Spaces:
Running
Running
Update translator.py
Browse files- translator.py +92 -53
translator.py
CHANGED
|
@@ -1,76 +1,83 @@
|
|
| 1 |
-
"""
|
| 2 |
-
translator.py
|
| 3 |
-
"""
|
| 4 |
import hashlib, hmac, json, os, random, time
|
| 5 |
from datetime import datetime
|
| 6 |
-
from typing import List, Sequence, Optional
|
| 7 |
|
| 8 |
import requests
|
| 9 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 10 |
TENCENT_TRANSLATE_URL = os.environ.get("TENCENT_TRANSLATE_URL", "https://tmt.tencentcloudapi.com")
|
| 11 |
-
BAIDU_TRANSLATE_URL = os.environ.get("BAIDU_TRANSLATE_URL", "https://fanyi-api.baidu.com/api/trans/vip/translate")
|
| 12 |
|
| 13 |
-
|
| 14 |
-
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
_baidu_idx = (_baidu_idx + 1) % len(creds_list)
|
| 19 |
-
return cred
|
| 20 |
|
| 21 |
def _sign(key: bytes, msg: str) -> bytes:
|
| 22 |
return hmac.new(key, msg.encode("utf-8"), hashlib.sha256).digest()
|
| 23 |
|
| 24 |
def _tc3_signature(secret_key: str, date: str, service: str, string_to_sign: str) -> str:
|
| 25 |
-
secret_date
|
| 26 |
-
secret_service
|
| 27 |
-
secret_signing
|
| 28 |
return hmac.new(secret_signing, string_to_sign.encode("utf-8"), hashlib.sha256).hexdigest()
|
| 29 |
|
| 30 |
-
def _translate_with_tencent(texts: Sequence[str], src: str, tgt: str,
|
| 31 |
-
|
| 32 |
-
secret_key: Optional[str]) -> Optional[List[str]]:
|
| 33 |
-
"""使用传入的密钥执行腾讯云翻译。如果密钥无效则失败。"""
|
| 34 |
-
if not (secret_id and secret_key):
|
| 35 |
-
return None
|
| 36 |
|
| 37 |
service, host, action, version, region = "tmt", "tmt.tencentcloudapi.com", "TextTranslate", "2018-03-21", "ap-beijing"
|
| 38 |
-
ts
|
| 39 |
-
|
| 40 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 41 |
|
| 42 |
-
canonical_request = f"POST\n/\n\ncontent-type:application/json; charset=utf-8\nhost:{host}\nx-tc-action:{action.lower()}\n\ncontent-type;host;x-tc-action\n{hashlib.sha256(payload_str.encode()).hexdigest()}"
|
| 43 |
credential_scope = f"{date}/{service}/tc3_request"
|
| 44 |
-
string_to_sign =
|
|
|
|
|
|
|
| 45 |
signature = _tc3_signature(secret_key, date, service, string_to_sign)
|
| 46 |
|
| 47 |
-
authorization = f"
|
| 48 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
|
| 50 |
try:
|
| 51 |
resp = requests.post(TENCENT_TRANSLATE_URL, headers=headers, data=payload_str, timeout=8)
|
| 52 |
resp.raise_for_status()
|
| 53 |
data = resp.json()
|
| 54 |
-
if "
|
| 55 |
-
|
|
|
|
|
|
|
| 56 |
return None
|
| 57 |
-
return data.get("Response", {}).get("TargetText", "").split("\n")
|
| 58 |
except Exception as e:
|
| 59 |
-
print(f"[translator] Tencent
|
| 60 |
return None
|
| 61 |
|
| 62 |
-
def _translate_with_baidu(texts: Sequence[str], src: str, tgt: str,
|
| 63 |
-
|
| 64 |
-
if not baidu_credentials_list:
|
| 65 |
-
return None
|
| 66 |
|
| 67 |
-
|
| 68 |
-
|
| 69 |
-
print("[translator] Baidu credentials format error.")
