|
|||
tokens = 0
|
||||
speaker = ''
|
||||
speakerVar = ''
|
||||
nametag = ''
|
||||
match = []
|
||||
syncIndext = 0
|
||||
global LOCK
|
||||
global NAMESLIST
|
||||
|
||||
|
|
@ -501,10 +503,6 @@ def searchCodes(page, pbar):
|
|||
# Grab String
|
||||
jaString = codeList[i]['parameters'][0]
|
||||
|
||||
# # Catch Speaker
|
||||
# if jaString.startswith(' '):
|
||||
# jaString = jaString + ': '
|
||||
|
||||
# Using this to keep track of 401's in a row. Throws IndexError at EndOfList (Expected Behavior)
|
||||
currentGroup.append(jaString)
|
||||
|
||||
|
|
@ -523,24 +521,69 @@ def searchCodes(page, pbar):
|
|||
|
||||
# Check for speaker
|
||||
if '\\N' in finalJAString or '\\n' in finalJAString:
|
||||
match = re.findall(r'([\\]+[Nn]<(.+)>)', finalJAString)
|
||||
match = re.findall(r'(.+[Nn]<(.+)>)', finalJAString)
|
||||
if len(match) != 0:
|
||||
response = translateGPT(match[0][1], 'Reply with only the english translation of the NPC name', True)
|
||||
tokens += response[1]
|
||||
speaker = response[0].strip('.')
|
||||
nametag = match[0][0].replace(match[0][1], speaker)
|
||||
finalJAString = finalJAString.replace(match[0][0], '')
|
||||
|
||||
# Put names in list
|
||||
if NAMES == True and speaker not in NAMESLIST:
|
||||
with LOCK:
|
||||
NAMESLIST.append(speaker)
|
||||
elif '\\kw' in finalJAString:
|
||||
match = re.findall(r'\\+kw\[[0-9]+\]', finalJAString)
|
||||
if len(match) != 0:
|
||||
if '1' in match[0]:
|
||||
speaker = 'Ayako Nagatsuki'
|
||||
if '2' in match[0]:
|
||||
speaker = 'Rei'
|
||||
|
||||
# Set name var to top of list
|
||||
codeList[j]['parameters'][0] = match[0]
|
||||
codeList[j]['code'] = 401
|
||||
|
||||
# Set next item as dialogue
|
||||
j += 1
|
||||
codeList[j]['parameters'][0] = match[0]
|
||||
codeList[j]['code'] = 401
|
||||
|
||||
# Remove from final string
|
||||
finalJAString = finalJAString.replace(match[0], '')
|
||||
elif '【' in finalJAString:
|
||||
if '\\z' in finalJAString:
|
||||
match = re.findall(r'(.+?【(.+?)】.+?\\z\[[0-9]+\])', finalJAString)
|
||||
else:
|
||||
match = re.findall(r'(.+?【(.+?)】)', finalJAString)
|
||||
|
||||
if len(match) != 0:
|
||||
response = translateGPT(match[0][1], 'Reply with only the english translation of the NPC name', True)
|
||||
tokens += response[1]
|
||||
speaker = response[0].strip('.')
