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

agent增加system_proxy_enable参数

Browse Source
This commit is contained in:
xunbu
2025-09-23 23:53:49 +08:00
parent 760b1b36dd
commit f611468a13
10 changed files with 83 additions and 72 deletions
Download Patch File Download Diff File Expand all files Collapse all files

128
docutranslate/agents/agent.py
View File

@@ -48,6 +48,7 @@ class AgentConfig:
timeout: int = 1200 # 单位(秒)，这个值是httpx.TimeOut中read的值,并非总的超时时间 timeout: int = 1200 # 单位(秒)，这个值是httpx.TimeOut中read的值,并非总的超时时间
thinking: ThinkingMode = "disable" thinking: ThinkingMode = "disable"
retry: int = 2 retry: int = 2
system_proxy_enable: bool = USE_PROXY
class TotalErrorCounter: class TotalErrorCounter:
@@ -111,14 +112,14 @@ def extract_token_info(response_data: dict) -> tuple[int, int, int, int]:
# 尝试从不同格式获取cached_tokens # 尝试从不同格式获取cached_tokens
# 格式1: input_tokens_details.cached_tokens # 格式1: input_tokens_details.cached_tokens
if ( if (
"input_tokens_details" in usage "input_tokens_details" in usage
and "cached_tokens" in usage["input_tokens_details"] and "cached_tokens" in usage["input_tokens_details"]
): ):
cached_tokens = usage["input_tokens_details"]["cached_tokens"] cached_tokens = usage["input_tokens_details"]["cached_tokens"]
# 格式2: prompt_tokens_details.cached_tokens # 格式2: prompt_tokens_details.cached_tokens
elif ( elif (
"prompt_tokens_details" in usage "prompt_tokens_details" in usage
and "cached_tokens" in usage["prompt_tokens_details"] and "cached_tokens" in usage["prompt_tokens_details"]
): ):
cached_tokens = usage["prompt_tokens_details"]["cached_tokens"] cached_tokens = usage["prompt_tokens_details"]["cached_tokens"]
# 格式3: prompt_cache_hit_tokens (直接在usage下) # 格式3: prompt_cache_hit_tokens (直接在usage下)
@@ -128,20 +129,20 @@ def extract_token_info(response_data: dict) -> tuple[int, int, int, int]:
# 尝试从不同格式获取reasoning_tokens # 尝试从不同格式获取reasoning_tokens
# 格式1: output_tokens_details.reasoning_tokens # 格式1: output_tokens_details.reasoning_tokens
if ( if (
"output_tokens_details" in usage "output_tokens_details" in usage
and "reasoning_tokens" in usage["output_tokens_details"] and "reasoning_tokens" in usage["output_tokens_details"]
): ):
reasoning_tokens = usage["output_tokens_details"]["reasoning_tokens"] reasoning_tokens = usage["output_tokens_details"]["reasoning_tokens"]
# 格式2: completion_tokens_details.reasoning_tokens # 格式2: completion_tokens_details.reasoning_tokens
elif ( elif (
"completion_tokens_details" in usage "completion_tokens_details" in usage
and "reasoning_tokens" in usage["completion_tokens_details"] and "reasoning_tokens" in usage["completion_tokens_details"]
): ):
reasoning_tokens = usage["completion_tokens_details"]["reasoning_tokens"] reasoning_tokens = usage["completion_tokens_details"]["reasoning_tokens"]
return input_tokens, cached_tokens, output_tokens, reasoning_tokens return input_tokens, cached_tokens, output_tokens, reasoning_tokens
except TypeError as e: except TypeError as e:
print(f"获取token发生错误:{e.__repr__()}") print(f"获取token发生错误:{e.__repr__()}")
return -1,-1,-1,-1 return -1, -1, -1, -1
class TokenCounter: class TokenCounter:
@@ -155,11 +156,11 @@ class TokenCounter:
self.logger = logger self.logger = logger
def add( def add(
self, self,
input_tokens: int, input_tokens: int,
cached_tokens: int, cached_tokens: int,
output_tokens: int, output_tokens: int,
reasoning_tokens: int, reasoning_tokens: int,
): ):
with self.lock: with self.lock:
self.input_tokens += input_tokens self.input_tokens += input_tokens
@@ -248,6 +249,7 @@ class Agent:
self.retry = config.retry self.retry = config.retry
self.system_proxy_enable=config.system_proxy_enable
def _add_thinking_mode(self, data: dict): def _add_thinking_mode(self, data: dict):
if self.domain not in self._think_factory: if self.domain not in self._think_factory:
return return
@@ -258,7 +260,7 @@ class Agent:
data[field_thinking] = val_disable data[field_thinking] = val_disable
def _prepare_request_data( def _prepare_request_data(
self, prompt: str, system_prompt: str, temperature=None, top_p=0.9 self, prompt: str, system_prompt: str, temperature=None, top_p=0.9
): ):
if temperature is None: if temperature is None:
temperature = self.temperature temperature = self.temperature
@@ -280,16 +282,16 @@ class Agent:
return headers, data return headers, data
async def send_async( async def send_async(
self, self,
client: httpx.AsyncClient, client: httpx.AsyncClient,
prompt: str, prompt: str,
system_prompt: None | str = None, system_prompt: None | str = None,
retry=True, retry=True,
retry_count=0, retry_count=0,
pre_send_handler: PreSendHandlerType = None, pre_send_handler: PreSendHandlerType = None,
result_handler: ResultHandlerType = None, result_handler: ResultHandlerType = None,
error_result_handler: ErrorResultHandlerType = None, error_result_handler: ErrorResultHandlerType = None,
best_partial_result: dict | None = None, best_partial_result: dict | None = None,
) -> Any: ) -> Any:
if system_prompt is None: if system_prompt is None:
system_prompt = self.system_prompt system_prompt = self.system_prompt
@@ -430,13 +432,13 @@ class Agent:
) )
async def send_prompts_async( async def send_prompts_async(