|
| 70 |
-
return None
|
| 71 |
|
| 72 |
-
app_id
|
| 73 |
-
|
|
|
|
|
|
|
| 74 |
sign = hashlib.md5((app_id + query + str(salt) + secret_key).encode()).hexdigest()
|
| 75 |
params = {"q": query, "from": src, "to": tgt, "appid": app_id, "salt": salt, "sign": sign}
|
| 76 |
|
|
@@ -78,25 +85,57 @@ def _translate_with_baidu(texts: Sequence[str], src: str, tgt: str,
|
|
| 78 |
resp = requests.get(BAIDU_TRANSLATE_URL, params=params, timeout=8)
|
| 79 |
resp.raise_for_status()
|
| 80 |
data = resp.json()
|
| 81 |
-
if "
|
| 82 |
-
|
|
|
|
|
|
|
| 83 |
return None
|
| 84 |
-
return [item["dst"] for item in data.get("trans_result", [])]
|
| 85 |
except Exception as e:
|
| 86 |
-
print(f"[translator] Baidu
|
| 87 |
return None
|
| 88 |
|
|
|
|
| 89 |
def translate_texts(texts: Sequence[str],
|
| 90 |
src_lang: str = "auto",
|
| 91 |
tgt_lang: str = "zh",
|
| 92 |
-
|
| 93 |
-
|
| 94 |
-
|
|
|
|
| 95 |
if not texts:
|
| 96 |
return []
|
| 97 |
|
| 98 |
-
|
| 99 |
-
|
| 100 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 101 |
|
| 102 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
import hashlib, hmac, json, os, random, time
|
| 2 |
from datetime import datetime
|
| 3 |
+
from typing import List, Sequence, Optional, Dict, Any
|
| 4 |
|
| 5 |
import requests
|
| 6 |
|
| 7 |
+
TRANSLATOR_ACCESS_KEY = os.environ.get("TRANSLATOR_ACCESS_KEY")
|
| 8 |
+
|
| 9 |
+
TENCENT_SECRET_ID = os.environ.get("TENCENT_SECRET_ID")
|
| 10 |
+
TENCENT_SECRET_KEY = os.environ.get("TENCENT_SECRET_KEY")
|
| 11 |
TENCENT_TRANSLATE_URL = os.environ.get("TENCENT_TRANSLATE_URL", "https://tmt.tencentcloudapi.com")
|
|
|
|
| 12 |
|
| 13 |
+
BAIDU_TRANSLATE_URL = os.environ.get("BAIDU_TRANSLATE_URL", "https://fanyi-api.baidu.com/api/trans/vip/translate")
|
| 14 |
+
try:
|
| 15 |
+
BAIDU_CREDENTIALS_DEFAULT = json.loads(os.environ.get("BAIDU_CREDENTIALS_JSON", "[]"))
|
| 16 |
+
except json.JSONDecodeError:
|
| 17 |
+
BAIDU_CREDENTIALS_DEFAULT = []
|
|
|
|
|
|
|
| 18 |
|
| 19 |
def _sign(key: bytes, msg: str) -> bytes:
|
| 20 |
return hmac.new(key, msg.encode("utf-8"), hashlib.sha256).digest()
|
| 21 |
|
| 22 |
def _tc3_signature(secret_key: str, date: str, service: str, string_to_sign: str) -> str:
|
| 23 |
+
secret_date = _sign(("TC3" + secret_key).encode(), date)
|
| 24 |
+
secret_service = _sign(secret_date, service)
|
| 25 |
+
secret_signing = _sign(secret_service, "tc3_request")
|
| 26 |
return hmac.new(secret_signing, string_to_sign.encode("utf-8"), hashlib.sha256).hexdigest()
|
| 27 |
|
| 28 |
+
def _translate_with_tencent(texts: Sequence[str], src: str, tgt: str, secret_id: str, secret_key: str) -> Optional[List[str]]:
|
| 29 |
+
if not (secret_id and secret_key): return None
|
|
|
|
|
|
|
|
|
|
|
|
|
| 30 |
|
| 31 |
service, host, action, version, region = "tmt", "tmt.tencentcloudapi.com", "TextTranslate", "2018-03-21", "ap-beijing"
|
| 32 |
+
ts = int(time.time())
|
| 33 |
+
date = datetime.utcfromtimestamp(ts).strftime("%Y-%m-%d")
|
| 34 |
+
|
| 35 |
+
payload = {"SourceText": "\n".join(texts), "Source": src, "Target": tgt, "ProjectId": 0}
|
| 36 |
+
payload_str = json.dumps(payload, ensure_ascii=False)
|
| 37 |
+
|
| 38 |
+
canonical_request = "POST\n/\n\n" + \
|
| 39 |
+
f"content-type:application/json; charset=utf-8\nhost:{host}\nx-tc-action:{action.lower()}\n\n" + \
|
| 40 |
+
"content-type;host;x-tc-action\n" + \
|
| 41 |
+
hashlib.sha256(payload_str.encode()).hexdigest()
|
| 42 |
|
|
|
|
| 43 |
credential_scope = f"{date}/{service}/tc3_request"
|
| 44 |
+
string_to_sign = "TC3-HMAC-SHA256\n" + str(ts) + "\n" + credential_scope + "\n" + \
|
| 45 |
+
hashlib.sha256(canonical_request.encode()).hexdigest()
|
| 46 |
+
|
| 47 |
signature = _tc3_signature(secret_key, date, service, string_to_sign)
|
| 48 |
|
| 49 |
+
authorization = f"TC3-HMAC-SHA256 Credential={secret_id}/{credential_scope}, " + \
|
| 50 |
+
f"SignedHeaders=content-type;host;x-tc-action, Signature={signature}"
|
| 51 |
+
|
| 52 |
+
headers = {
|
| 53 |
+
"Authorization": authorization, "Content-Type": "application/json; charset=utf-8",
