Get help with the detect

Satyam Rai 0

Hi all, I am currently trying to translate Turkish text into English and back and forth. The translation itself works fine, but the input language detection is not so good. If suppose, I pass mixed text that contains both Turkish and English text, then it detects the whole as English and does not translate it. How can I fix this issue?

Saideep Anchuri 1,190 Microsoft Vendor

Hi Satyam Rai

Welcome to Microsoft Q&A Forum, thank you for posting your query here!

If your input text contains both Turkish and English, the language detection feature may come handy identifying the current dominant language context. Once the paragraph has been split to sentences or words (if you want accuracy on word level), you can use if else logic to translate the English words and retain the Turkish words without translating to Turkish again if predominant language is English. Similar way, you can translate if predominant language is Turkish. detection-results

from azure.core.credentials import AzureKeyCredential
from azure.ai.translation.text import TextTranslationClient
from azure.ai.textanalytics import TextAnalyticsClient
from azure.ai.textanalytics import DetectLanguageInput
import os
 
# Replace with your own subscription key and endpoint
subscription_key = "YOUR_SUBSCRIPTION_KEY"
endpoint_translate = "https://YOUR_TRANSLATE_RESOURCE_NAME.cognitiveservices.azure.com/"
endpoint_language = "https://YOUR_LANGUAGE_RESOURCE_NAME.cognitiveservices.azure.com/"
 
# Initialize the clients
translator_client = TextTranslationClient(endpoint_translate, AzureKeyCredential(subscription_key))
language_client = TextAnalyticsClient(endpoint_language, AzureKeyCredential(subscription_key))
 
def detect_language(text):
    response = language_client.detect_language(documents=[DetectLanguageInput(id="1", text=text)])
    return response[0].primary_language.iso6391_name
 
def translate_mixed_text(text):
    translated_text = []
 
    # Split the text into words
    words = text.split()
 
    for word in words:
        try:
            # Detect the language of the word
            language = detect_language(word)
            if language == 'en':
                # Translate English word to Turkish
                response = translator_client.translate(content=[word], to=["tr"])
                translated_word = response[0].translations[0].text
                translated_text.append(translated_word)
            else:
                # Retain Turkish word
                translated_text.append(word)
        except Exception as e:
            print(f"Processing failed for word '{word}': {e}")
            translated_text.append(word)
 
    return ' '.join(translated_text)
 
text = "This is English. Bu Türkçedir."
translated_result = translate_mixed_text(text)
print(translated_result)

Kindly refer below documentation: detect-language

Thank You.

Saideep Anchuri 1,190 Reputation points Microsoft Vendor

2025-01-22T00:32:54.0233333+00:00

Hi Satyam Rai

Following up to see if the given response was helpful.

Thank You.

Share via

Get help with the detect

Your answer