chore: initialize lawyers crawler project
This commit is contained in:
@@ -0,0 +1,19 @@
|
||||
from request.requests_client import (
|
||||
RequestClientError,
|
||||
RequestConnectTimeout,
|
||||
RequestConnectionError,
|
||||
RequestSSLError,
|
||||
RequestTimeout,
|
||||
RequestsClient,
|
||||
ResponseData,
|
||||
)
|
||||
|
||||
__all__ = [
|
||||
"RequestsClient",
|
||||
"ResponseData",
|
||||
"RequestClientError",
|
||||
"RequestConnectTimeout",
|
||||
"RequestTimeout",
|
||||
"RequestConnectionError",
|
||||
"RequestSSLError",
|
||||
]
|
||||
@@ -0,0 +1,97 @@
|
||||
import json
|
||||
import os
|
||||
from typing import Dict, Optional
|
||||
|
||||
CONFIG_PATH = os.path.join(os.path.dirname(__file__), "proxy_settings.json")
|
||||
|
||||
DEFAULT_CONFIG = {
|
||||
"enabled": True,
|
||||
"tunnel": "t133.kdltps.com:15818",
|
||||
"username": "t16766298346583",
|
||||
"password": "zyn0vb20",
|
||||
"scheme": "http",
|
||||
}
|
||||
|
||||
_PROXY_STATUS_REPORTED = False
|
||||
|
||||
|
||||
def _normalize_bool(value, default: bool = True) -> bool:
|
||||
if value is None:
|
||||
return default
|
||||
if isinstance(value, bool):
|
||||
return value
|
||||
text = str(value).strip().lower()
|
||||
return text not in ("0", "false", "no", "off", "")
|
||||
|
||||
|
||||
def _load_config() -> Dict[str, str]:
|
||||
if not os.path.exists(CONFIG_PATH):
|
||||
return dict(DEFAULT_CONFIG)
|
||||
try:
|
||||
with open(CONFIG_PATH, "r", encoding="utf-8") as f:
|
||||
data = json.load(f) or {}
|
||||
except Exception as exc:
|
||||
print(f"[proxy] 配置读取失败: {exc}, 使用默认配置")
|
||||
return dict(DEFAULT_CONFIG)
|
||||
|
||||
config = dict(DEFAULT_CONFIG)
|
||||
for key, value in data.items():
|
||||
if value is not None:
|
||||
config[key] = value
|
||||
return config
|
||||
|
||||
|
||||
def report_proxy_status() -> None:
|
||||
global _PROXY_STATUS_REPORTED
|
||||
if _PROXY_STATUS_REPORTED:
|
||||
return
|
||||
_PROXY_STATUS_REPORTED = True
|
||||
|
||||
config = _load_config()
|
||||
enabled = _normalize_bool(config.get("enabled"), True)
|
||||
if not enabled:
|
||||
print("[proxy] disabled by config")
|
||||
return
|
||||
|
||||
missing = [key for key in ("tunnel", "username", "password") if not config.get(key)]
|
||||
if missing:
|
||||
print(f"[proxy] enabled but missing fields: {', '.join(missing)}")
|
||||
return
|
||||
print(f"[proxy] enabled=True tunnel={config.get('tunnel')}")
|
||||
|
||||
|
||||
def get_proxies() -> Optional[Dict[str, str]]:
|
||||
"""
|
||||
返回统一的代理配置;当配置 enabled=false 时返回 None。
|
||||
代理配置从 proxy_settings.json 读取,不依赖环境变量。
|
||||
"""
|
||||
config = _load_config()
|
||||
if not _normalize_bool(config.get("enabled"), True):
|
||||
return None
|
||||
|
||||
tunnel = str(config.get("tunnel") or "").strip()
|
||||
username = str(config.get("username") or "").strip()
|
||||
password = str(config.get("password") or "").strip()
|
||||
scheme = str(config.get("scheme") or "http").strip().lower()
|
||||
|
||||
if not tunnel or not username or not password:
|
||||
print("[proxy] missing proxy credentials, proxy disabled")
