add baidu translator

deep_translator/__init__.py

@@ -2,6 +2,7 @@
 
 __copyright__ = "Copyright (C) 2020 Nidhal Baccouri"
 
+from deep_translator.baidu import BaiduTranslator
 from deep_translator.chatgpt import ChatGptTranslator
 from deep_translator.deepl import DeeplTranslator
 from deep_translator.detection import batch_detection, single_detection
@@ -31,6 +32,7 @@ __all__ = [
     "LibreTranslator",
     "PapagoTranslator",
     "ChatGptTranslator",
+    "BaiduTranslator",
     "single_detection",
     "batch_detection",
 ]

deep_translator/baidu.py

@@ -0,0 +1,127 @@
+"""
+baidu translator API
+"""
+
+__copyright__ = "Copyright (C) 2020 Nidhal Baccouri"
+
+import hashlib
+import os
+import random
+from typing import List, Optional
+
+import requests
+
+from deep_translator.base import BaseTranslator
+from deep_translator.constants import (
+    BAIDU_ID_VAR,
+    BAIDU_KEY_VAR,
+    BAIDU_LANGUAGE_TO_CODE,
+    BASE_URLS,
+)
+from deep_translator.exceptions import (
+    ApiKeyException,
+    ServerException,
+    TencentAPIerror,
+    TranslationNotFound,
+)
+from deep_translator.validate import is_empty, is_input_valid
+
+
+class BaiduTranslator(BaseTranslator):
+    """
+    class that wraps functions, which use the TentCentTranslator translator
+    under the hood to translate word(s)
+    """
+
+    def __init__(
+        self,
+        source: str = "en",
+        target: str = "zh",
+        secret_id: Optional[str] = os.getenv(BAIDU_ID_VAR, None),
+        secret_key: Optional[str] = os.getenv(BAIDU_KEY_VAR, None),
+        **kwargs
+    ):
+        """
+        @param secret_id: your baidu cloud api secret id.
+        Get one here: https://fanyi-api.baidu.com/choose
+        @param secret_key: your baidu cloud api secret key.
+        @param source: source language
+        @param target: target language
+        """
+        if not secret_id:
+            raise ApiKeyException(env_var=BAIDU_ID_VAR)
+
+        if not secret_key:
+            raise ApiKeyException(env_var=BAIDU_KEY_VAR)
+
+        self.secret_id = secret_id
+        self.secret_key = secret_key
+        super().__init__(
+            base_url=BASE_URLS.get("BAIDU"),
+            source=source,
+            target=target,
+            languages=BAIDU_LANGUAGE_TO_CODE,
+            **kwargs
+        )
+
+    def translate(self, text: str, **kwargs) -> str:
+        """
+        @param text: text to translate
+        @return: translated text
+        """
+        if is_input_valid(text):
+            if self._same_source_target() or is_empty(text):
+                return text
+
+            # Create the request parameters.
+            salt = random.randint(32768, 65536)
+            sign = hashlib.md5(
+                (self.secret_id + text + str(salt) + self.secret_key).encode(
+                    "utf-8"
+                )
+            ).hexdigest()
+            headers = {"Content-Type": "application/x-www-form-urlencoded"}
+            payload = {
+                "appid": self.secret_id,
+                "q": text,
+                "from": self.source,
+                "to": self.target,
+                "salt": salt,
+                "sign": sign,
+            }
+
+            # Do the request and check the connection.
+            try:
+                response = requests.post(
+                    self._base_url, params=payload, headers=headers
+                )
+            except ConnectionError:
+                raise ServerException(503)
+            if response.status_code != 200:
+                raise ServerException(response.status_code)
+            # Get the response and check is not empty.
+            res = response.json()
+            if not res:
+                raise TranslationNotFound(text)
+            # Process and return the response.
+            if "error_code" in res:
+                raise TencentAPIerror(res["error_msg"])
+            return res["trans_result"]["dst"]
+
+    def translate_file(self, path: str, **kwargs) -> str:
+        return self._translate_file(path, **kwargs)
+
+    def translate_batch(self, batch: List[str], **kwargs) -> List[str]:
+        """
+        @param batch: list of texts to translate
+        @return: list of translations
+        """
+        return self._translate_batch(batch, **kwargs)
+
+
+if __name__ == "__main__":
+    d = BaiduTranslator(
+        target="zh", secret_id="some-id", secret_key="some-key"
+    )
+    t = d.translate("Ich habe keine ahnung")
+    print("text: ", t)

