Merge pull request #578 from A0nameless0man/main

feat: 增加 Sakura 模型支持

README.md

@@ -389,7 +389,7 @@ Colorizer: **mc2**
 --upscale-ratio UPSCALE_RATIO                Image upscale ratio applied before detection. Can
                                              improve text detection.
 --colorizer {mc2}                            Colorization model to use.
---translator {google,youdao,baidu,deepl,papago,caiyun,gpt3,gpt3.5,gpt4,none,original,offline,nllb,nllb_big,sugoi,jparacrawl,jparacrawl_big,m2m100,m2m100_big}
+--translator {google,youdao,baidu,deepl,papago,caiyun,gpt3,gpt3.5,gpt4,none,original,offline,nllb,nllb_big,sugoi,jparacrawl,jparacrawl_big,m2m100,m2m100_big,sakura}
                                              Language translator to use
 --translator-chain TRANSLATOR_CHAIN          Output of one translator goes in another. Example:
                                              --translator-chain "google:JPN;sugoi:ENG".

README_CN.md

@@ -140,7 +140,7 @@ IND: Indonesian
 --upscale-ratio UPSCALE_RATIO                Image upscale ratio applied before detection. Can
                                              improve text detection.
 --colorizer {mc2}                            Colorization model to use.
---translator {google,youdao,baidu,deepl,papago,caiyun,gpt3,gpt3.5,gpt4,none,original,offline,nllb,nllb_big,sugoi,jparacrawl,jparacrawl_big,m2m100,m2m100_big}
+--translator {google,youdao,baidu,deepl,papago,caiyun,gpt3,gpt3.5,gpt4,none,original,offline,nllb,nllb_big,sugoi,jparacrawl,jparacrawl_big,m2m100,sakura}
                                              Language translator to use
 --translator-chain TRANSLATOR_CHAIN          Output of one translator goes in another. Example:
                                              --translator-chain "google:JPN;sugoi:ENG".

manga_translator/server/web_main.py

@@ -58,6 +58,7 @@
     'jparacrawl_big',
     'm2m100',
     'm2m100_big',
+    'sakura',
     'none',
     'original',
 ]

manga_translator/translators/__init__.py

@@ -7,7 +7,7 @@
 from .deepl import DeeplTranslator
 from .papago import PapagoTranslator
 from .caiyun import CaiyunTranslator
-from .chatgpt import GPT3Translator, GPT35TurboTranslator, GPT4Translator
+from .chatgpt import GPT3Translator, GPT35TurboTranslator, GPT4Translator, SakuraTranslator
 from .nllb import NLLBTranslator, NLLBBigTranslator
 from .sugoi import JparacrawlTranslator, JparacrawlBigTranslator, SugoiTranslator
 from .m2m100 import M2M100Translator, M2M100BigTranslator
@@ -24,6 +24,7 @@
     'jparacrawl_big': JparacrawlBigTranslator,
     'm2m100': M2M100Translator,
     'm2m100_big': M2M100BigTranslator,
+    'sakura': SakuraTranslator,
 }
 
 TRANSLATORS = {

manga_translator/translators/chatgpt.py

@@ -9,7 +9,7 @@
 from typing import List, Dict
 
 from .common import CommonTranslator, MissingAPIKeyException
-from .keys import OPENAI_API_KEY, OPENAI_HTTP_PROXY, OPENAI_API_BASE
+from .keys import OPENAI_API_KEY, OPENAI_HTTP_PROXY, OPENAI_API_BASE, SAKURA_API_BASE, SAKURA_API_KEY
 
 CONFIG = None
 
@@ -54,11 +54,11 @@ class GPT3Translator(CommonTranslator):
     _INCLUDE_TEMPLATE = True
     _PROMPT_TEMPLATE = 'Please help me to translate the following text from a manga to {to_lang} (if it\'s already in {to_lang} or looks like gibberish you have to output it as it is instead):\n'
 