|
||||
nametag = match[0][0].replace(match[0][1], speaker)
|
||||
finalJAString = finalJAString.replace(match[0][0], '')
|
||||
|
||||
# Set name var to top of list
|
||||
codeList[j]['parameters'][0] = nametag
|
||||
codeList[j]['code'] = 401
|
||||
|
||||
# Set next item as dialogue
|
||||
j += 1
|
||||
codeList[j]['parameters'][0] = finalJAString
|
||||
codeList[j]['code'] = 401
|
||||
|
||||
# Put names in list
|
||||
if NAMES == True and speaker not in NAMESLIST:
|
||||
with LOCK:
|
||||
NAMESLIST.append(speaker)
|
||||
|
||||
# Need to remove outside code and put it back later
|
||||
startString = re.search(r'^[^一-龠ぁ-ゔァ-ヴー【】（）[]<>「」『』a-zA-Z0-9Ａ-Ｚ０-９\\]+', finalJAString)
|
||||
finalJAString = re.sub(r'^[^一-龠ぁ-ゔァ-ヴー【】（）[]<>「」『』a-zA-Z0-9Ａ-Ｚ０-９\\]+', '', finalJAString)
|
||||
if startString is None: startString = ''
|
||||
else: startString = startString.group()
|
||||
else: startString = startString.group()
|
||||
|
||||
# Remove any textwrap
|
||||
finalJAString = re.sub(r'\n', ' ', finalJAString)
|
||||
|
|
@ -565,17 +608,18 @@ def searchCodes(page, pbar):
|
|||
translatedText = response[0]
|
||||
textHistory.append('\"' + translatedText + '\"')
|
||||
|
||||
# Remove added speaker
|
||||
# Remove added speaker and add nametag
|
||||
translatedText = re.sub(r'^.+?:\s', '', translatedText)
|
||||
translatedText = nametag + translatedText
|
||||
speaker = ''
|
||||
speakerVar = ''
|
||||
nametag = ''
|
||||
else:
|
||||
translatedText = finalJAString
|
||||
|
||||
# Textwrap
|
||||
translatedText = textwrap.fill(translatedText, width=WIDTH)
|
||||
|
||||
# Resub start and end
|
||||
# Start String
|
||||
translatedText = startString + translatedText
|
||||
|
||||
# Set Data
|
||||
|
|
@ -601,7 +645,7 @@ def searchCodes(page, pbar):
|
|||
if codeList[i]['code'] == 122 and CODE122 == True:
|
||||
# This is going to be the var being set. (IMPORTANT)
|
||||
varNum = codeList[i]['parameters'][0]
|
||||
if varNum != 1238:
|
||||
if varNum != 53:
|
||||
continue
|
||||
|
||||
jaString = codeList[i]['parameters'][4]
|
||||
|
|
@ -620,7 +664,7 @@ def searchCodes(page, pbar):
|
|||
matchList = re.findall(r"\'(.*?)\'", jaString)
|
||||
|
||||
for match in matchList:
|
||||
response = translateGPT(match, 'Sex Positions', True)
|
||||
response = translateGPT(match, '', True)
|
||||
translatedText = response[0]
|
||||
tokens += response[1]
|
||||
|
||||
|
|
@ -850,7 +894,7 @@ def searchCodes(page, pbar):
|
|||
continue
|
||||
|
||||
# Want to translate this script
|
||||
if 'ask:' not in jaString:
|
||||
if 'event_text :' not in jaString:
|
||||
continue
|
||||
|
||||
# Need to remove outside code and put it back later
|
||||
|
|
@ -864,7 +908,7 @@ def searchCodes(page, pbar):
|
|||
else: endString = endString.group()
|
||||
|
||||
# Translate
|
||||
response = translateGPT(jaString, '', True)
|
||||
response = translateGPT(jaString, 'Reply with the English translation of the Event Title', True)
|
||||
tokens += response[1]
|
||||
translatedText = response[0]
|
||||
|
||||
|
|
@ -1045,11 +1089,11 @@ def searchCodes(page, pbar):
|
|||
|
||||
except IndexError as e:
|
||||
# This is part of the logic so we just pass it
|
||||
tracebackLineNo = str(traceback.extract_tb(sys.exc_info()[2])[-1].lineno)
|
||||
traceback.print_exc()
|
||||
# raise Exception(str(e) + '|Line:' + tracebackLineNo)
|
||||
except Exception as e:
|
||||
tracebackLineNo = str(traceback.extract_tb(sys.exc_info()[2])[-1].lineno)
|
||||
raise Exception(str(e) + '|Line:' + tracebackLineNo + '| Failed to translate: ' + oldjaString)
