Spaces:
Sleeping
Sleeping
Commit
·
01d0236
1
Parent(s):
7eccaaf
introduce googletrans w reqs mend qoogletrans
Browse files
app.py
CHANGED
|
@@ -30,7 +30,16 @@ class TranslationSystem:
|
|
| 30 |
"""
|
| 31 |
self.method = method
|
| 32 |
self.llm = llm
|
| 33 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 34 |
|
| 35 |
def translate_text(self, text, src='ru', dest='en'):
|
| 36 |
"""
|
|
@@ -44,51 +53,96 @@ class TranslationSystem:
|
|
| 44 |
Returns:
|
| 45 |
str: Translated text
|
| 46 |
"""
|
| 47 |
-
if pd.isna(text) or not text.strip():
|
| 48 |
return text
|
| 49 |
|
| 50 |
try:
|
| 51 |
-
if self.method == 'googletrans':
|
| 52 |
return self._translate_with_googletrans(text, src, dest)
|
| 53 |
else:
|
| 54 |
return self._translate_with_llm(text, src, dest)
|
| 55 |
except Exception as e:
|
| 56 |
-
st.warning(f"Translation error: {str(e)}")
|
| 57 |
return text
|
| 58 |
|
| 59 |
def _translate_with_googletrans(self, text, src='ru', dest='en'):
|
| 60 |
"""
|
| 61 |
-
Translate using googletrans library.
|
| 62 |
"""
|
| 63 |
try:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 64 |
# Add delay to avoid rate limits
|
| 65 |
time.sleep(0.5)
|
| 66 |
-
|
| 67 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 68 |
except Exception as e:
|
|
|
|
|
|
|
|
|
|
|
|
|
| 69 |
raise Exception(f"Googletrans error: {str(e)}")
|
| 70 |
|
| 71 |
def _translate_with_llm(self, text, src='ru', dest='en'):
|
| 72 |
"""
|
| 73 |
-
Translate using LangChain LLM.
|
| 74 |
"""
|
| 75 |
if not self.llm:
|
| 76 |
raise Exception("LLM not initialized for translation")
|
| 77 |
|
| 78 |
-
messages = [
|
| 79 |
-
{"role": "system", "content": "You are a translator. Translate the given Russian text to English accurately and concisely."},
|
| 80 |
-
{"role": "user", "content": f"Translate this Russian text to English: {text}"}
|
| 81 |
-
]
|
| 82 |
-
|
| 83 |
try:
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 84 |
response = self.llm.invoke(messages)
|
| 85 |
|
|
|
|
| 86 |
if hasattr(response, 'content'):
|
| 87 |
-
|
| 88 |
elif isinstance(response, str):
|
| 89 |
-
|
| 90 |
else:
|
| 91 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 92 |
except Exception as e:
|
| 93 |
raise Exception(f"LLM translation error: {str(e)}")
|
| 94 |
|
|
@@ -564,7 +618,7 @@ def create_output_file(df, uploaded_file, llm):
|
|
| 564 |
|
| 565 |
def main():
|
| 566 |
with st.sidebar:
|
| 567 |
-
st.title("::: AI-анализ мониторинга новостей (v.3.
|
| 568 |
st.subheader("по материалам СКАН-ИНТЕРФАКС ")
|
| 569 |
|
| 570 |
model_choice = st.radio(
|
|
|
|
| 30 |
"""
|
| 31 |
self.method = method
|
| 32 |
self.llm = llm
|
| 33 |
+
if method == 'googletrans':
|
| 34 |
+
try:
|
| 35 |
+
self.google_translator = GoogleTranslator()
|
| 36 |
+
# Test the translator with a simple string
|
| 37 |
+
self.google_translator.translate('test', src='en', dest='ru')
|
| 38 |
+
except Exception as e:
|
| 39 |
+
st.warning(f"Error initializing Google Translator: {str(e)}. Falling back to LLM translation.")
|
| 40 |
+
self.method = 'llm'
|
| 41 |
+
else:
|
| 42 |
+
self.google_translator = None
|
| 43 |
|
| 44 |
def translate_text(self, text, src='ru', dest='en'):
|
| 45 |
"""
|
|
|
|
| 53 |
Returns:
|
| 54 |
str: Translated text
|
| 55 |
"""
|
| 56 |
+
if pd.isna(text) or not isinstance(text, str) or not text.strip():
|
| 57 |
return text
|
| 58 |
|
| 59 |
try:
|
| 60 |
+
if self.method == 'googletrans' and self.google_translator:
|
| 61 |
return self._translate_with_googletrans(text, src, dest)
|
| 62 |
else:
|
| 63 |
return self._translate_with_llm(text, src, dest)
|
| 64 |
except Exception as e:
|
| 65 |
+
st.warning(f"Translation error: {str(e)}. Returning original text.")