|
||||
return None
|
||||
|
||||
proxy = f"{scheme}://{username}:{password}@{tunnel}/"
|
||||
return {"http": proxy, "https": proxy}
|
||||
|
||||
|
||||
def apply_proxy(session) -> Optional[Dict[str, str]]:
|
||||
"""为 requests.Session 应用统一代理配置,返回最终代理字典或 None。"""
|
||||
report_proxy_status()
|
||||
proxies = get_proxies()
|
||||
session.trust_env = False
|
||||
if proxies:
|
||||
session.proxies.update(proxies)
|
||||
else:
|
||||
session.proxies.clear()
|
||||
return proxies
|
||||
|
||||
|
||||
__all__ = ["get_proxies", "apply_proxy", "report_proxy_status"]
|
||||
@@ -0,0 +1,7 @@
|
||||
{
|
||||
"enabled": true,
|
||||
"tunnel": "t133.kdltps.com:15818",
|
||||
"username": "t16766298346583",
|
||||
"password": "zyn0vb20",
|
||||
"scheme": "http"
|
||||
}
|
||||
@@ -0,0 +1,168 @@
|
||||
from dataclasses import dataclass
|
||||
from typing import Any, Dict, Iterable, Mapping, Optional, Tuple, Union
|
||||
|
||||
import requests
|
||||
from requests.adapters import HTTPAdapter
|
||||
from urllib3.util.retry import Retry
|
||||
|
||||
from request.proxy_config import apply_proxy
|
||||
|
||||
TimeoutType = Union[float, Tuple[float, float]]
|
||||
|
||||
|
||||
class RequestClientError(Exception):
|
||||
"""请求客户端通用异常。"""
|
||||
|
||||
|
||||
class RequestConnectTimeout(RequestClientError):
|
||||
"""连接超时。"""
|
||||
|
||||
|
||||
class RequestTimeout(RequestClientError):
|
||||
"""请求超时。"""
|
||||
|
||||
|
||||
class RequestConnectionError(RequestClientError):
|
||||
"""连接错误。"""
|
||||
|
||||
|
||||
class RequestSSLError(RequestClientError):
|
||||
"""SSL 错误。"""
|
||||
|
||||
|
||||
@dataclass
|
||||
class ResponseData:
|
||||
# 只保留采集侧稳定需要的字段,避免直接向上层泄露原始 Response 对象
|
||||
status_code: int
|
||||
text: str
|
||||
url: str
|
||||
headers: Dict[str, str]
|
||||
|
||||
|
||||
class RequestsClient:
|
||||
"""
|
||||
统一 requests 客户端:
|
||||
- 自动应用代理配置
|
||||
- 支持可选重试
|
||||
- 对外抛出统一异常类型
|
||||
"""
|
||||
|
||||
def __init__(
|
||||
self,
|
||||
headers: Optional[Mapping[str, str]] = None,
|
||||
*,
|
||||
retry_total: int = 0,
|
||||
retry_backoff_factor: float = 0.0,
|
||||
retry_status_forcelist: Optional[Iterable[int]] = None,
|
||||
retry_allowed_methods: Optional[Iterable[str]] = None,
|
||||
default_timeout: Optional[TimeoutType] = None,
|
||||
) -> None:
|
||||
self._base_headers: Dict[str, str] = dict(headers or {})
|
||||
self.retry_total = int(retry_total)
|
||||
self.retry_backoff_factor = float(retry_backoff_factor)
|
||||
self.retry_status_forcelist = tuple(retry_status_forcelist or ())
|
||||
self.retry_allowed_methods = tuple(retry_allowed_methods or ("GET", "POST"))
|
||||
self.default_timeout = default_timeout
|
||||
self._session = self._build_session()
|
||||
|
||||
def _build_session(self) -> requests.Session:
|
||||
session = requests.Session()
|
||||
# 统一从 proxy_settings.json 注入代理,并屏蔽系统环境代理干扰
|
||||
apply_proxy(session)
|
||||
if self.retry_total > 0:
|
||||
# 适配器级重试:主要处理连接波动与指定状态码的瞬时失败
|
||||
retries = Retry(
|
||||
total=self.retry_total,
|
||||
backoff_factor=self.retry_backoff_factor,
|
||||
