Spaces:
Running
Running
Create translator.py
Browse files- translator.py +160 -0
translator.py
ADDED
@@ -0,0 +1,160 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
"""
|
2 |
+
translator.py
|
3 |
+
腾讯云 + 百度翻译 API 轮询封装
|
4 |
+
⚠️ 需在 HF 空间的 “Variables” 页设置以下环境变量
|
5 |
+
------------------------------------------------------------------
|
6 |
+
TENCENT_SECRET_ID 腾讯云 SecretId
|
7 |
+
TENCENT_SECRET_KEY 腾讯云 SecretKey
|
8 |
+
TENCENT_TRANSLATE_URL (可选) 默认 https://tmt.tencentcloudapi.com
|
9 |
+
BAIDU_TRANSLATE_URL (可选) 默认 https://fanyi-api.baidu.com/api/trans/vip/translate
|
10 |
+
BAIDU_CREDENTIALS_JSON 形如:
|
11 |
+
[
|
12 |
+
{"app_id": "xxxx", "secret_key": "yyyy"},
|
13 |
+
{"app_id": "aaaa", "secret_key": "bbbb"}
|
14 |
+
]
|
15 |
+
------------------------------------------------------------------
|
16 |
+
"""
|
17 |
+
import hashlib, hmac, json, os, random, time
|
18 |
+
from datetime import datetime
|
19 |
+
from typing import List, Sequence, Optional
|
20 |
+
|
21 |
+
import requests
|
22 |
+
|
23 |
+
# ------------------------------------------------------------------
|
24 |
+
# 读取环境变量
|
25 |
+
# ------------------------------------------------------------------
|
26 |
+
TENCENT_SECRET_ID = os.environ.get("TENCENT_SECRET_ID")
|
27 |
+
TENCENT_SECRET_KEY = os.environ.get("TENCENT_SECRET_KEY")
|
28 |
+
TENCENT_TRANSLATE_URL = os.environ.get("TENCENT_TRANSLATE_URL", "https://tmt.tencentcloudapi.com")
|
29 |
+
|
30 |
+
BAIDU_TRANSLATE_URL = os.environ.get("BAIDU_TRANSLATE_URL", "https://fanyi-api.baidu.com/api/trans/vip/translate")
|
31 |
+
BAIDU_CREDENTIALS = json.loads(os.environ.get("BAIDU_CREDENTIALS_JSON", "[]"))
|
32 |
+
|
33 |
+
# 内部轮询索引
|
34 |
+
_baidu_idx: int = 0
|
35 |
+
def _next_baidu_cred():
|
36 |
+
global _baidu_idx
|
37 |
+
if not BAIDU_CREDENTIALS:
|
38 |
+
return None
|
39 |
+
cred = BAIDU_CREDENTIALS[_baidu_idx]
|
40 |
+
_baidu_idx = (_baidu_idx + 1) % len(BAIDU_CREDENTIALS)
|
41 |
+
return cred
|
42 |
+
|
43 |
+
# ------------------------------------------------------------------
|
44 |
+
# 腾讯翻译
|
45 |
+
# ------------------------------------------------------------------
|
46 |
+
def _sign(key: bytes, msg: str) -> bytes:
|
47 |
+
return hmac.new(key, msg.encode("utf-8"), hashlib.sha256).digest()
|
48 |
+
|
49 |
+
def _tc3_signature(secret_key: str, date: str, service: str, string_to_sign: str) -> str:
|
50 |
+
secret_date = _sign(("TC3" + secret_key).encode(), date)
|
51 |
+
secret_service = _sign(secret_date, service)
|
52 |
+
secret_signing = _sign(secret_service, "tc3_request")
|
53 |
+
return hmac.new(secret_signing, string_to_sign.encode("utf-8"), hashlib.sha256).hexdigest()
|
54 |
+
|
55 |
+
def _translate_with_tencent(texts: Sequence[str], src="auto", tgt="zh") -> Optional[List[str]]:
|
56 |
+
"""优先使用腾讯云翻译。失败返回 None"""
|
57 |
+
if not (TENCENT_SECRET_ID and TENCENT_SECRET_KEY):
|
58 |
+
return None # 未配置凭证
|
59 |
+
service = "tmt"
|
60 |
+
host = "tmt.tencentcloudapi.com"
|
61 |
+
action = "TextTranslate"
|
62 |
+
version = "2018-03-21"
|
63 |
+
region = "ap-beijing"
|
64 |
+
ts = int(time.time())
|
65 |
+
date = datetime.utcfromtimestamp(ts).strftime("%Y-%m-%d")
|
66 |
+
algorithm = "TC3-HMAC-SHA256"
|
67 |
+
|
68 |
+
payload = {
|
69 |
+
"SourceText": "\n".join(texts),
|
70 |
+
"Source": src,
|
71 |
+
"Target": tgt,
|
72 |
+
"ProjectId": 0,
|
73 |
+
}
|
74 |
+
payload_str = json.dumps(payload, ensure_ascii=False)
|
75 |
+
|
76 |
+