|
| 66 |
return text
|
| 67 |
|
| 68 |
def _translate_with_googletrans(self, text, src='ru', dest='en'):
|
| 69 |
"""
|
| 70 |
+
Translate using googletrans library with improved error handling.
|
| 71 |
"""
|
| 72 |
try:
|
| 73 |
+
# Clean and validate input text
|
| 74 |
+
text = text.strip()
|
| 75 |
+
if not text:
|
| 76 |
+
return text
|
| 77 |
+
|
| 78 |
# Add delay to avoid rate limits
|
| 79 |
time.sleep(0.5)
|
| 80 |
+
|
| 81 |
+
# Attempt translation with retry logic
|
| 82 |
+
max_retries = 3
|
| 83 |
+
for attempt in range(max_retries):
|
| 84 |
+
try:
|
| 85 |
+
result = self.google_translator.translate(text, src=src, dest=dest)
|
| 86 |
+
if result and result.text:
|
| 87 |
+
return result.text
|
| 88 |
+
raise Exception("Empty translation result")
|
| 89 |
+
except Exception as e:
|
| 90 |
+
if attempt == max_retries - 1:
|
| 91 |
+
raise
|
| 92 |
+
time.sleep(1) # Wait before retry
|
| 93 |
+
|
| 94 |
+
raise Exception("All translation attempts failed")
|
| 95 |
+
|
| 96 |
except Exception as e:
|
| 97 |
+
# If googletrans fails, fall back to LLM translation
|
| 98 |
+
if self.llm:
|
| 99 |
+
st.warning(f"Googletrans error: {str(e)}. Falling back to LLM translation.")
|
| 100 |
+
return self._translate_with_llm(text, src, dest)
|
| 101 |
raise Exception(f"Googletrans error: {str(e)}")
|
| 102 |
|
| 103 |
def _translate_with_llm(self, text, src='ru', dest='en'):
|
| 104 |
"""
|
| 105 |
+
Translate using LangChain LLM with improved error handling.
|
| 106 |
"""
|
| 107 |
if not self.llm:
|
| 108 |
raise Exception("LLM not initialized for translation")
|
| 109 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 110 |
try:
|
| 111 |
+
# Clean input text
|
| 112 |
+
text = text.strip()
|
| 113 |
+
if not text:
|
| 114 |
+
return text
|
| 115 |
+
|
| 116 |
+
# Prepare system message based on language direction
|
| 117 |
+
if src == 'ru' and dest == 'en':
|
| 118 |
+
system_msg = "You are a translator. Translate the given Russian text to English accurately and concisely."
|
| 119 |
+
user_msg = f"Translate this Russian text to English: {text}"
|
| 120 |
+
elif src == 'en' and dest == 'ru':
|
| 121 |
+
system_msg = "You are a translator. Translate the given English text to Russian accurately and concisely."
|
| 122 |
+
user_msg = f"Translate this English text to Russian: {text}"
|
| 123 |
+
else:
|
| 124 |
+
raise Exception(f"Unsupported language pair: {src} to {dest}")
|
| 125 |
+
|
| 126 |
+
messages = [
|
| 127 |
+
{"role": "system", "content": system_msg},
|
| 128 |
+
{"role": "user", "content": user_msg}
|
| 129 |
+
]
|
| 130 |
+
|
| 131 |
response = self.llm.invoke(messages)
|
| 132 |
|
| 133 |
+
# Handle different response types
|
| 134 |
if hasattr(response, 'content'):
|
| 135 |
+
translation = response.content.strip()
|
| 136 |
elif isinstance(response, str):
|
| 137 |
+
translation = response.strip()
|
| 138 |
else:
|
| 139 |
+
translation = str(response).strip()
|
| 140 |
+
|
| 141 |
+
if not translation:
|
| 142 |
+
raise Exception("Empty translation result")
|
| 143 |
+
|
| 144 |
+
return translation
|
| 145 |
+
|
| 146 |
except Exception as e:
|
| 147 |
raise Exception(f"LLM translation error: {str(e)}")
|
| 148 |
|
|
|
|
| 618 |
|
| 619 |
def main():
|
| 620 |
with st.sidebar:
|
| 621 |
+
st.title("::: AI-анализ мониторинга новостей (v.3.33 ):::")
|
| 622 |
st.subheader("по материалам СКАН-ИНТЕРФАКС ")
|
| 623 |
|
| 624 |
model_choice = st.radio(
|