deep_translator/constants.py

@@ -7,6 +7,8 @@ LIBRE_ENV_VAR = "LIBRE_API_KEY"
 MSFT_ENV_VAR = "MICROSOFT_API_KEY"
 QCRI_ENV_VAR = "QCRI_API_KEY"
 YANDEX_ENV_VAR = "YANDEX_API_KEY"
+BAIDU_ID_VAR = "BAIDU_ID"
+BAIDU_KEY_VAR = "BAIDU_KEY"
 
 
 BASE_URLS = {
@@ -23,6 +25,7 @@ BASE_URLS = {
     "PAPAGO_API": "https://openapi.naver.com/v1/papago/n2mt",
     "LIBRE": "https://libretranslate.com/",
     "LIBRE_FREE": "https://libretranslate.de/",
+    "BAIDU": "https://fanyi-api.baidu.com/api/trans/vip/translate",
 }
 
 GOOGLE_LANGUAGES_TO_CODES = {
@@ -280,3 +283,23 @@ LIBRE_LANGUAGES_TO_CODES = {
     "Turkish": "tr",
     "Vietnamese": "vi",
 }
+
+BAIDU_LANGUAGE_TO_CODE = {
+    "arabic": "ar",
+    "chinese (simplified)": "zh",
+    "chinese (traditional)": "zh-TW",
+    "english": "en",
+    "french": "fr",
+    "german": "de",
+    "hindi": "hi",
+    "indonesian": "id",
+    "japanese": "ja",
+    "korean": "ko",
+    "malay": "ms",
+    "portuguese": "pt",
+    "russian": "ru",
+    "spanish": "es",
+    "thai": "th",
+    "turkish": "tr",
+    "vietnamese": "vi",
+}

deep_translator/exceptions.py

@@ -180,3 +180,16 @@ class AuthorizationException(Exception):
     def __init__(self, api_key, *args):
         msg = "Unauthorized access with the api key " + api_key
         super().__init__(msg, *args)
+
+
+class BaiduAPIerror(Exception):
+    """
+    exception thrown if Baidu API returns one of its errors
+    """
+
+    def __init__(self, api_message):
+        self.api_message = str(api_message)
+        self.message = "Baidu API returned the following error"
+
+    def __str__(self):
+        return "{}: {}".format(self.message, self.api_message)

tests/test_baidu.py

@@ -0,0 +1,37 @@
+from unittest.mock import Mock, patch
+
+import pytest
+
+from deep_translator import BaiduTranslator
+from deep_translator.exceptions import AuthorizationException
+
+
+@patch("deep_translator.baidu.requests")
+def test_simple_translation(mock_requests):
+    translator = BaiduTranslator(
+        secret_id="this-is-an-valid-api-id",
+        source="en",
+        target="zh",
+        secret_key="this-is-an-valid-api-key",
+    )
+    # Set the request response mock.
+    mock_response = Mock()
+    mock_response.status_code = 200
+    mock_response.json.return_value = {"translations": [{"text": "hola"}]}
+    mock_requests.get.return_value = mock_response
+    translation = translator.translate("hello")
+    assert translation == "你好"
+
+
+@patch("deep_translator.tencent.requests.get")
+def test_wrong_api_key(mock_requests):
+    translator = BaiduTranslator(
+        secret_id="this-is-a-wrong-api-id",
+        source="en",
+        target="zh",
+        secret_key="this-is-a-wrong-api-key",
+    )
+    # Set the response status_code only.
+    mock_requests.return_value = Mock(status_code=403)
+    with pytest.raises(AuthorizationException):
+        translator.translate("Hello")