DazedAnon
65a7ee51c2
- Skip candidate batches when every target is translated - Keep mixed batches and use global BATCHSIZE for checks - Add CSV setting checkbox and unit coverage
491 lines
17 KiB
Python
491 lines
17 KiB
Python
# Libraries
|
||
import json
|
||
import os
|
||
import re
|
||
import util.dazedwrap as dazedwrap
|
||
import threading
|
||
import time
|
||
import traceback
|
||
import tiktoken
|
||
import csv
|
||
from pathlib import Path
|
||
from colorama import Fore
|
||
from dotenv import load_dotenv
|
||
from retry import retry
|
||
from tqdm import tqdm
|
||
from util.translation import TranslationConfig, translateAI as sharedtranslateAI, getPricingConfig, calculateCost
|
||
from util.speakers import strip_speaker_prefix
|
||
import tempfile
|
||
|
||
# Globals
|
||
MODEL = os.getenv("model")
|
||
TIMEOUT = int(os.getenv("timeout"))
|
||
LANGUAGE = os.getenv("language").capitalize()
|
||
from util.paths import PROMPT_PATH, VOCAB_PATH
|
||
|
||
PROMPT = PROMPT_PATH.read_text(encoding="utf-8")
|
||
VOCAB = VOCAB_PATH.read_text(encoding="utf-8")
|
||
LOCK = threading.Lock()
|
||
WIDTH = int(os.getenv("width"))
|
||
LISTWIDTH = int(os.getenv("listWidth"))
|
||
NOTEWIDTH = int(os.getenv("noteWidth"))
|
||
MAXHISTORY = 10
|
||
ESTIMATE = ""
|
||
TOKENS = [0, 0]
|
||
NAMESLIST = []
|
||
NAMES = False # Output a list of all the character names found
|
||
BRFLAG = False # If the game uses <br> instead
|
||
FIXTEXTWRAP = True # Overwrites textwrap
|
||
IGNORETLTEXT = True # Ignores all translated text.
|
||
MISMATCH = [] # Lists files that thdata a mismatch error (Length of GPT list response is wrong)
|
||
FILENAME = None
|
||
BRACKETNAMES = False
|
||
|
||
# CSV Configuration Settings (configurable via GUI)
|
||
CSV_DELIMITER = " " # CSV delimiter character (comma, semicolon, tab)
|
||
SOURCE_COLUMN = 2 # Which column has the source text to translate (0-indexed)
|
||
TARGET_COLUMN = 3 # Which column to write translations to
|
||
SPEAKER_COLUMN = 1 # Which column has speaker names (-1 = none)
|
||
SKIP_HEADER_ROW = False # Skip the first row (header)
|
||
USE_TARGET_IF_NOT_EMPTY = False # Use target column text if not empty (T++ style)
|
||
SKIP_IF_TARGET_TRANSLATED = False # Skip batches whose targets are already translated
|
||
WRITE_TO_NEXT_COLUMN = False # Write to column after target instead of overwriting
|
||
PARSE_NAME_TAGS = False # Parse :name[] tags in text
|
||
PARSE_M_MARKERS = False # Parse \M markers in text
|
||
REMOVE_FURIGANA = True # Remove furigana annotations ＜＝＞
|
||
SKIP_COMMENT_ROWS = False # Skip rows starting with 'comment'
|
||
|
||
# Regex - Need to change this if you want to translate from/to other languages. Default is Japanese Regex
|
||
LANGREGEX = r"[一-龠ぁ-ゔァ-ヴーａ-ｚＡ-Ｚ０-９\uFF61-\uFF9F]+"
|
||
|
||
# Get pricing configuration based on the model
|
||
PRICING_CONFIG = getPricingConfig(MODEL)
|
||
INPUTAPICOST = PRICING_CONFIG["inputAPICost"]
|
||
OUTPUTAPICOST = PRICING_CONFIG["outputAPICost"]
|
||
BATCHSIZE = PRICING_CONFIG["batchSize"]
|
||
FREQUENCY_PENALTY = PRICING_CONFIG["frequencyPenalty"]
|
||
|
||
# tqdm Globals
|
||
BAR_FORMAT = "{l_bar}{bar:10}{r_bar}{bar:-10b}"
|
||
POSITION = 0
|
||
LEAVE = False
|
||
PBAR = None
|
||
ENCODING = "utf8"
|
||
|
||
# Initialize Translation Config
|
||
TRANSLATION_CONFIG = TranslationConfig(
|
||
model=MODEL,
|
||
language=LANGUAGE,
|
||
prompt=PROMPT,
|
||
vocab=VOCAB,
|
||
langRegex=LANGREGEX,
|
||
batchSize=BATCHSIZE,
|
||
maxHistory=MAXHISTORY,
|
||
estimateMode=False # Will be set dynamically based on ESTIMATE
|
||
)
|
||
LEAVE = False
|
||
|
||
def handleCSV(filename, estimate):
|
||
global ESTIMATE, TOKENS, FILENAME
|
||
ESTIMATE = estimate
|
||
FILENAME = filename
|
||
|
||
if not ESTIMATE:
|
||
with open("translated/" + filename, "w+t", newline="", encoding=ENCODING, errors="xmlcharrefreplace") as writeFile:
|
||
# Translate
|
||
start = time.time()
|
||
translatedData = openFiles(filename, writeFile)
|
||
|
||
# Print Result
|
||
end = time.time()
|
||
tqdm.write(getResultString(translatedData, end - start, filename))
|
||
with LOCK:
|
||
TOKENS[0] += translatedData[1][0]
|
||
TOKENS[1] += translatedData[1][1]
|
||
else:
|
||
# Translate
|
||
start = time.time()
|
||
translatedData = openFilesEstimate(filename)
|
||
|
||
# Print Result
|
||
end = time.time()
|
||
tqdm.write(getResultString(translatedData, end - start, filename))
|
||
with LOCK:
|
||
TOKENS[0] += translatedData[1][0]
|
||
TOKENS[1] += translatedData[1][1]
|
||
|
||
# Print Total
|
||
totalString = getResultString(["", TOKENS, None], end - start, "TOTAL")
|
||
|
||
# Print any errors on maps
|
||
if len(MISMATCH) > 0:
|
||
return totalString + Fore.RED + f"\nMismatch Errors: {MISMATCH}" + Fore.RESET
|
||
else:
|
||
return totalString
|
||
|
||
|
||
def openFiles(filename, writeFile):
|
||
with open("files/" + filename, "r", encoding=ENCODING) as readFile, writeFile:
|
||
translatedData = parseCSV(readFile, writeFile, filename)
|
||
|
||
return translatedData
|
||
|
||
|
||
def openFilesEstimate(filename):
|
||
with open("files/" + filename, "r", encoding="utf8") as readFile:
|
||
translatedData = parseCSV(readFile, "", filename)
|
||
|
||
return translatedData
|
||
|
||
|
||
def getResultString(translatedData, translationTime, filename):
|
||
# File Print String
|
||
cost = calculateCost(translatedData[1][0], translatedData[1][1], MODEL)
|
||
totalTokenstring = (
|
||
Fore.YELLOW + "[Input: " + str(translatedData[1][0]) + "]"
|
||
"[Output: "
|
||
+ str(translatedData[1][1])
|
||
+ "]" "[Cost: ${:,.4f}".format(cost)
|
||
+ "]"
|
||
)
|
||
timeString = Fore.BLUE + "[" + str(round(translationTime, 1)) + "s]"
|
||
|
||
if translatedData[2] is None:
|
||
# Success
|
||
return filename + ": " + totalTokenstring + timeString + Fore.GREEN + " \u2713 " + Fore.RESET
|
||
else:
|
||
# Fail
|
||
try:
|
||
raise translatedData[2]
|
||
except Exception as e:
|
||
traceback.print_exc()
|
||
errorString = str(e) + Fore.RED
|
||
return filename + ": " + totalTokenstring + timeString + Fore.RED + " \u2717 " + errorString + Fore.RESET
|
||
|
||
|
||
def parseCSV(readFile, writeFile, filename):
|
||
totalTokens = [0, 0]
|
||
totalLines = 0
|
||
global LOCK
|
||
|
||
# Get total for progress bar
|
||
totalLines = len(readFile.readlines())
|
||
readFile.seek(0)
|
||
|
||
reader = csv.reader(readFile, delimiter=CSV_DELIMITER)
|
||
if not ESTIMATE:
|
||
writer = csv.writer(
|
||
writeFile,
|
||
delimiter=CSV_DELIMITER,
|
||
)
|
||
else:
|
||
writer = ""
|
||
|
||
with tqdm(bar_format=BAR_FORMAT, position=POSITION, total=totalLines, leave=LEAVE) as pbar:
|
||
pbar.desc = filename
|
||
pbar.total = totalLines
|
||
|
||
# Grab All Rows
|
||
data = []
|
||
for row in reader:
|
||
data.append(row)
|
||
|
||
try:
|
||
response = translateCSV(data, pbar, writeFile, writer, filename, None)
|
||
totalTokens[0] = response[0]
|
||
totalTokens[1] = response[1]
|
||
except Exception:
|
||
traceback.print_exc()
|
||
return [data, totalTokens, None]
|
||
|
||
|
||
def flush_progress_csv(writeFile, writer, rows):
|
||
"""Flush current CSV progress to the already-open output file (Windows-safe)."""