self, self,
prompts: list[str], prompts: list[str],
system_prompt: str | None = None, system_prompt: str | None = None,
max_concurrent: int | None = None, max_concurrent: int | None = None,
pre_send_handler: PreSendHandlerType = None, pre_send_handler: PreSendHandlerType = None,
result_handler: ResultHandlerType = None, result_handler: ResultHandlerType = None,
error_result_handler: ErrorResultHandlerType = None, error_result_handler: ErrorResultHandlerType = None,
) -> list[Any]: ) -> list[Any]:
max_concurrent = ( max_concurrent = (
self.max_concurrent if max_concurrent is None else max_concurrent self.max_concurrent if max_concurrent is None else max_concurrent
@@ -447,7 +449,7 @@ class Agent:
) )
self.logger.info(f"预计发送{total}个请求，并发请求数:{max_concurrent}") self.logger.info(f"预计发送{total}个请求，并发请求数:{max_concurrent}")
self.total_error_counter.max_errors_count = ( self.total_error_counter.max_errors_count = (
len(prompts) // MAX_REQUESTS_PER_ERROR len(prompts) // MAX_REQUESTS_PER_ERROR
) )
# 新增：在每次批量发送前重置计数器 # 新增：在每次批量发送前重置计数器
@@ -459,7 +461,7 @@ class Agent:
semaphore = asyncio.Semaphore(max_concurrent) semaphore = asyncio.Semaphore(max_concurrent)
tasks = [] tasks = []
proxies = get_httpx_proxies() if USE_PROXY else None proxies = get_httpx_proxies() if self.system_proxy_enable else None
limits = httpx.Limits( limits = httpx.Limits(
max_connections=self.max_concurrent * 2, # 为重试和并发预留空间 max_connections=self.max_concurrent * 2, # 为重试和并发预留空间
@@ -467,7 +469,7 @@ class Agent:
) )
async with httpx.AsyncClient( async with httpx.AsyncClient(
trust_env=False, proxies=proxies, verify=False, limits=limits trust_env=False, proxies=proxies, verify=False, limits=limits
) as client: ) as client:
async def send_with_semaphore(p_text: str): async def send_with_semaphore(p_text: str):
@@ -498,7 +500,7 @@ class Agent:
# 新增：打印token使用统计 # 新增：打印token使用统计
token_stats = self.token_counter.get_stats() token_stats = self.token_counter.get_stats()
if token_stats['input_tokens']<0: if token_stats['input_tokens'] < 0:
self.logger.info("Token统计失败") self.logger.info("Token统计失败")
else: else:
self.logger.info( self.logger.info(
@@ -510,16 +512,16 @@ class Agent:
return results return results
def send( def send(
self, self,
client: httpx.Client, client: httpx.Client,
prompt: str, prompt: str,
system_prompt: None | str = None, system_prompt: None | str = None,
retry=True, retry=True,
retry_count=0, retry_count=0,
pre_send_handler=None, pre_send_handler=None,
result_handler=None, result_handler=None,
error_result_handler=None, error_result_handler=None,
best_partial_result: dict | None = None, best_partial_result: dict | None = None,
) -> Any: ) -> Any:
if system_prompt is None: if system_prompt is None:
system_prompt = self.system_prompt system_prompt = self.system_prompt
@@ -656,14 +658,14 @@ class Agent:
) )
def _send_prompt_count( def _send_prompt_count(
self, self,
client: httpx.Client, client: httpx.Client,
prompt: str, prompt: str,
system_prompt: None | str, system_prompt: None | str,
count: PromptsCounter, count: PromptsCounter,
pre_send_handler, pre_send_handler,
result_handler, result_handler,
error_result_handler, error_result_handler,
) -> Any: ) -> Any:
result = self.send( result = self.send(
client, client,
@@ -677,12 +679,12 @@ class Agent:
return result return result
def send_prompts( def send_prompts(
self, self,
prompts: list[str], prompts: list[str],
system_prompt: str | None = None, system_prompt: str | None = None,
pre_send_handler: PreSendHandlerType = None, pre_send_handler: PreSendHandlerType = None,
result_handler: ResultHandlerType = None, result_handler: ResultHandlerType = None,
error_result_handler: ErrorResultHandlerType = None, error_result_handler: ErrorResultHandlerType = None,
) -> list[Any]: ) -> list[Any]:
self.logger.info( self.logger.info(
f"base-url:{self.baseurl},model-id:{self.model_id},concurrent:{self.max_concurrent},temperature:{self.temperature}" f"base-url:{self.baseurl},model-id:{self.model_id},concurrent:{self.max_concurrent},temperature:{self.temperature}"
@@ -691,7 +693,7 @@ class Agent:
f"预计发送{len(prompts)}个请求，并发请求数:{self.max_concurrent}" f"预计发送{len(prompts)}个请求，并发请求数:{self.max_concurrent}"
) )
self.total_error_counter.max_errors_count = ( self.total_error_counter.max_errors_count = (
len(prompts) // MAX_REQUESTS_PER_ERROR len(prompts) // MAX_REQUESTS_PER_ERROR
) )
# 新增：在每次批量发送前重置计数器 # 新增：在每次批量发送前重置计数器
@@ -710,9 +712,9 @@ class Agent:
max_connections=self.max_concurrent * 2, # 允许连接复用 max_connections=self.max_concurrent * 2, # 允许连接复用
max_keepalive_connections=self.max_concurrent, # 保持活跃连接 max_keepalive_connections=self.max_concurrent, # 保持活跃连接
) )
proxies = get_httpx_proxies() if USE_PROXY else None proxies = get_httpx_proxies() if self.system_proxy_enable else None
with httpx.Client( with httpx.Client(
trust_env=False, proxies=proxies, verify=False, limits=limits trust_env=False, proxies=proxies, verify=False, limits=limits
) as client: ) as client:
clients = itertools.repeat(client, len(prompts)) clients = itertools.repeat(client, len(prompts))
with ThreadPoolExecutor(max_workers=self.max_concurrent) as executor: with ThreadPoolExecutor(max_workers=self.max_concurrent) as executor:

