Leon/docutranslate
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

增加对mineru的支持

Browse Source
This commit is contained in:
xunbu
2025-05-20 18:16:58 +08:00
parent a55d20af77
commit 85573561e4
10 changed files with 814 additions and 263 deletions
Download Patch File Download Diff File Expand all files Collapse all files

93
docutranslate/app.py
View File

@@ -11,7 +11,7 @@ from fastapi import FastAPI, File, Form, UploadFile, Request, HTTPException
from fastapi.responses import HTMLResponse, JSONResponse, StreamingResponse,FileResponse
from fastapi.templating import Jinja2Templates
from fastapi.staticfiles import StaticFiles
from docutranslate import FileTranslater
from docutranslate import FileTranslater # Assuming FileTranslater is in docutranslate module
from docutranslate.logger import translater_logger
from docutranslate.utils.resource_utils import resource_path
@@ -19,11 +19,10 @@ app = FastAPI()
STATIC_DIR=resource_path("static")
# print(f"__file__:{Path(__file__).resolve()}")
app.mount("/static",StaticFiles(directory=STATIC_DIR), name="static")
# --- 全局配置 ---
log_queue: Optional[asyncio.Queue] = None # Will be initialized in startup_event
log_queue: Optional[asyncio.Queue] = None
current_state: Dict[str, Any] = {
"is_processing": False,
"status_message": "空闲",
@@ -36,9 +35,9 @@ current_state: Dict[str, Any] = {
"task_end_time": 0,
"current_task_ref": None,
}
templates = Jinja2Templates(directory=".")
MAX_LOG_HISTORY = 200 # Max items for the persistent log_history list
log_history: List[str] = [] # Keeps a longer history, not directly for "unread"
templates = Jinja2Templates(directory=".") # Not strictly used if index.html is served as FileResponse
MAX_LOG_HISTORY = 200
log_history: List[str] = []
# --- 日志处理器 ---
@@ -51,27 +50,22 @@ class QueueAndHistoryHandler(logging.Handler):
def emit(self, record: logging.LogRecord):
log_entry = self.format(record)
# Add to the persistent history (capped)
print(log_entry) # Keep console log for server visibility
self.history_list.append(log_entry)
if len(self.history_list) > self.max_history:
del self.history_list[:len(self.history_list) - self.max_history]
# Add to the "unread" queue for frontend consumption
try:
# Ensure self.queue is not None (it's initialized at startup)
if self.queue is not None:
if self.queue is not None:
try:
main_loop = getattr(app.state, "main_event_loop", None)
if main_loop and main_loop.is_running():
main_loop.call_soon_threadsafe(self.queue.put_nowait, log_entry)
else:
self.queue.put_nowait(log_entry) # Fallback
else:
print(f"CRITICAL: Log queue not initialized. Log: {log_entry}")
except asyncio.QueueFull:
print(f"Log queue is full. Log dropped: {log_entry}") # Or handle differently
except Exception as e:
print(f"Error putting log to queue: {e}. Log: {log_entry}")
self.queue.put_nowait(log_entry)
except asyncio.QueueFull:
print(f"Log queue is full. Log dropped: {log_entry}")
except Exception as e:
print(f"Error putting log to queue: {e}. Log: {log_entry}")
# --- 应用生命周期事件 ---
@@ -79,7 +73,7 @@ class QueueAndHistoryHandler(logging.Handler):
async def startup_event():
global log_queue
app.state.main_event_loop = asyncio.get_running_loop()
log_queue = asyncio.Queue() # Initialize the global log_queue
log_queue = asyncio.Queue()
for handler in translater_logger.handlers[:]:
translater_logger.removeHandler(handler)
@@ -93,7 +87,7 @@ async def startup_event():
translater_logger.setLevel(logging.INFO)
log_history.clear()
while not log_queue.empty(): # Clear queue just in case
while not log_queue.empty():
try:
log_queue.get_nowait()
except asyncio.QueueEmpty:
@@ -112,6 +106,7 @@ async def _perform_translation(params: Dict[str, Any], file_contents: bytes, ori
try:
translater_logger.info(f"使用 Base URL: {params['base_url']}, Model: {params['model_id']}")
translater_logger.info(f"文件大小: {len(file_contents)} 字节。目标语言: {params['to_lang']}")
translater_logger.info(f"使用转换引擎: {params['convert_engin']}")
translater_logger.info(
f"选项 - 公式: {params['formula_ocr']}, 代码: {params['code_ocr']}, 修正: {params['refine_markdown']}")
@@ -119,7 +114,9 @@ async def _perform_translation(params: Dict[str, Any], file_contents: bytes, ori
base_url=params['base_url'],
key=params['apikey'],
model_id=params['model_id'],
tips=False
convert_engin=params['convert_engin'],
mineru_token=params['mineru_token'],
tips=False # Assuming tips are not needed for server-side processing
)
await ft.translate_bytes_async(
name=original_filename,
@@ -152,7 +149,7 @@ async def _perform_translation(params: Dict[str, Any], file_contents: bytes, ori