|
||||
traceback.print_exc()
|
||||
raise Exception(str(e) + 'Failed to translate: ' + oldjaString)
|
||||
|
||||
# Append leftover groups in 401
|
||||
if len(currentGroup) > 0:
|
||||
|
|
@ -1099,8 +1143,8 @@ def searchSS(state, pbar):
|
|||
message4Response = translateGPT(state['message4'], 'reply with only the english translation of the text.', True) if 'message4' in state else ''
|
||||
|
||||
# if 'note' in state:
|
||||
if 'DOBBY' in state['note']:
|
||||
tokens += translateNote(state, r'<DOBBY_SKILLINFO>\n(.+)\n</DOBBY_SKILLINFO>')
|
||||
if 'Extended Description' in state['note']:
|
||||
tokens += translateNote(state, r'<Extended Description:([^>]*)>')
|
||||
|
||||
# Count Tokens
|
||||
tokens += nameResponse[1] if nameResponse != '' else 0
|
||||
|
|
@ -1203,7 +1247,7 @@ def subVars(jaString):
|
|||
varList = re.findall(varRegex, jaString)
|
||||
if len(varList) != 0:
|
||||
for var in varList:
|
||||
jaString = jaString.replace(var, '<x' + str(count) + '>')
|
||||
jaString = jaString.replace(var, '{x' + str(count) + '}')
|
||||
count += 1
|
||||
|
||||
return [jaString, varList]
|
||||
|
|
@ -1213,12 +1257,12 @@ def resubVars(translatedText, varList):
|
|||
|
||||
if len(varList) != 0:
|
||||
for var in varList:
|
||||
translatedText = translatedText.replace('<x' + str(count) + '>', var)
|
||||
translatedText = translatedText.replace('{x' + str(count) + '}', var)
|
||||
count += 1
|
||||
|
||||
# Remove Color Variables Spaces
|
||||
if '\\c' in translatedText:
|
||||
translatedText = re.sub(r'\s*(\\+c\[[1-9]+\])\s*', r'\1', translatedText)
|
||||
translatedText = re.sub(r'\s*(\\+c\[[1-9]+\])\s*', r' \1', translatedText)
|
||||
translatedText = re.sub(r'\s*(\\+c\[0+\])', r'\1', translatedText)
|
||||
return translatedText
|
||||
|
||||
|
|
@ -1241,7 +1285,7 @@ def translateGPT(t, history, fullPromptFlag):
|
|||
return(t, 0)
|
||||
|
||||
"""Translate text using GPT"""
|
||||
context = 'Eroge Names Context: カレン == Karen | Female, エリス == Eris | Female, コレット == Colette | Female, テオ == Teo | Male, メイヴィス == Mavis | Female, '
|
||||
context = 'Eroge Names Context: レイン == Rain | Female, フィルス == Phils | Female, メイル == Meryl | Female, ジャス == Jazz | Male'
|
||||
if fullPromptFlag:
|
||||
system = PROMPT
|
||||
user = 'Line to Translate: ' + subbedT
|
||||
|
|
|
|||
|
|
@ -25,7 +25,7 @@ APICOST = .002 # Depends on the model https://openai.com/pricing
|
|||
PROMPT = Path('prompt.txt').read_text(encoding='utf-8')
|
||||
THREADS = 20
|
||||
LOCK = threading.Lock()
|
||||
WIDTH = 72
|
||||
WIDTH = 75
|
||||
LISTWIDTH = 75
|
||||
MAXHISTORY = 10
|
||||
ESTIMATE = ''
|
||||
|
|
@ -122,6 +122,7 @@ def parseText(data, filename):
|
|||
try:
|
||||
response = translateText(linesList, pbar)
|
||||
except Exception as e:
|
||||
traceback.print_exc()
|
||||
return [linesList, 0, e]
|
||||
return [response[0], response[1], None]
|
||||
|
||||
|
|
@ -138,7 +139,7 @@ def translateText(data, pbar):
|
|||
if i != syncIndex:
|
||||
continue
|
||||
|
||||
match = re.findall(r'm\[[0-9]+\] = \"(.+?)\"', data[i])
|
||||
match = re.findall(r'm\[[0-9]+\] = \"(.*)\"', data[i])
|
||||
if len(match) > 0:
|
||||
jaString = match[0]
|
||||
|
||||
|
|
@ -148,19 +149,25 @@ def translateText(data, pbar):
|
|||
|
||||
# Grab Speaker
|
||||
speakerMatch = re.findall(r's\[[0-9]+\] = \"(.+?)[／\"]', data[i-1])
|
||||
if len(match) != 0:
|
||||
response = translateGPT(speakerMatch[0], 'Reply with only the english translation of the NPC name', True)
|
||||
tokens += response[1]
|
||||
speaker = response[0].strip('.')
|
||||
if len(speakerMatch) > 0:
|
||||
# If there isn't any Japanese in the text just skip
|
||||
if re.search(r'[一-龠]+|[ぁ-ゔ]+|[ァ-ヴー]+', jaString) and '_' not in speakerMatch[0]:
|
||||
speaker = ''
|
||||
else:
|
||||
speaker = ''
|
||||
else:
|
||||
speaker = ''
|
||||
|
||||
# Grab rest of the messages
|
||||
currentGroup.append(jaString)
|
||||
start = i
|
||||
while (re.search(r'm\[[0-9]+\] = \"(.+?)\"', data[i+1]) != None):
|
||||
data[i] = re.sub(r'(m\[[0-9]+\]) = \"(.+)\"', rf'\1 = ""', data[i])
|
||||
while (len(data) > i+1 and re.search(r'm\[[0-9]+\] = \"(.*)\"', data[i+1]) != None):
|
||||
i+=1
|
||||
match = re.findall(r'm\[[0-9]+\] = \"(.+?)\"', data[i])
|
||||
match = re.findall(r'm\[[0-9]+\] = \"(.*)\"', data[i])
|
||||
currentGroup.append(match[0])
|
||||
finalJAString = ''.join(currentGroup)
|
||||
data[i] = re.sub(r'(m\[[0-9]+\]) = \"(.+)\"', rf'\1 = ""', data[i])
|
||||
finalJAString = ' '.join(currentGroup)
|
||||
|
||||
# Translate
|
||||
if speaker != '':
|
||||
|
|
@ -170,7 +177,7 @@ def translateText(data, pbar):
|
|||
tokens += response[1]
|
||||
translatedText = response[0]
|
||||
|
||||
# Remove added speaker
|
||||
# Remove added speaker and quotes
|
||||
translatedText = re.sub(r'^.+?:\s', '', translatedText)
|
||||
|
||||
# TextHistory is what we use to give GPT Context, so thats appended here.
|
||||
|
|
@ -186,22 +193,15 @@ def translateText(data, pbar):
|
|||
currentGroup = []
|
||||
|
||||
# Textwrap
|
||||
translatedText = translatedText.replace('\"', '\\"')
|
||||
translatedText = textwrap.fill(translatedText, width=WIDTH)
|
||||
translatedText = translatedText.replace('\n','\\n')
|
||||
speaker = ''
|
||||
|
||||
# Setup Speaker
|
||||
if speakerFlag == True:
|
||||
# Remove characters that may break scripts
|
||||
charList = ['.', '\"']
|
||||
for char in charList:
|
||||
translatedText = translatedText.replace(char, '')
|
||||
|
||||
speaker = translatedText
|
||||
speakerFlag = False
|
||||
|
||||
# Write
|
||||
data[i] = translatedText
|
||||
textList = translatedText.split("\n")
|
||||
for t in textList:
|
||||
data[start] = re.sub(r'(m\[[0-9]+\]) = \"(.*)\"', rf'\1 = "{t}"', data[start])
|
||||
start+=1
|
||||
|
||||
syncIndex = i + 1
|
||||
pbar.update()
|
||||
return [data, tokens]
|
||||
|
|
@ -247,11 +247,12 @@ def translateGPT(t, history, fullPromptFlag):
|
|||
subbedT = varResponse[0]
|
||||
|
||||
# If there isn't any Japanese in the text just skip
|
||||
return (t,0)
|
||||
if not re.search(r'[一-龠]+|[ぁ-ゔ]+|[ァ-ヴー]+', subbedT):
|
||||
return(t, 0)
|
||||
|
||||
"""Translate text using GPT"""
|
||||
context = 'Eroge Names Context: カレン == Karen | Female, エリス == Eris | Female, コレット == Colette | Female, テオ == Teo | Male, メイヴィス == Mavis | Female, '
|
||||
context = 'Eroge Names Context: 夏樹 明人 == Natsuki Akito | Male, 朝露 砂夜子 == Asatsuyu Sayoko | Female, 神野 藍 == Jinno Ai | Female, 神野 菫 | Jinno Sumire | Female, 夏樹 海夕里 == Natsuki Miyuri | Female, 水森 陽太 == Mizumori Youta | Male, 夏樹 和人 == Natsuki Kazuto | Male, 野崎 博也 == Nozaki Hiroya | Male, 大菊 ジュン == Oogiku Jun | Male, 酒井 絹代 == Sakai Kinuyo | Female, 酒井 豊 == Sakai Yutaka | Male, 竜生 春義 == Tatsuki Haruyoshi | Male, 竜生 潤 == Tatsuki Jun | Male, 吉沢 英玄 == Yoshizawa Eigen | Male, 吉沢 武雄 == Yoshizawa Takeo | Male'
|
||||
if fullPromptFlag:
|
||||
system = PROMPT
|
||||
user = 'Line to Translate: ' + subbedT
|
||||
|
|
|
|||
Loading…
Reference in a new issue