Update Lune

This commit is contained in:
Dazed 2023-11-07 11:26:15 -06:00
parent ddee81460e
commit 02c3ec1a89
2 changed files with 117 additions and 163 deletions

View file

@ -1,4 +1,3 @@
from concurrent.futures import ThreadPoolExecutor, as_completed
import json
import os
from pathlib import Path
@ -21,36 +20,29 @@ load_dotenv()
openai.organization = os.getenv('org')
openai.api_key = os.getenv('key')
APICOST = .002 # Depends on the model https://openai.com/pricing
INPUTAPICOST = .002 # Depends on the model https://openai.com/pricing
OUTPUTAPICOST = .002
PROMPT = Path('prompt.txt').read_text(encoding='utf-8')
THREADS = 20
THREADS = 10 # Controls how many threads are working on a single file (May have to drop this)
LOCK = threading.Lock()
WIDTH = 60
LISTWIDTH = 75
WIDTH = 50
LISTWIDTH = 90
NOTEWIDTH = 50
MAXHISTORY = 10
ESTIMATE = ''
TOTALCOST = 0
TOKENS = 0
TOTALTOKENS = 0
totalTokens = [0, 0]
NAMESLIST = []
#tqdm Globals
BAR_FORMAT='{l_bar}{bar:10}{r_bar}{bar:-10b}'
POSITION=0
LEAVE=False
# Flags
CODE401 = True
CODE102 = True
CODE122 = False
CODE101 = False
CODE355655 = False
CODE357 = False
CODE356 = False
CODE320 = False
CODE111 = False
BRFLAG = False # If the game uses <br> instead
FIXTEXTWRAP = True
IGNORETLTEXT = False
def handleLune(filename, estimate):
global ESTIMATE, TOKENS, TOTALTOKENS, TOTALCOST
global ESTIMATE, totalTokens
ESTIMATE = estimate
if estimate:
@ -61,185 +53,139 @@ def handleLune(filename, estimate):
end = time.time()
tqdm.write(getResultString(translatedData, end - start, filename))
with LOCK:
TOTALCOST += translatedData[1] * .001 * APICOST
TOTALTOKENS += translatedData[1]
totalTokens[0] += translatedData[1][0]
totalTokens[1] += translatedData[1][1]
return getResultString(['', totalTokens, None], end - start, 'TOTAL')
else:
with open('translated/' + filename, 'w', encoding='shiftjis', newline='\n') as outFile:
start = time.time()
translatedData = openFiles(filename)
try:
with open('translated/' + filename, 'w', encoding='UTF-8') as outFile:
start = time.time()
translatedData = openFiles(filename)
# Print Result
end = time.time()
outFile.writelines(translatedData[0])
tqdm.write(getResultString(translatedData, end - start, filename))
with LOCK:
TOTALCOST += translatedData[1] * .001 * APICOST
TOTALTOKENS += translatedData[1]
# Print Result
end = time.time()
json.dump(translatedData[0], outFile, ensure_ascii=False)
tqdm.write(getResultString(translatedData, end - start, filename))
with LOCK:
totalTokens[0] += translatedData[1][0]
totalTokens[1] += translatedData[1][1]
except Exception as e:
return 'Fail'
return getResultString(['', TOTALTOKENS, None], end - start, 'TOTAL')
return getResultString(['', totalTokens, None], end - start, 'TOTAL')
def openFiles(filename):
with open('files/' + filename, 'r', encoding='shiftjis') as f:
translatedData = parseText(f, filename)
with open('files/' + filename, 'r', encoding='UTF-8-sig') as f:
data = json.load(f)
# Map Files
if '.json' in filename:
translatedData = parseJSON(data, filename)
else:
raise NameError(filename + ' Not Supported')
return translatedData
def getResultString(translatedData, translationTime, filename):
# File Print String
tokenString = Fore.YELLOW + '[' + str(translatedData[1]) + \
' Tokens/${:,.4f}'.format(translatedData[1] * .001 * APICOST) + ']'
totalTokenstring =\
Fore.YELLOW +\
'[Input: ' + str(translatedData[1][0]) + ']'\
'[Output: ' + str(translatedData[1][1]) + ']'\
'[Cost: ${:,.4f}'.format((translatedData[1][0] * .001 * INPUTAPICOST) +\
(translatedData[1][1] * .001 * OUTPUTAPICOST)) + ']'
timeString = Fore.BLUE + '[' + str(round(translationTime, 1)) + 's]'
if translatedData[2] == None:
# Success
return filename + ': ' + tokenString + timeString + Fore.GREEN + u' \u2713 ' + Fore.RESET