-    def __init__(self):
+    def __init__(self, check_openai_key = True):
         super().__init__()
         openai.api_key = openai.api_key or OPENAI_API_KEY
         openai.api_base = OPENAI_API_BASE
-        if not openai.api_key:
+        if not openai.api_key and check_openai_key:
             raise MissingAPIKeyException('Please set the OPENAI_API_KEY environment variable before using the chatgpt translator.')
         if OPENAI_HTTP_PROXY:
             proxies = {
@@ -320,3 +320,77 @@ async def _request_translation(self, to_lang: str, prompt: str) -> str:
 
         # If no response with text is found, return the first response's content (which may be empty)
         return response.choices[0].message.content
+
+
+class SakuraTranslator(GPT3Translator):
+    _CONFIG_KEY = 'sakura'
+    _MAX_REQUESTS_PER_MINUTE = 200
+    _RETRY_ATTEMPTS = 5
+    _MAX_TOKENS = 8192
+    _CHAT_SYSTEM_TEMPLATE = (
+        '你是一个轻小说翻译模型，可以流畅通顺地以日本轻小说的风格将日文翻译成简体中文，并联系上下文正确使用人称代词，不擅自添加原文中没有的代词。'
+    )
+    def __init__(self):
+        super().__init__(check_openai_key=False)
+
+    async def _request_translation(self, to_lang: str, prompt: str) -> str:
+        messages = [
+            {'role': 'system', 'content': self._CHAT_SYSTEM_TEMPLATE},
+            {'role': 'user', 'content': '将下面的日文文本翻译成中文：'+prompt},
+        ]
+
+        response = await openai.ChatCompletion.acreate(
+            model='gpt-4-0613',
+            messages=messages,
+            max_tokens=self._MAX_TOKENS // 2,
+            temperature=self.temperature,
+            top_p=self.top_p,
+            api_key=SAKURA_API_KEY,
+            api_base=SAKURA_API_BASE,
+        )
+
+        self.token_count += response.usage['total_tokens']
+        self.token_count_last = response.usage['total_tokens']
+        for choice in response.choices:
+            if 'text' in choice:
+                return choice.text
+
+        # If no response with text is found, return the first response's content (which may be empty)
+        return response.choices[0].message.content
+    async def _translate(self, from_lang: str, to_lang: str, queries: List[str]) -> List[str]:
+        translations = []
+        self.logger.debug(f'Temperature: {self.temperature}, TopP: {self.top_p}')
+
+        for query in queries:
+
+            ratelimit_attempt = 0
+            server_error_attempt = 0
+            timeout_attempt = 0
+            while True:
+                request_task = asyncio.create_task(self._request_translation(to_lang, query))
+                try:
+                    response = await request_task
+                    break
+                except openai.error.RateLimitError: # Server returned ratelimit response
+                    ratelimit_attempt += 1
+                    if ratelimit_attempt >= self._RATELIMIT_RETRY_ATTEMPTS:
+                        raise
+                    self.logger.warn(f'Restarting request due to ratelimiting by openai servers. Attempt: {ratelimit_attempt}')
+                    await asyncio.sleep(2)
+                except openai.error.APIError: # Server returned 500 error (probably server load)
+                    server_error_attempt += 1
+                    if server_error_attempt >= self._RETRY_ATTEMPTS:
+                        self.logger.error('Sakura encountered a server error, possibly due to high server load. Use a different translator or try again later.')
+                        raise
+                    self.logger.warn(f'Restarting request due to a server error. Attempt: {server_error_attempt}')
+                    await asyncio.sleep(1)
+
+            self.logger.debug('-- Sakura Response --\n' + response)
+
+            translations.extend([response])
+
+        self.logger.debug(translations)
+        if self.token_count_last:
+            self.logger.info(f'Used {self.token_count_last} tokens (Total: {self.token_count})')
+
+        return translations

manga_translator/translators/keys.py

@@ -17,3 +17,6 @@
 OPENAI_API_BASE = os.getenv('OPENAI_API_BASE', 'https://api.openai.com/v1') #使用api-for-open-llm例子 http://127.0.0.1:8000/v1
 
 CAIYUN_TOKEN = os.getenv('CAIYUN_TOKEN', '') # 彩云小译API访问令牌
+
+SAKURA_API_KEY = os.getenv('SAKURA_API_KEY', '')
+SAKURA_API_BASE = os.getenv('SAKURA_API_BASE', 'http://127.0.0.1:5000/v1')