|
||
try:
|
||
if ESTIMATE or writeFile is None:
|
||
return
|
||
with LOCK:
|
||
writeFile.seek(0)
|
||
# Recreate writer at current position to avoid state issues
|
||
tmp_writer = csv.writer(writeFile, delimiter=CSV_DELIMITER)
|
||
tmp_writer.writerows(rows)
|
||
writeFile.truncate()
|
||
writeFile.flush()
|
||
os.fsync(writeFile.fileno())
|
||
except Exception:
|
||
traceback.print_exc()
|
||
|
||
|
||
def _actual_target_column():
|
||
"""Column index translations are written to."""
|
||
return TARGET_COLUMN + 1 if WRITE_TO_NEXT_COLUMN else TARGET_COLUMN
|
||
|
||
|
||
def _target_is_translated(row):
|
||
"""True when the write-target cell is non-empty and has no Japanese left."""
|
||
actual_target = _actual_target_column()
|
||
if len(row) <= actual_target:
|
||
return False
|
||
target = row[actual_target]
|
||
if not isinstance(target, str) or not target.strip():
|
||
return False
|
||
return not re.search(LANGREGEX, target)
|
||
|
||
|
||
def _row_source_text(row):
|
||
"""Source text for a row, honoring USE_TARGET_IF_NOT_EMPTY."""
|
||
if USE_TARGET_IF_NOT_EMPTY and len(row) > TARGET_COLUMN and row[TARGET_COLUMN]:
|
||
return row[TARGET_COLUMN]
|
||
if len(row) > SOURCE_COLUMN and row[SOURCE_COLUMN]:
|
||
return row[SOURCE_COLUMN]
|
||
return ""
|
||
|
||
|
||
def _is_candidate_row(data, i):
|
||
"""True when this row would normally be collected for translation."""
|
||
if SKIP_HEADER_ROW and i == 0:
|
||
return False
|
||
if SKIP_COMMENT_ROWS and len(data[i]) > 0 and "comment" in str(data[i][0]).lower():
|
||
return False
|
||
if len(data[i]) <= SOURCE_COLUMN:
|
||
return False
|
||
return bool(_row_source_text(data[i]))
|
||
|
||
|
||
def _collect_process_indices(data):
|
||
"""Row indices to translate.
|
||
|
||
When SKIP_IF_TARGET_TRANSLATED is on, candidates are checked in groups of
|
||
BATCHSIZE (same as the translation batch size). A group is skipped only if
|
||
every target in it is already translated; if any row still needs work, the
|
||
whole group runs.
|
||
"""
|
||
candidates = [i for i in range(len(data)) if _is_candidate_row(data, i)]
|
||
if not SKIP_IF_TARGET_TRANSLATED:
|
||
return candidates
|
||
|
||
process = []
|
||
batch_size = max(1, int(BATCHSIZE))
|
||
for start in range(0, len(candidates), batch_size):
|
||
batch = candidates[start:start + batch_size]
|
||
if all(_target_is_translated(data[i]) for i in batch):
|
||
continue
|
||
process.extend(batch)
|
||
return process
|
||
|
||
|
||
def translateCSV(data, pbar, writeFile, writer, filename, translatedList):
|
||
"""
|
||
Unified CSV translation function using configurable settings.
|
||
|
||
Uses global settings:
|
||
- SOURCE_COLUMN: column index for source text
|
||
- TARGET_COLUMN: column index to write translations
|
||
- SPEAKER_COLUMN: column index for speaker names (-1 = none)
|
||
- SKIP_HEADER_ROW: whether to skip first row
|
||
- USE_TARGET_IF_NOT_EMPTY: use existing target text if present (T++ style)
|
||
- SKIP_IF_TARGET_TRANSLATED: skip fully-translated candidate batches
|
||
- WRITE_TO_NEXT_COLUMN: write to column after target
|
||
- PARSE_NAME_TAGS: parse :name[] tags
|
||
- PARSE_M_MARKERS: parse \\M markers
|
||
- REMOVE_FURIGANA: remove furigana
|
||
- SKIP_COMMENT_ROWS: skip rows with 'comment' in first column
|
||
"""
|
||
global LOCK, ESTIMATE, PBAR
|
||
PBAR = pbar
|
||
translatedText = ""
|
||
totalTokens = [0, 0]
|
||
stringList = []
|
||
process_indices = _collect_process_indices(data)
|
||
|
||
try:
|
||
# Translate
|
||
for i in process_indices:
|
||
# Get source text
|
||
jaString = _row_source_text(data[i])
|
||
speaker = ""
|
||
|
||
# Handle speaker column if configured
|
||
if SPEAKER_COLUMN >= 0 and len(data[i]) > SPEAKER_COLUMN and data[i][SPEAKER_COLUMN]:
|
||
speakerResponse = getSpeaker(data[i][SPEAKER_COLUMN])
|
||
totalTokens[0] += speakerResponse[1][0]
|
||
totalTokens[1] += speakerResponse[1][1]
|
||
speaker = speakerResponse[0]
|
||
data[i][SPEAKER_COLUMN] = speaker
|
||
|
||
# Parse :name[] tags if configured
|
||
if PARSE_NAME_TAGS and ':name' in jaString:
|
||
match = re.search(r":name\[([^\]]+?)\]\n([\w\W]*)", jaString)
|
||
if match:
|
||
# Translate speaker name
|
||
response = getSpeaker(match.group(1))
|
||
speaker = response[0]
|
||
totalTokens[0] += response[1][0]
|
||
totalTokens[1] += response[1][1]
|
||
data[i][SOURCE_COLUMN] = data[i][SOURCE_COLUMN].replace(match.group(1), speaker)
|
||
|
||
# Extract text portion
|
||
jaString = match.group(2)
|
||
# Remove voice markers
|
||
voMatch = re.search(r"\\[vfF]+\[.+]", jaString)
|
||
if voMatch:
|
||
jaString = jaString.replace(voMatch.group(0), "")
|
||
|
||
# Parse \M markers if configured
|
||
if PARSE_M_MARKERS and '\\M' in jaString:
|
||
match = re.search(r"\\M.+\n([\w\W]*)", jaString)
|
||
if match:
|
||
jaString = match.group(1)
|
||
voMatch = re.search(r"\\[vfF]+\[.+]", jaString)
|
||
if voMatch:
|
||
jaString = jaString.replace(voMatch.group(0), "")
|
||
|
||
# Remove furigana if configured
|
||
if REMOVE_FURIGANA:
|
||
jaString = re.sub(r"＜(.*)＝.*＞", r"\1", jaString)
|
||
|
||
# Store original for replacement
|
||
ojaString = jaString
|
||
|
||
# Remove textwrap
|
||
jaString = jaString.replace("\\n", " ")
|
||
jaString = jaString.replace("\n", " ")
|
||
|
||
# Pass 1: Collect strings
|
||
if not translatedList:
|
||
if speaker:
|
||
stringList.append(f"[{speaker}]: {jaString}")
|
||
else:
|
||
stringList.append(jaString)
|
||
|
||
# Pass 2: Apply translations
|
||
else:
|
||
# Grab and pop translation
|
||
translatedText = translatedList[0]
|
||
translatedList.pop(0)
|
||
|
||
# Remove speaker prefix from translation if present
|
||
translatedText = strip_speaker_prefix(translatedText)
|
||
|
||
# Add wordwrap
|
||
translatedText = dazedwrap.wrapText(translatedText, WIDTH)
|
||
translatedText = translatedText.replace("\n", "\\n")
|
||
|
||
# Determine target column
|
||
actual_target = _actual_target_column()
|
||
|