|
| 54 |
+
"Host": host, "X-TC-Action": action, "X-TC-Timestamp": str(ts),
|
| 55 |
+
"X-TC-Version": version, "X-TC-Region": region,
|
| 56 |
+
}
|
| 57 |
|
| 58 |
try:
|
| 59 |
resp = requests.post(TENCENT_TRANSLATE_URL, headers=headers, data=payload_str, timeout=8)
|
| 60 |
resp.raise_for_status()
|
| 61 |
data = resp.json()
|
| 62 |
+
if "Response" in data and "TargetText" in data["Response"]:
|
| 63 |
+
return data["Response"]["TargetText"].split("\n")
|
| 64 |
+
else:
|
| 65 |
+
print(f"[translator] Tencent API abnormal response: {data}")
|
| 66 |
return None
|
|
|
|
| 67 |
except Exception as e:
|
| 68 |
+
print(f"[translator] Tencent API error → {e}")
|
| 69 |
return None
|
| 70 |
|
| 71 |
+
def _translate_with_baidu(texts: Sequence[str], src: str, tgt: str, baidu_credentials: List[Dict[str, str]]) -> Optional[List[str]]:
|
| 72 |
+
if not baidu_credentials: return None
|
|
|
|
|
|
|
| 73 |
|
| 74 |
+
cred = random.choice(baidu_credentials)
|
| 75 |
+
app_id, secret_key = cred.get("app_id"), cred.get("secret_key")
|
|
|
|
|
|
|
| 76 |
|
| 77 |
+
if not (app_id and secret_key): return None
|
| 78 |
+
|
| 79 |
+
salt = random.randint(32768, 65536)
|
| 80 |
+
query = "\n".join(texts)
|
| 81 |
sign = hashlib.md5((app_id + query + str(salt) + secret_key).encode()).hexdigest()
|
| 82 |
params = {"q": query, "from": src, "to": tgt, "appid": app_id, "salt": salt, "sign": sign}
|
| 83 |
|
|
|
|
| 85 |
resp = requests.get(BAIDU_TRANSLATE_URL, params=params, timeout=8)
|
| 86 |
resp.raise_for_status()
|
| 87 |
data = resp.json()
|
| 88 |
+
if "trans_result" in data:
|
| 89 |
+
return [item["dst"] for item in data["trans_result"]]
|
| 90 |
+
else:
|
| 91 |
+
print(f"[translator] Baidu API abnormal response: {data}")
|
| 92 |
return None
|
|
|
|
| 93 |
except Exception as e:
|
| 94 |
+
print(f"[translator] Baidu API error → {e}")
|
| 95 |
return None
|
| 96 |
|
| 97 |
+
|
| 98 |
def translate_texts(texts: Sequence[str],
|
| 99 |
src_lang: str = "auto",
|
| 100 |
tgt_lang: str = "zh",
|
| 101 |
+
system_key_input: Optional[str] = None,
|
| 102 |
+
tencent_id: Optional[str] = None,
|
| 103 |
+
tencent_key: Optional[str] = None,
|
| 104 |
+
baidu_creds_json_str: Optional[str] = None) -> List[str]:
|
| 105 |
if not texts:
|
| 106 |
return []
|
| 107 |
|
| 108 |
+
use_tencent_id, use_tencent_key = None, None
|
| 109 |
+
use_baidu_creds = []
|
| 110 |
+
|
| 111 |
+
if tencent_id and tencent_key:
|
| 112 |
+
use_tencent_id, use_tencent_key = tencent_id, tencent_key
|
| 113 |
+
print("[translator] Using custom Tencent API key.")
|
| 114 |
+
if baidu_creds_json_str:
|
| 115 |
+
try:
|
| 116 |
+
creds = json.loads(baidu_creds_json_str)
|
| 117 |
+
if isinstance(creds, list) and all(isinstance(d, dict) for d in creds):
|
| 118 |
+
use_baidu_creds = creds
|
| 119 |
+
print("[translator] Using custom Baidu API key(s).")
|
| 120 |
+
else:
|
| 121 |
+
print("[translator] Warning: Custom Baidu credentials format is incorrect.")
|
| 122 |
+
except json.JSONDecodeError:
|
| 123 |
+
print("[translator] Warning: Failed to parse custom Baidu credentials JSON.")
|
| 124 |
+
|
| 125 |
+
elif TRANSLATOR_ACCESS_KEY and system_key_input == TRANSLATOR_ACCESS_KEY:
|
| 126 |
+
print("[translator] System access key validated. Using system-configured API keys.")
|
| 127 |
+
use_tencent_id, use_tencent_key = TENCENT_SECRET_ID, TENCENT_SECRET_KEY
|
| 128 |
+
use_baidu_creds = BAIDU_CREDENTIALS_DEFAULT
|
| 129 |
+
|
| 130 |
+
else:
|
| 131 |
+
print("[translator] Translation disabled: No valid API keys or system key provided.")
|
| 132 |
+
return list(texts)
|
| 133 |
+
|
| 134 |
+
translated_texts = None
|
| 135 |
+
if use_tencent_id and use_tencent_key:
|
| 136 |
+
translated_texts = _translate_with_tencent(texts, src_lang, tgt_lang, use_tencent_id, use_tencent_key)
|
| 137 |
|
| 138 |
+
if translated_texts is None and use_baidu_creds:
|
| 139 |
+
translated_texts = _translate_with_baidu(texts, src_lang, tgt_lang, use_baidu_creds)
|
| 140 |
+
|
| 141 |
+
return translated_texts or list(texts)
|