3
docutranslate/translator/ai_translator/ass_translator.py
View File

@@ -38,7 +38,8 @@ class AssTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/docx_translator.py
View File

@@ -53,7 +53,8 @@ class DocxTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/epub_translator.py
View File

@@ -44,7 +44,8 @@ class EpubTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/html_translator.py
View File

@@ -101,7 +101,8 @@ class HtmlTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/json_translator.py
View File

@@ -34,7 +34,8 @@ class JsonTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.json_paths = config.json_paths self.json_paths = config.json_paths

3
docutranslate/translator/ai_translator/md_translator.py
View File

@@ -34,7 +34,8 @@ class MDTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry) retry=config.retry,
system_proxy_enable=config.system_proxy_enable)
self.translate_agent = MDTranslateAgent(agent_config) self.translate_agent = MDTranslateAgent(agent_config)
def translate(self, document: MarkdownDocument) -> Self: def translate(self, document: MarkdownDocument) -> Self:

3
docutranslate/translator/ai_translator/srt_translator.py
View File

@@ -40,7 +40,8 @@ class SrtTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/txt_translator.py
View File

@@ -58,7 +58,8 @@ class TXTTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode

3
docutranslate/translator/ai_translator/xlsx_translator.py
View File

@@ -42,7 +42,8 @@ class XlsxTranslator(AiTranslator):
timeout=config.timeout, timeout=config.timeout,
logger=self.logger, logger=self.logger,
glossary_dict=config.glossary_dict, glossary_dict=config.glossary_dict,
retry=config.retry retry=config.retry,
system_proxy_enable=config.system_proxy_enable
) )
self.translate_agent = SegmentsTranslateAgent(agent_config) self.translate_agent = SegmentsTranslateAgent(agent_config)
self.insert_mode = config.insert_mode self.insert_mode = config.insert_mode