return filename + ': ' + totalTokenstring + timeString + Fore.GREEN + u' \u2713 ' + Fore.RESET
else:
# Fail
try:
raise translatedData[2]
except Exception as e:
traceback.print_exc()
errorString = str(e) + Fore.RED
return filename + ': ' + tokenString + timeString + Fore.RED + u' \u2717 ' +\
return filename + ': ' + totalTokenstring + timeString + Fore.RED + u' \u2717 ' +\
errorString + Fore.RESET
def parseText(data, filename):
totalTokens = 0
def parseJSON(data, filename):
totalTokens = [0, 0]
totalLines = 0
totalLines = len(data)
global LOCK
# Get total for progress bar
linesList = data.readlines()
totalLines = len(linesList)
with tqdm(bar_format=BAR_FORMAT, position=POSITION, total=totalLines, leave=LEAVE) as pbar:
pbar.desc=filename
pbar.total=totalLines
try:
response = translateText(linesList, pbar)
result = translateJSON(data, pbar)
totalTokens[0] += result[0]
totalTokens[1] += result[1]
except Exception as e:
traceback.print_exc()
return [linesList, 0, e]
return [response[0], response[1], None]
return [data, totalTokens, e]
return [data, totalTokens, None]
def translateText(data, pbar):
def translateJSON(data, pbar):
textHistory = []
maxHistory = MAXHISTORY
tokens = 0
speaker = ''
speakerFlag = False
syncIndex = 0
tokens = [0, 0]
speaker = 'None'
### Translation
for i in range(len(data)):
if syncIndex > i:
i = syncIndex
for item in data:
# Speaker
if 'name' in item:
if item['name'] not in [None, '-']:
response = translateGPT(item['name'], 'Reply with only the english translation of the NPC name', False)
speaker = response[0]
tokens[0] += response[1][0]
tokens[1] += response[1][1]
item['name'] = speaker
else:
speaker = 'None'
# Finish if at end
if i+1 > len(data):
return [data, tokens]
# Text
if 'message' in item:
if item['message'] != None:
jaString = item['message']
# Remove newlines
jaString = data[i]
jaString = jaString.replace('\\n', '')
jaString = jaString.replace('\n', '')
# Remove any textwrap
if FIXTEXTWRAP == True:
jaString = jaString.replace('\n', '')
# Choices
if '0100410000000' in jaString:
decodedBITCH = bytes.fromhex(jaString).decode('shiftjis')
matchList = re.findall(r'd(.+?),', decodedBITCH)
if len(matchList) > 0:
for match in matchList:
response = translateGPT(match, 'Keep your translation as brief as possible. Previous text: ' + textHistory[len(textHistory)-1] + '\n\nReply in the style of a dialogue option.', True)
tokens += response[1]
# Translate
if jaString != '':
response = translateGPT(f'{speaker} | {jaString}', textHistory, True)
tokens[0] += response[1][0]
tokens[1] += response[1][1]
translatedText = response[0]
textHistory.append('\"' + translatedText + '\"')
else:
translatedText = jaString
textHistory.append('\"' + translatedText + '\"')
# Remove characters that may break scripts
charList = ['.', '\"', '\\n']
for char in charList:
translatedText = translatedText.replace(char, '')
decodedBITCH = decodedBITCH.replace(match, translatedText.replace(' ', '\u3000'))
data[i] = decodedBITCH.encode('shift-jis').hex() + '\n'
continue
# Remove added speaker
translatedText = re.sub(r'^.+?\s\|\s?', '', translatedText)
# Reset Speaker
if '00000000' == jaString:
i += 1
pbar.update(1)
speaker = ''
jaString = data[i]
# Textwrap
translatedText = textwrap.fill(translatedText, width=WIDTH)
# Grab and Translate Speaker
elif re.search(r'^0000[1-9]000$', jaString):
i += 1
pbar.update(1)
jaString = data[i].replace('\n', '')
jaString = jaString.replace('拓海', 'Takumi')
jaString = jaString.replace('こはる', 'Koharu')
jaString = jaString.replace('理央', 'Rio')
jaString = jaString.replace('アリサ', 'Arisa')
jaString = jaString.replace('友里子', 'Yuriko')
# Translate Speaker
response = translateGPT(jaString, 'Reply with only the english translation of the NPC name', True)
tokens += response[1]
speaker = response[0].strip('.')
