study-tool
305 lines
13 KiB
Python
305 lines
13 KiB
Python
"""广东公需课：jsglpt 登录 → jsxx 学习。
|
||
|
||
登录在 `jsglpt.gds.edu.cn`（账号 / 密码都是 base64），成功后用
|
||
`ssoItlmsGxkTicket.action` 换 ITLMS 票，带票进 `jsxx.gds.edu.cn` 拿学习会话；
|
||
两个域名的 cookie 都留在同一个 requests.Session 里。
|
||
|
||
学习页是服务端渲染的老系统，视频是 CDN 直链 MP4，进度靠 `updateViewTime1`
|
||
心跳上报（`lastViewTime` 是已看秒数，页面上「要求观看视频 N 分钟」才是完成门槛）。
|
||
平台按真实流逝时间放行（实测一次报 +300 只记 +60），所以只能 1 倍速慢慢刷。
|
||
|
||
本文件只留 `Platform` 子类与注册，工具按阶段拆到 `utils/`（登录 / 课程解析 / 心跳）。
|
||
"""
|
||
|
||
import re
|
||
|
||
import requests
|
||
|
||
from study_tool import database, notify, runtime
|
||
from study_tool.platforms import registry
|
||
from study_tool.platforms.base import Platform
|
||
|
||
from .utils import course, exam, http, learn, login, register
|
||
|
||
|
||
MAX_ROUNDS = 5 # 最多「刷一轮 → 重取进度页」的轮数
|
||
MAX_STALL = 3 # 连续这么多轮平台完成数没增加就放弃，避免非视频任务点死循环
|
||
ROUND_WAIT = 60 # 两轮之间等平台刷新完成状态的秒数
|
||
EXAM_MARKS = ("考核", "考试") # 课程目录里这两类活动是试卷
|
||
|
||
|
||
def _exam_activity(courseware):
|
||
"""课程目录里的考核活动：考核是独立活动页（/<aid>/study/course/<cid>），题目只在它上面。"""
|
||
return next((c for c in courseware or [] if any(k in (c.get("title") or "") for k in EXAM_MARKS)), None)
|
||
|
||
|
||
def _courseware(cid, acts):
|
||
"""进度页的活动清单 → 引擎用的 courseware 条目。"""
|
||
return [
|
||
{
|
||
"aid": a["aid"],
|
||
"title": a["title"],
|
||
"done": a["done"],
|
||
"href": http.X + "/" + a["aid"] + "/study/course/" + cid,
|
||
"total": 0,
|
||
"viewTime": 0,
|
||
}
|
||
for a in acts
|
||
]
|
||
|
||
|
||
class Gdgxk(Platform):
|
||
"""广东公需课：jsglpt 登录（账号 / 密码都是 base64 + 图形验证码），jsxx 学习。
|
||
|
||
账号级能力都实现在这里；反馈表 / 测验这类平台没有的阶段不实现，保持基类的 None。
|
||
"""
|
||
|
||
id = http.PLATFORM_ID
|
||
label = "广东公需课"
|
||
entries = ("jsglpt.gds.edu.cn", "jsxx.gds.edu.cn")
|
||
|
||
def session(self, cookies=None):
|
||
s = requests.Session()
|
||
s.headers.update({"User-Agent": http.UA, "Accept-Language": "zh-CN,zh;q=0.9"})
|
||
for c in cookies or []:
|
||
if "gds.edu.cn" not in (c.get("domain") or ""):
|
||
continue
|
||
s.cookies.set(
|
||
c["name"], c.get("value") or "", domain=(c.get("domain") or "").lstrip("."), path=c.get("path") or "/"
|
||
)
|
||
return s
|
||
|
||
def dump_cookies(self, s):
|
||
return [{"name": c.name, "value": c.value, "domain": c.domain, "path": c.path} for c in s.cookies]
|
||
|
||
def login(self, username, password):
|
||
s = self.session()
|
||
s.get(http.G + "/login.jsp", timeout=30)
|
||
err = ""
|
||
for _ in range(5):
|
||
data = {
|
||
"loginUser.userName": login.b64(username),
|
||
"loginUser.loginType": "normal",
|
||
"loginUser.password": login.b64(password),
|
||
"captcha": login.ocr().classification(login.captcha(s)),
|
||
}
|
||
r = s.post(http.G + "/loginIn.do", data=data, allow_redirects=False, timeout=30)
|
||
if r.status_code == 302 and "/default.jsp" in (r.headers.get("location") or ""):
|
||
login.sso(s)
|
||
return s, self.profile(s) or username
|
||
err = login.login_error(s)
|
||
if "验证码" not in err: # 不是验证码的问题，重试没意义
|
||
break
|
||
raise RuntimeError("登录失败：%s" % (err or "账号或密码不对"))
|
||
|
||
def renew_session(self, s):
|
||
try:
|
||
login.sso(s)
|
||
return True
|
||
except Exception:
|
||
return False
|
||
|
||
def profile(self, s):
|
||
"""个人中心里平台显示的用户名：学员是打码姓名（如「*栋」），老师是完整姓名。"""
|
||
html = s.get(http.G + "/teacher/index.do", timeout=30).text
|
||
# 左上角头像区：<a href="javascript:;" class="name" ...>姓名</a>
|
||