# ---------- step‑1 canonical request ----------
|
77 |
+
canonical_request = "\n".join([
|
78 |
+
"POST",
|
79 |
+
"/",
|
80 |
+
"",
|
81 |
+
f"content-type:application/json; charset=utf-8\nhost:{host}\nx-tc-action:{action.lower()}\n",
|
82 |
+
"content-type;host;x-tc-action",
|
83 |
+
hashlib.sha256(payload_str.encode()).hexdigest(),
|
84 |
+
])
|
85 |
+
|
86 |
+
# ---------- step‑2 string to sign ----------
|
87 |
+
credential_scope = f"{date}/{service}/tc3_request"
|
88 |
+
string_to_sign = "\n".join([
|
89 |
+
algorithm, str(ts), credential_scope,
|
90 |
+
hashlib.sha256(canonical_request.encode()).hexdigest(),
|
91 |
+
])
|
92 |
+
|
93 |
+
# ---------- step‑3 signature ----------
|
94 |
+
signature = _tc3_signature(TENCENT_SECRET_KEY, date, service, string_to_sign)
|
95 |
+
|
96 |
+
# ---------- step‑4 headers ----------
|
97 |
+
authorization = (
|
98 |
+
f"{algorithm} Credential={TENCENT_SECRET_ID}/{credential_scope}, "
|
99 |
+
f"SignedHeaders=content-type;host;x-tc-action, Signature={signature}"
|
100 |
+
)
|
101 |
+
headers = {
|
102 |
+
"Authorization": authorization,
|
103 |
+
"Content-Type": "application/json; charset=utf-8",
|
104 |
+
"Host": host,
|
105 |
+
"X-TC-Action": action,
|
106 |
+
"X-TC-Timestamp": str(ts),
|
107 |
+
"X-TC-Version": version,
|
108 |
+
"X-TC-Region": region,
|
109 |
+
}
|
110 |
+
|
111 |
+
# ---------- request ----------
|
112 |
+
try:
|
113 |
+
resp = requests.post(TENCENT_TRANSLATE_URL, headers=headers, data=payload_str, timeout=8)
|
114 |
+
resp.raise_for_status()
|
115 |
+
data = resp.json()
|
116 |
+
return data["Response"]["TargetText"].split("\n")
|
117 |
+
except Exception as e:
|
118 |
+
print(f"[translator] Tencent API error → {e}")
|
119 |
+
return None
|
120 |
+
|
121 |
+
# ------------------------------------------------------------------
|
122 |
+
# 百度翻译
|
123 |
+
# ------------------------------------------------------------------
|
124 |
+
def _translate_with_baidu(texts: Sequence[str], src="auto", tgt="zh") -> Optional[List[str]]:
|
125 |
+
creds = _next_baidu_cred()
|
126 |
+
if creds is None:
|
127 |
+
return None # 未配置凭证
|
128 |
+
app_id, secret_key = creds["app_id"], creds["secret_key"]
|
129 |
+
salt = random.randint(32768, 65536)
|
130 |
+
query = "\n".join(texts)
|
131 |
+
sign = hashlib.md5((app_id + query + str(salt) + secret_key).encode()).hexdigest()
|
132 |
+
params = {
|
133 |
+
"q": query, "from": src, "to": tgt,
|
134 |
+
"appid": app_id, "salt": salt, "sign": sign,
|
135 |
+
}
|
136 |
+
try:
|
137 |
+
resp = requests.get(BAIDU_TRANSLATE_URL, params=params, timeout=8)
|
138 |
+
resp.raise_for_status()
|
139 |
+
data = resp.json()
|
140 |
+
return [item["dst"] for item in data["trans_result"]]
|
141 |
+
except Exception as e:
|
142 |
+
print(f"[translator] Baidu API error → {e}")
|
143 |
+
return None
|
144 |
+
|
145 |
+
# ------------------------------------------------------------------
|
146 |
+
# 对外统一函数
|
147 |
+
# ------------------------------------------------------------------
|
148 |
+
def translate_texts(texts: Sequence[str],
|
149 |
+
src_lang: str = "auto",
|
150 |
+
tgt_lang: str = "zh") -> List[str]:
|
151 |
+
"""
|
152 |
+
优先 Tencent → 失败再 Baidu → 如果都失败,返回原文。
|
153 |
+
"""
|
154 |
+
if not texts:
|
155 |
+
return []
|
156 |
+
|
157 |
+
out = _translate_with_tencent(texts, src_lang, tgt_lang)
|
158 |
+
if out is None:
|
159 |
+
out = _translate_with_baidu(texts, src_lang, tgt_lang)
|
160 |
+
return out or list(texts)
|