data[i] = speaker + '\n'
# Set Data
item['message'] = translatedText
# Set index to line
i += 1
# Keep textHistory list at length maxHistory
if len(textHistory) > maxHistory:
textHistory.pop(0)
currentGroup = []
pbar.update(1)
else:
pbar.update(1)
continue
# Translate
finalJAString = data[i]
return tokens
# Remove Textwrap
finalJAString = finalJAString.replace('\\n', ' ')
finalJAString = finalJAString.replace('\n', ' ')
if speaker == '':
speaker = 'Takumi'
response = translateGPT(speaker + ': ' + finalJAString, textHistory, True)
tokens += response[1]
translatedText = response[0]
# Remove Textwrap
translatedText = translatedText.replace('\\n', ' ')
translatedText = translatedText.replace('\n', ' ')
# Remove added speaker and quotes
translatedText = re.sub(r'^.+?:\s', '', translatedText)
# TextHistory is what we use to give GPT Context, so thats appended here.
if speaker != '':
textHistory.append(speaker + ': ' + translatedText)
elif speakerFlag == False:
textHistory.append('\"' + translatedText + '\"')
# Keep textHistory list at length maxHistory
if len(textHistory) > maxHistory:
textHistory.pop(0)
currentGroup = []
# Textwrap
translatedText = textwrap.fill(translatedText, width=WIDTH)
translatedText = translatedText.replace(',\n', ', \n')
translatedText = translatedText.replace('\n', '\\n')
translatedText = translatedText.replace(',\\n', ', \\n')
# Set Data
data[i] = translatedText + '\n'
syncIndex = i + 1
pbar.update(1)
return [data, tokens]
def subVars(jaString):
jaString = jaString.replace('\u3000', ' ')
@ -347,9 +293,18 @@ def resubVars(translatedText, allList):
def translateGPT(t, history, fullPromptFlag):
# If ESTIMATE is True just count this as an execution and return.
if ESTIMATE:
enc = tiktoken.encoding_for_model("gpt-4")
tokens = len(enc.encode(t)) * 2 + len(enc.encode(str(history))) + len(enc.encode(PROMPT))
return (t, tokens)
enc = tiktoken.encoding_for_model("gpt-3.5-turbo")
historyRaw = ''
if isinstance(history, list):
for line in history:
historyRaw += line
else:
historyRaw = history
inputTotalTokens = len(enc.encode(historyRaw)) + len(enc.encode(PROMPT))
outputTotalTokens = len(enc.encode(t)) * 2 # Estimating 2x the size of the original text
totalTokens = [inputTotalTokens, outputTotalTokens]
return (t, totalTokens)
# Sub Vars
varResponse = subVars(t)
@ -357,16 +312,15 @@ def translateGPT(t, history, fullPromptFlag):
# If there isn't any Japanese in the text just skip
if not re.search(r'[一-龠]+|[ぁ-ゔ]+|[ァ-ヴ]+|[\uFF00-\uFFEF]', subbedT):
return(t, 0)
return(t, [0,0])
# Characters
context = '```\
Game Characters:\
Character: 池ノ上 拓海 == Ikenoue Takumi - Gender: Male\
Character: 福永 こはる == Fukunaga Koharu - Gender: Female\
Character: ソル == Sol - Gender: Female\
Character: ェニ先生 == Eni-sensei - Gender: Female\
Character: 神泉 理央 == Kamiizumi Rio - Gender: Female\
Character: 吉祥寺 アリサ == Kisshouji Arisa - Gender: Female\
Character: 久我 友里子 == Kuga Yuriko - Gender: Female\
```'
# Prompt
@ -392,14 +346,14 @@ def translateGPT(t, history, fullPromptFlag):
temperature=0.1,
frequency_penalty=0.2,
presence_penalty=0.2,
model="gpt-3.5-turbo",
model="gpt-3.5-turbo-1106",
messages=msg,
request_timeout=30,
)
# Save Translated Text
translatedText = response.choices[0].message.content
tokens = response.usage.total_tokens
totalTokens = [response.usage.prompt_tokens, response.usage.completion_tokens]
# Resub Vars
translatedText = resubVars(translatedText, varResponse[1])
@ -422,4 +376,4 @@ def translateGPT(t, history, fullPromptFlag):
if len(translatedText) > 15 * len(t) or "I'm sorry, but I'm unable to assist with that translation" in translatedText:
raise Exception
else:
return [translatedText, tokens]
return [translatedText, totalTokens]

View file

@ -138,7 +138,7 @@ def main():
# Open File (Threads)
with ThreadPoolExecutor(max_workers=THREADS) as executor:
futures = [executor.submit(handleLune, filename, estimate) \
for filename in os.listdir("files") if filename.endswith('txt')]
for filename in os.listdir("files") if filename.endswith('json')]
for future in as_completed(futures):
try: