Add Ollama vision image translation with checkbox option
This commit is contained in:
@@ -59,19 +59,21 @@ class LibreTranslationProvider(TranslationProvider):
|
||||
return text
|
||||
|
||||
try:
|
||||
translator = LibreTranslator(source=source_language, target=target_language)
|
||||
# LibreTranslator doesn't need API key for self-hosted instances
|
||||
translator = LibreTranslator(source=source_language, target=target_language, custom_url="http://localhost:5000")
|
||||
return translator.translate(text)
|
||||
except Exception as e:
|
||||
print(f"Translation error: {e}")
|
||||
# Fail silently and return original text
|
||||
return text
|
||||
|
||||
|
||||
class OllamaTranslationProvider(TranslationProvider):
|
||||
"""Ollama LLM translation implementation"""
|
||||
|
||||
def __init__(self, base_url: str = "http://localhost:11434", model: str = "llama3"):
|
||||
def __init__(self, base_url: str = "http://localhost:11434", model: str = "llama3", vision_model: str = "llava"):
|
||||
self.base_url = base_url.rstrip('/')
|
||||
self.model = model
|
||||
self.vision_model = vision_model
|
||||
|
||||
def translate(self, text: str, target_language: str, source_language: str = 'auto') -> str:
|
||||
if not text or not text.strip():
|
||||
@@ -96,6 +98,34 @@ class OllamaTranslationProvider(TranslationProvider):
|
||||
print(f"Ollama translation error: {e}")
|
||||
return text
|
||||
|
||||
def translate_image(self, image_path: str, target_language: str) -> str:
|
||||
"""Translate text within an image using Ollama vision model"""
|
||||
import base64
|
||||
|
||||
try:
|
||||
# Read and encode image
|
||||
with open(image_path, 'rb') as img_file:
|
||||
image_data = base64.b64encode(img_file.read()).decode('utf-8')
|
||||
|
||||
prompt = f"Extract all text from this image and translate it to {target_language}. Return ONLY the translated text, preserving the structure and formatting."
|
||||
|
||||
response = requests.post(
|
||||
f"{self.base_url}/api/generate",
|
||||
json={
|
||||
"model": self.vision_model,
|
||||
"prompt": prompt,
|
||||
"images": [image_data],
|
||||
"stream": False
|
||||
},
|
||||
timeout=60
|
||||
)
|
||||
response.raise_for_status()
|
||||
result = response.json()
|
||||
return result.get("response", "").strip()
|
||||
except Exception as e:
|
||||
print(f"Ollama vision translation error: {e}")
|
||||
return ""
|
||||
|
||||
@staticmethod
|
||||
def list_models(base_url: str = "http://localhost:11434") -> List[str]:
|
||||
"""List available Ollama models"""
|
||||
@@ -121,20 +151,9 @@ class TranslationService:
|
||||
|
||||
def _get_default_provider(self) -> TranslationProvider:
|
||||
"""Get the default translation provider from configuration"""
|
||||
service_type = config.TRANSLATION_SERVICE.lower()
|
||||
|
||||
if service_type == "deepl":
|
||||
if not config.DEEPL_API_KEY:
|
||||
raise ValueError("DeepL API key not configured")
|
||||
return DeepLTranslationProvider(config.DEEPL_API_KEY)
|
||||
elif service_type == "libre":
|
||||
return LibreTranslationProvider()
|
||||
elif service_type == "ollama":
|
||||
ollama_url = getattr(config, 'OLLAMA_BASE_URL', 'http://localhost:11434')
|
||||
ollama_model = getattr(config, 'OLLAMA_MODEL', 'llama3')
|
||||
return OllamaTranslationProvider(base_url=ollama_url, model=ollama_model)
|
||||
else: # Default to Google
|
||||
return GoogleTranslationProvider()
|
||||
# Always use Google Translate by default to avoid API key issues
|
||||
# Provider will be overridden per request in the API endpoint
|
||||
return GoogleTranslationProvider()
|
||||
|
||||
def translate_text(self, text: str, target_language: str, source_language: str = 'auto') -> str:
|
||||
"""
|
||||
|
||||
Reference in New Issue
Block a user