||
# Ensure row has enough columns
|
||
while len(data[i]) <= actual_target:
|
||
data[i].append("")
|
||
|
||
# Set data
|
||
if PARSE_NAME_TAGS or PARSE_M_MARKERS:
|
||
# Replace original text portion
|
||
data[i][actual_target] = data[i][actual_target].replace(ojaString, translatedText) if data[i][actual_target] else translatedText
|
||
else:
|
||
data[i][actual_target] = translatedText
|
||
|
||
flush_progress_csv(writeFile, writer, data)
|
||
|
||
# EOF - Process collected strings
|
||
if len(stringList) > 0:
|
||
# Set Progress
|
||
pbar.total = len(stringList)
|
||
pbar.refresh()
|
||
|
||
# Translate
|
||
response = translateAI(stringList, "")
|
||
totalTokens[0] += response[1][0]
|
||
totalTokens[1] += response[1][1]
|
||
translatedList = response[0]
|
||
|
||
# Set Strings
|
||
if len(stringList) == len(translatedList):
|
||
translateCSV(data, pbar, writeFile, writer, filename, translatedList)
|
||
# Mismatch
|
||
else:
|
||
with LOCK:
|
||
if filename not in MISMATCH:
|
||
MISMATCH.append(filename)
|
||
|
||
# Write all Data
|
||
with LOCK:
|
||
if not ESTIMATE:
|
||
for row in data:
|
||
writer.writerow(row)
|
||
flush_progress_csv(writeFile, writer, data)
|
||
|
||
except Exception:
|
||
traceback.print_exc()
|
||
|
||
# Write all Data
|
||
with LOCK:
|
||
if not ESTIMATE:
|
||
for row in data:
|
||
writer.writerow(row)
|
||
flush_progress_csv(writeFile, writer, data)
|
||
return totalTokens
|
||
|
||
return totalTokens
|
||
|
||
# Save some money and enter the character before translation
|
||
def getSpeaker(speaker):
|
||
match speaker:
|
||
case "ファイン":
|
||
return ["Fine", [0, 0]]
|
||
case "":
|
||
return ["", [0, 0]]
|
||
case _:
|
||
# Find Speaker
|
||
for i in range(len(NAMESLIST)):
|
||
if speaker == NAMESLIST[i][0]:
|
||
return [NAMESLIST[i][1], [0, 0]]
|
||
|
||
if speaker == "？？？":
|
||
return ["???", [0, 0]]
|
||
|
||
# Translate and Store Speaker
|
||
response = translateAI(
|
||
f"{speaker}",
|
||
"Reply with the " + LANGUAGE + " translation of the NPC name.",
|
||
False,
|
||
)
|
||
response[0] = response[0].title()
|
||
response[0] = response[0].replace("'S", "'s")
|
||
response[0] = response[0].replace("Speaker: ", "")
|
||
|
||
# Retry if name doesn't translate for some reason
|
||
if re.search(r"([a-zA-Z？?])", response[0]) == None:
|
||
response = translateAI(
|
||
f"{speaker}",
|
||
"Reply with the " + LANGUAGE + " translation of the NPC name.",
|
||
False,
|
||
)
|
||
response[0] = response[0].title()
|
||
response[0] = response[0].replace("'S", "'s")
|
||
|
||
speakerList = [speaker, response[0]]
|
||
NAMESLIST.append(speakerList)
|
||
return response
|
||
return [speaker, [0, 0]]
|
||
|
||
def translateAI(text, history, history_ctx=None):
|
||
"""
|
||
Legacy wrapper function for the new shared translation utility.
|
||
This maintains compatibility with existing code while using the new shared implementation.
|
||
"""
|
||
global PBAR, MISMATCH, FILENAME
|
||
|
||
# Update config estimate mode based on global ESTIMATE
|
||
TRANSLATION_CONFIG.estimateMode = bool(ESTIMATE)
|
||
|
||
# Call the new shared translation function
|
||
return sharedtranslateAI(
|
||
text=text,
|
||
history=history,
|
||
config=TRANSLATION_CONFIG,
|
||
filename=FILENAME,
|
||
pbar=PBAR,
|
||
lock=LOCK,
|
||
mismatchList=MISMATCH
|
||
)
|