A Streamlit app that helps a pet owner build and manage a smart daily care plan for one or more pets.
.png)
Set up owner profile and add pets with species, breed, and age.
.png)
Add care tasks with duration, priority, and recurrence (once, daily, or weekly).
View the daily schedule with sorting, filtering, and automatic conflict detection.
- Owner profile — enter your name and contact info; the profile persists across the entire session.
- Multi-pet support — add as many pets as you need, each with a name, species, breed, and age.
- Flexible task entry — record any care task (walk, feed, meds, grooming, vet visit, etc.) with a duration, a priority score (1–10), and an optional preferred time.
- Recurring tasks — mark a task as
dailyorweekly; when you complete it the next occurrence is automatically scheduled using Python'stimedelta(+1 day or +7 days) and added to the pet's queue. - One-time tasks — tasks set to
onceare removed from the plan when completed with no follow-up.
- Priority-first planning —
Scheduler.generate_plan()sorts pending tasks by priority (high → low), using preferred time as a tiebreaker. High-urgency tasks always appear at the top of the plan. - Sort by time —
Scheduler.sort_by_time()re-orders the plan chronologically using alambdakey onpreferred_time, so you can view the day in clock order. Tasks without a preferred time appear last. - Filter by pet or status —
Scheduler.filter_tasks(pet_name, status)returns a focused subset of the plan. Filters can be used independently or combined (e.g. "show only Buddy's pending tasks"). - Conflict detection —
Scheduler.get_conflict_warnings()groups tasks bypreferred_timeand returns a human-readable warning message for every time slot that has more than one task assigned. Each conflict names the exact pets and tasks involved so the owner knows what to reschedule.
- Live schedule table — the plan updates immediately after every change; sort order and filters are applied with inline controls (no page reload needed).
- Per-conflict warning banners — each scheduling conflict surfaces as its own
st.warning()block, naming the time slot and the clashing tasks. - Mark done in-app — select any pending task from a dropdown and mark it complete; recurring tasks spawn their next occurrence and a confirmation banner appears automatically.
- Session persistence — all owner, pet, task, and schedule data is held in
st.session_state, so nothing is lost when you interact with the UI.
pawpal_system.py — core logic: Task, Pet, Owner, Scheduler classes
app.py — Streamlit UI wired to the logic layer
main.py — terminal demo showing sorting, filtering, and recurring tasks
tests/
test_pawpal.py — pytest tests for mark_done() and add_task()
uml_final.png — final UML class diagram (updated from initial design)
generate_uml.py — script used to render uml_final.png
reflection.md — design decisions, tradeoffs, and AI collaboration notes
python -m venv .venv
source .venv/bin/activate # Windows: .venv\Scripts\activate
pip install -r requirements.txtstreamlit run app.pypython main.pypython -m pytestFour algorithms power the scheduling layer:
| Feature | Method | How it works |
|---|---|---|
| Priority plan | Scheduler.generate_plan() |
Sorts by -priority, then preferred_time as tiebreaker |
| Time ordering | Scheduler.sort_by_time() |
sorted() with a lambda key; no-time tasks placed at 23:59 |
| Filtering | Scheduler.filter_tasks() |
Two-pass list comprehension; each filter is optional |
| Conflict detection | Scheduler.get_conflict_warnings() |
defaultdict buckets tasks by time; warns per busy slot |
| Recurring tasks | Task.next_occurrence() + Scheduler.mark_task_complete() |
timedelta advances due_date; new Task added to pet's list |
Key tradeoff: conflict detection compares exact preferred_time values rather than true time intervals (start + duration). This keeps the algorithm simple and readable; see reflection.md section 2b for the full reasoning.