m = re.search(r'class="name"[^>]*>([^<]{1,20})</a>', html)
|
||
if m and m.group(1).strip():
|
||
return m.group(1).strip()
|
||
m = re.search(r"你好[，,]?\s*([^\s<]+)\s*老师", re.sub(r"\s+", " ", html))
|
||
return m.group(1) if m else ""
|
||
|
||
def list_courses(self, s):
|
||
html = s.get(http.X + "/uc/store/courseRegister", timeout=30).text
|
||
out = course.parse_courses(html)
|
||
if not out and "logout" not in html:
|
||
raise http.expired("学习平台会话已失效，请重新登录")
|
||
# 没有目标年度的课：去选课超市选一门，再重新读列表
|
||
if not any(register.config()[0] in c["section"] for c in out) and register.ensure(s):
|
||
out = course.parse_courses(s.get(http.X + "/uc/store/courseRegister", timeout=30).text)
|
||
return out
|
||
|
||
def discover(self, s, plan_id, project_id, name=""):
|
||
cid = plan_id
|
||
page = http.X + "/study/course/" + cid
|
||
s.headers["Referer"] = page
|
||
r = s.get(page + "?v=1", timeout=30)
|
||
if "loginBtn" in r.text and "md5Uid" not in r.text:
|
||
raise http.expired("学习平台会话已失效，请重新登录")
|
||
# 一门课是「章 / 节 / 活动」的目录，不是单个视频；活动清单和完成情况都从进度页取
|
||
pr = course.progress(s, cid)
|
||
if not pr["acts"]:
|
||
raise RuntimeError("进度页里没有找到活动列表，页面结构可能变了")
|
||
courseware = _courseware(cid, pr["acts"])
|
||
return {
|
||
"user": self.profile(s),
|
||
"course": name or "公需课程",
|
||
"progress": "已完成 %d/%d 个活动" % (pr["done"], pr["total"]),
|
||
"info": page,
|
||
"courseware": courseware,
|
||
"feedback": [],
|
||
"test": [],
|
||
"platform": http.PLATFORM_ID,
|
||
"generation": "gxk",
|
||
"plan_id": cid,
|
||
"project_id": "",
|
||
}
|
||
|
||
def study(self, key, s, d, **opts):
|
||
"""按进度页的活动清单逐个心跳；平台只按真实时间放行，所以默认 1 倍速慢慢刷。"""
|
||
speed = float(opts.get("speed") or 1)
|
||
beat = int(opts.get("beat") or 60)
|
||
cid = d["plan_id"]
|
||
cws = d["courseware"]
|
||
runtime.set_state(key, status="studying", total=len(cws), done=0)
|
||
best, stall, pr, remain = -1, 0, {"done": 0, "total": len(cws)}, []
|
||
for rnd in range(1, MAX_ROUNDS + 1):
|
||
todo = [c for c in cws if not c.get("done")]
|
||
base = len(cws) - len(todo)
|
||
for i, cw in enumerate(todo, 1):
|
||
if learn.stopped(key):
|
||
runtime.log(key, "已按要求停止")
|
||
return
|
||
for attempt in (1, 2):
|
||
try:
|
||
learn.study_one(key, s, cw, speed, beat)
|
||
break
|
||
except Exception as e:
|
||
expired = e.__class__.__name__ == "SessionExpired"
|
||
if expired and attempt == 1:
|
||
s = self._relogin(key) # 掉线：重登后把当前活动重新走一遍
|
||
continue
|
||
if expired:
|
||
raise
|
||
runtime.log(key, "活动「%s」本轮跳过：%s" % (cw["title"], e), level="warn")
|
||
break
|
||
runtime.set_state(key, done=base + i)
|
||
# 以平台进度页为准：不管本轮刷了几个（包括一个都不用刷），都重取任务点状态
|
||
s, pr = self._progress(key, s, cid)
|
||
cws = _courseware(cid, pr["acts"])
|
||
remain = [c["title"] for c in cws if not c["done"]]
|
||
runtime.set_state(key, done=pr["done"], total=pr["total"])
|
||
if pr["done"] >= pr["total"] and not remain:
|
||
break
|
||
stall = 0 if pr["done"] > best else stall + 1
|
||
best = pr["done"]
|
||
if stall >= MAX_STALL:
|
||
break
|
||
runtime.log(
|
||
key,
|
||
"第 %d 轮结束，平台进度 %d/%d，仍有 %d 个任务点未完成，稍后自动重新学习：%s"
|
||
% (rnd, pr["done"], pr["total"], len(remain), "、".join(remain) or "（进度页未列出具体任务点）"),
|
||
level="warn",
|
||
)
|
||
notify.retry(
|
||
key,
|
||
"未完成任务点自动重学",
|
||
"平台进度 %d/%d，第 %d 轮仍有 %d 个任务点未完成" % (pr["done"], pr["total"], rnd, len(remain)),
|
||
)
|
||
if not learn.wait(key, ROUND_WAIT):
|
||
runtime.log(key, "已按要求停止")
|
||
return
|
||
