data models

Typed wrappers around portal JSON. Attendance: Attendance models. Exams: Exam models. Registration: Registration models. Captcha: Tokens and captcha.

Many methods still return raw dict (get_attendance, get_exam_schedule, grade card, fees). Models exist where the wrapper bothers to parse.

flowchart TB
  subgraph attendance [attendance.py]
    AH["AttendanceHeader"]
    Sem["Semester"]
    AM["AttendanceMeta"]
  end
  subgraph exam [exam.py]
    EE["ExamEvent"]
  end
  subgraph registration [registration.py]
    RS["RegisteredSubject"]
    Reg["Registrations"]
  end
  subgraph tokens [tokens.py]
    Cap["Captcha"]
  end
  AM --> AH
  AM --> Sem
  Reg --> RS

Every dataclass except AttendanceMeta and Registrations has from_json(resp: dict). Those two take the full response object in __init__ and map child lists.

flowchart LR
  JSON["resp['response']"] --> Model["from_json or __init__"]
  Model --> Obj["typed object"]

Semester is defined in attendance.py even when used for exams, marks, and grade cards.

Portal keys are lowercase mashed words (registrationid, employeename, exameventid). Python attributes use underscores (registration_id, employee_name, exam_event_id). from_json is the only translation layer.

classDiagram
  class AttendanceMeta {
    headers
    semesters
    latest_header()
    latest_semester()
  }
  class Registrations {
    total_credits
    subjects
  }
  class ExamEvent {
    exam_event_id
    registration_id
  }
  class Captcha {
    captcha
    hidden
    image
    payload()
  }

Bank info, fee summary, fines, attendance detail rows, exam schedule, grade card, SGPA/CGPA, subject choices. Those stay as the portal’s dicts, typos included (LTpercantage, registationid on the way out).