status_forcelist=self.retry_status_forcelist,
|
||||
allowed_methods=frozenset(self.retry_allowed_methods),
|
||||
raise_on_status=False,
|
||||
)
|
||||
adapter = HTTPAdapter(max_retries=retries)
|
||||
session.mount("https://", adapter)
|
||||
session.mount("http://", adapter)
|
||||
if self._base_headers:
|
||||
# 基础头只在建 session 时注入,业务请求可通过 headers 临时覆盖
|
||||
session.headers.update(self._base_headers)
|
||||
return session
|
||||
|
||||
@property
|
||||
def headers(self):
|
||||
return self._session.headers
|
||||
|
||||
@property
|
||||
def proxies(self) -> Dict[str, str]:
|
||||
return dict(self._session.proxies)
|
||||
|
||||
def refresh(self) -> None:
|
||||
# 强制重建 session,用于 403/连接异常后的“换连接”场景
|
||||
self.close()
|
||||
self._session = self._build_session()
|
||||
|
||||
def close(self) -> None:
|
||||
try:
|
||||
self._session.close()
|
||||
except Exception:
|
||||
pass
|
||||
|
||||
def clone(self) -> "RequestsClient":
|
||||
# 线程场景建议 clone:复用同配置,但使用独立连接池
|
||||
clone_client = RequestsClient(
|
||||
headers=dict(self.headers),
|
||||
retry_total=self.retry_total,
|
||||
retry_backoff_factor=self.retry_backoff_factor,
|
||||
retry_status_forcelist=self.retry_status_forcelist,
|
||||
retry_allowed_methods=self.retry_allowed_methods,
|
||||
default_timeout=self.default_timeout,
|
||||
)
|
||||
return clone_client
|
||||
|
||||
def request_text(
|
||||
self,
|
||||
method: str,
|
||||
url: str,
|
||||
*,
|
||||
timeout: Optional[TimeoutType] = None,
|
||||
verify: bool = True,
|
||||
headers: Optional[Mapping[str, str]] = None,
|
||||
**kwargs: Any,
|
||||
) -> ResponseData:
|
||||
response = None
|
||||
# 调用方未传 timeout 时,回退到客户端默认超时
|
||||
real_timeout = self.default_timeout if timeout is None else timeout
|
||||
try:
|
||||
response = self._session.request(
|
||||
method=method,
|
||||
url=url,
|
||||
timeout=real_timeout,
|
||||
verify=verify,
|
||||
headers=headers,
|
||||
**kwargs,
|
||||
)
|
||||
return ResponseData(
|
||||
status_code=response.status_code,
|
||||
text=response.text,
|
||||
url=response.url,
|
||||
headers=dict(response.headers),
|
||||
)
|
||||
# 把 requests 的具体异常统一收敛,业务层无需依赖 requests.exceptions
|
||||
except requests.exceptions.ConnectTimeout as exc:
|
||||
raise RequestConnectTimeout(str(exc)) from exc
|
||||
except requests.exceptions.Timeout as exc:
|
||||
raise RequestTimeout(str(exc)) from exc
|
||||
except requests.exceptions.ConnectionError as exc:
|
||||
raise RequestConnectionError(str(exc)) from exc
|
||||
except requests.exceptions.SSLError as exc:
|
||||
raise RequestSSLError(str(exc)) from exc
|
||||
except requests.exceptions.RequestException as exc:
|
||||
raise RequestClientError(str(exc)) from exc
|
||||
finally:
|
||||
if response is not None:
|
||||
# 立即释放底层连接,避免大量采集时连接堆积
|
||||
response.close()
|
||||
|
||||
def get_text(self, url: str, **kwargs: Any) -> ResponseData:
|
||||
return self.request_text("GET", url, **kwargs)
|
||||
|
||||
def post_text(self, url: str, **kwargs: Any) -> ResponseData:
|
||||
return self.request_text("POST", url, **kwargs)
|
||||
Reference in New Issue
Block a user