if remain: # 视频都刷完了还有缺口，多半是课程考核；交完卷再按进度页复核一次
|
||
cw = _exam_activity(cws)
|
||
if cw:
|
||
s = self._exam(key, s, cw["href"], opts)
|
||
s, pr = self._progress(key, s, cid)
|
||
remain = [c["title"] for c in _courseware(cid, pr["acts"]) if not c["done"]]
|
||
runtime.set_state(key, status="done", done=pr["done"], total=pr["total"])
|
||
if pr["done"] >= pr["total"] and not remain:
|
||
runtime.log(key, "平台进度：%d/%d 个活动已完成" % (pr["done"], pr["total"]))
|
||
else:
|
||
runtime.log(
|
||
key,
|
||
"平台进度：%d/%d 个活动已完成，自动重学后仍未完成，需人工处理%s"
|
||
% (pr["done"], pr["total"], ("（" + "、".join(remain) + "）") if remain else ""),
|
||
level="warn",
|
||
)
|
||
|
||
def _exam(self, key, s, url, opts):
|
||
"""课程考核：题库优先、大模型兜底；题库没覆盖全部题目就不交卷。掉线重登后重试一次。"""
|
||
auto = opts.get("exam_auto")
|
||
auto = True if auto is None else bool(auto)
|
||
try:
|
||
s, msg = exam.run(key, s, url, auto)
|
||
except Exception as e:
|
||
if e.__class__.__name__ != "SessionExpired":
|
||
runtime.log(key, "考核处理失败：%s" % e, level="warn")
|
||
return s
|
||
s = self._relogin(key)
|
||
try:
|
||
s, msg = exam.run(key, s, url, auto)
|
||
except Exception as e2:
|
||
runtime.log(key, "考核处理失败：%s" % e2, level="warn")
|
||
return s
|
||
if msg:
|
||
runtime.set_state(key, status="exam", total=1, done=1)
|
||
return s
|
||
|
||
def quiz(self, s, d, b):
|
||
"""课程考核的题目：考核是独立活动页，解析它上面的 testForm。已交卷也返回题目，方便回看。"""
|
||
cw = _exam_activity(d.get("courseware"))
|
||
if not cw:
|
||
return {"ok": False, "error": "这门课没有考核"}
|
||
cid = d["plan_id"]
|
||
html = s.get(cw["href"], timeout=30).text
|
||
done = exam.finished(html)
|
||
ex = exam.parse(html)
|
||
questions = [
|
||
{
|
||
"no": str(q["no"]),
|
||
"type": exam.TYPE_CODE.get(q["type"], "01"),
|
||
"title": q["title"],
|
||
"options": [{"key": o["key"], "text": o["text"]} for o in q["options"]],
|
||
}
|
||
for q in (ex or {}).get("questions") or []
|
||
]
|
||
if not done and not questions:
|
||
return {"ok": False, "error": "考核页里没有找到题目（页面结构可能变了）"}
|
||
return {
|
||
"ok": True,
|
||
"platform": http.PLATFORM_ID,
|
||
"plan_id": cid,
|
||
"object_id": cw["aid"],
|
||
"title": "课程考核",
|
||
"done": done,
|
||
"questions": questions,
|
||
}
|
||
|
||
def quiz_submit(self, s, b):
|
||
"""弹窗里点交卷：按用户选的字母答案提交，返回得分。"""
|
||
if not b.object_id:
|
||
return {"ok": False, "error": "缺少考核活动 ID，请重新打开测验"}
|
||
url = http.X + "/" + b.object_id + "/study/course/" + b.plan_id
|
||
return exam.submit_letters(s, url, b.answers)
|
||
|
||
def _progress(self, key, s, cid):
|
||
"""取进度页；掉线就重登后再取一次，返回 (会话, 进度)。"""
|
||
try:
|
||
return s, course.progress(s, cid)
|
||
except Exception as e:
|
||
if e.__class__.__name__ != "SessionExpired":
|
||
raise
|
||
notify.retry(key, "会话掉线重登", "取进度页时会话失效，自动重新登录")
|
||
s = self._relogin(key)
|
||
return s, course.progress(s, cid)
|
||
|
||
def _relogin(self, key):
|
||
"""掉线后按库里存的账号密码重登，返回新会话。"""
|
||
acc = database.account_get(key) or {}
|
||
pw = database.account_password(key)
|
||
if not pw:
|
||
raise http.expired("会话掉线，且没有保存密码，无法自动重连")
|
||
runtime.log(key, "会话掉线，正在用账号密码自动重新登录…", level="warn")
|
||
s, _ = self.login(acc.get("username") or "", pw)
|
||
# 新会话要写回账号和运行状态：否则库里留的还是刚失效的那份，
|
||
# 下次「刷新课程」立刻又判过期，反复触发重登。
|
||
cookies = self.dump_cookies(s)
|
||
database.account_save_cookies(key, cookies)
|
||
runtime.set_state(key, cookies=cookies)
|
||
runtime.log(key, "重新登录成功，继续学习")
|
||
return s
|
||
|
||
|
||
PLATFORM = registry.register(Gdgxk())
|