translater_logger.info(f"翻译任务 '{original_filename}' 已被取消 (用时 {duration:.2f} 秒).")
current_state.update({
"status_message": f"翻译任务已取消（若有转换任务仍会后台进行） (用时 {duration:.2f} 秒).",
"error_flag": False, # Cancellation is not an error in this context
"error_flag": False,
"download_ready": False,
"markdown_content": None,
"html_content": None,
@@ -180,11 +177,25 @@ async def _perform_translation(params: Dict[str, Any], file_contents: bytes, ori
# --- API Endpoints ---
@app.get("/", response_class=HTMLResponse)
async def main_page(request: Request):
return FileResponse(STATIC_DIR/"index.html")
# Serve index.html from the static directory or root project directory
# Assuming index.html is at the same level as app.py or in STATIC_DIR
# For simplicity, if index.html is at root:
# return FileResponse(Path(__file__).parent / "index.html")
# If using Jinja2Templates and index.html is in "templates" folder:
# return templates.TemplateResponse("index.html", {"request": request})
# Using FileResponse for index.html directly:
index_path = Path("index.html") # Adjust if index.html is elsewhere
if not index_path.exists():
# Fallback to static dir if not in root
index_path = STATIC_DIR / "index.html"
if not index_path.exists():
raise HTTPException(status_code=404, detail="index.html not found")
return FileResponse(index_path)
@app.post("/translate")
async def handle_translate(
request: Request, # Added request for potential future use, not strictly needed now
base_url: str = Form(...),
apikey: str = Form(...),
model_id: str = Form(...),
@@ -192,6 +203,8 @@ async def handle_translate(
formula_ocr: bool = Form(False),
code_ocr: bool = Form(False),
refine_markdown: bool = Form(False),
convert_engin: str = Form(...), # New parameter
mineru_token: Optional[str] = Form(None), # New parameter
file: UploadFile = File(...)
):
global current_state, log_queue, log_history
@@ -209,6 +222,12 @@ async def handle_translate(
content={"task_started": False, "message": "没有选择文件或文件无效。"}
)
if convert_engin == "mineru" and (not mineru_token or not mineru_token.strip()):
return JSONResponse(
status_code=400,
content={"task_started": False, "message": "使用 Mineru 引擎时必须提供有效的 Mineru Token。"}
)
current_state["is_processing"] = True
original_filename_for_init = file.filename or "uploaded_file"
@@ -224,26 +243,22 @@ async def handle_translate(
"current_task_ref": None,
})
# Clear logs for the new task
log_history.clear()
if log_queue: # Ensure log_queue is initialized
if log_queue:
while not log_queue.empty():
try:
log_queue.get_nowait()
except asyncio.QueueEmpty:
break
# Add initial log entry for the new task
# We create a LogRecord manually to ensure it goes through the formatter and handler
initial_log_msg = f"收到新的翻译请求: {original_filename_for_init}"
if translater_logger.handlers and isinstance(translater_logger.handlers[0], QueueAndHistoryHandler):
# Use the existing handler to format and queue/store the log
record = logging.LogRecord(
name=translater_logger.name, level=logging.INFO, pathname="", lineno=0,
msg=initial_log_msg, args=(), exc_info=None, func=""
)
translater_logger.handlers[0].emit(record) # This will add to both queue and history
else: # Fallback if handler setup is unusual
translater_logger.handlers[0].emit(record)
else:
translater_logger.info(initial_log_msg)
try:
@@ -255,6 +270,8 @@ async def handle_translate(
"base_url": base_url, "apikey": apikey, "model_id": model_id,
"to_lang": to_lang, "formula_ocr": formula_ocr,
"code_ocr": code_ocr, "refine_markdown": refine_markdown,
"convert_engin": convert_engin, # Pass to task
"mineru_token": mineru_token, # Pass to task
}
loop = asyncio.get_running_loop()
@@ -332,18 +349,17 @@ async def get_status():
@app.get("/get-logs")
async def get_logs_from_queue(): # Renamed for clarity, though path is the same
async def get_logs_from_queue():
global log_queue
new_logs = []
if log_queue: # Ensure log_queue is initialized
if log_queue:
while not log_queue.empty():
try:
log_entry = log_queue.get_nowait() # Consume from queue
log_entry = log_queue.get_nowait()
new_logs.append(log_entry)
log_queue.task_done() # Important for queue management if using join() elsewhere
log_queue.task_done()
except asyncio.QueueEmpty:
break
# No total_count, as the frontend just appends what it receives
return JSONResponse(content={"logs": new_logs})
@@ -384,10 +400,11 @@ async def download_html(filename_with_ext: str):
def run_app():
print("正在启动 DocuTranslate")
print("正在启动 DocuTranslate WebUI")
print("请访问 http://127.0.0.1:8010")
uvicorn.run(app, host="127.0.0.1", port=8010, workers=1)
if __name__ == "__main__":
run_app()
run_app()