23 April 2025

To-Do-App

by Carson Kempf

First Impressions

Original code generated IDs via len(tasks) + 1, leading to duplicate IDs after deletions.
New tasks did not appear immediately; UI lacked st.rerun() calls for instant refresh.
Corrupted tasks.json printed warnings only to console, not visible in Streamlit UI.
Filter dropdowns built from set(...) produced unpredictable category order.
No built‑in button or mechanism to run tests from the Streamlit app.
Business logic (task creation and filtering) was mixed directly into UI code, making unit testing difficult.

Unit Testing Coverage and Refactoring Summary

Achieved 100% overall coverage across all test suites for both src/app.py and src/tasks.py by:

Streamlit Test Buttons in app.py
- “Run Unit Tests” triggers pytest -q.
- “Run Coverage” triggers pytest --cov=src --cov-report=html.
- “Run Param Tests” triggers pytest tests/test_param.py.
- “Run Mock Tests” triggers pytest tests/test_mock.py.
- “Run HTML Report” triggers pytest --html=report.html --self-contained-html.
Comprehensive Test Suites
Consolidated and expanded test coverage with:
- tests/test_basic.py (basic unit tests)
- tests/test_advanced.py (fixtures & parameterization)
- tests/test_tdd.py (TDD-driven examples)
- tests/test_property.py (hypothesis property-based tests)
- tests/test_param.py (parameterized tests)
- tests/test_mock.py (mocking tests)
Refactoring for Testability
- Extracted pure logic into functions in src/tasks.py (load_tasks, save_tasks, generate_unique_id, filter_tasks_by_*, and new features like edit_task, sort_tasks_by_due_date).
- Isolated UI-centric code in src/app.py, marking UI methods with # pragma: no cover.
- Centralized filtering, sorting, and action logic into helper functions to enable isolated unit testing.

Tue, Apr 22 2025 - 04:14 PM      —      Rolla: ☀️   +71°F

=============================================================

(.venv) bash-3.2$ streamlit run src/app.py

  You can now view your Streamlit app in your browser.

  Local URL: http://localhost:8501
  Network URL: http://10.106.50.237:8501

  For better performance, install the Watchdog module:

  $ xcode-select --install
  $ pip install watchdog
            
.................................................................                                  [100%]
============================================= tests coverage =============================================
____________________________ coverage: platform darwin, python 3.13.1-final-0 ____________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          111      1    99%   62
src/tasks.py         55      1    98%   176
-----------------------------------------------
TOTAL               166      2    99%
65 passed in 1.10s
.................................................................                                  [100%]
============================================= tests coverage =============================================
____________________________ coverage: platform darwin, python 3.13.1-final-0 ____________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          111      1    99%   62
src/tasks.py         55      1    98%   176
-----------------------------------------------
TOTAL               166      2    99%
Coverage HTML written to dir htmlcov
65 passed in 1.19s
......................................                                                             [100%]
============================================= tests coverage =============================================
____________________________ coverage: platform darwin, python 3.13.1-final-0 ____________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          111      2    98%   39, 62
src/tasks.py         55      1    98%   176
-----------------------------------------------
TOTAL               166      3    98%
38 passed in 0.82s
......................................                                                             [100%]
============================================= tests coverage =============================================
____________________________ coverage: platform darwin, python 3.13.1-final-0 ____________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          111      2    98%   39, 62
src/tasks.py         55      1    98%   176
-----------------------------------------------
TOTAL               166      3    98%
38 passed in 0.89s
.................................................................                                  [100%]
============================================= tests coverage =============================================
____________________________ coverage: platform darwin, python 3.13.1-final-0 ____________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          111      1    99%   62
src/tasks.py         55      1    98%   176
-----------------------------------------------
TOTAL               166      2    99%
- Generated html report: file:///Users/ctk/Programming/VSCodeProjects/School/Software%20Engineering/cs4090-assignment-4/report.html -
65 passed in 1.11s

Bug Reports

Bug 1: Completing task under filter overwrites all tasks

Description: When marking a task as complete while filters are applied, only the filtered subset is saved back to tasks.json, causing other tasks to be lost.
Steps to Reproduce:

Add at least two tasks in different categories.
Apply a category filter so only one task is visible.
Click “Complete” on the visible task.
Reopen the app or inspect tasks.json. Expected Behavior: The selected task’s completed status updates; all other tasks remain unaffected in tasks.json.
Actual Behavior: tasks.json is overwritten with only the filtered task list—other tasks disappear.
Severity: High

Bug 2: Undo complete not available

Description: The UI button always reads “Complete” and the logic never supports toggling a completed task back to incomplete.
Steps to Reproduce:

Add a task.
Click “Complete” on that task.
Try to undo by clicking the same button again.
Expected Behavior: After completion, the button label changes to “Undo” and clicking it toggles the task back to incomplete.
Actual Behavior: Button remains “Complete” and decide_task_action does not handle undo.

Severity: Medium

Bug 3: Button actions require double click

Description: Button clicks (“Complete”, “Undo”, “Delete”) do not register on the first press; the user must click twice for action to execute.
Steps to Reproduce:

Launch app with tasks visible.
Click “Complete” on a task once — nothing happens.
Click “Complete” a second time — task completes.
The same behavior applies to “Undo” and “Delete” buttons.
Expected Behavior: Single click on any action button immediately triggers its associated action.
Actual Behavior: First click only registers internally; second click is required for effect.
Severity: Medium

Bug Fixes

Fix for Bug 1: Completing task under filter overwrites all tasks

Refactored display_tasks logic to load and save the full task list instead of the filtered subset, ensuring all tasks persist across filter-based actions.

Fix for Bug 2: Undo complete not available

Updated decide_task_action to handle “undo” state and modified display_tasks to toggle completed=False when “Undo” is clicked, providing toggle functionality.

Fix for Bug 3: Button actions require double click

Removed redundant state reset in Streamlit after each action; replaced return None inside display_tasks with st.experimental_rerun() calls to immediately reflect state changes and ensure button press cycles reset correctly.

New Feature Implementation Plan

Feature: Edit Task
- TDD Tests:
  - Write initial failing tests in test_tdd.py for editing a task’s title, description, priority, category, and due_date.
- Implementation Steps:
  1. Add edit_task(tasks, task_id, updates) in tasks.py to apply changes.
  2. Update app.py to include an “Edit” button per task that opens a form pre-filled with current values.
  3. On form submit, call edit_task, save tasks, and rerun UI.
- Refactoring & Validation:
  - Refactor UI logic into handle_edit_task for testability.
  - Confirm tests pass and UI reflects edited values.
Feature: Overdue Task Highlighting
- TDD Tests:
  - Write failing tests to verify get_overdue_tasks returns only tasks with due_date < today and completed=False.
  - Write a UI test to confirm overdue tasks are rendered with a specific CSS class or markdown style.
- Implementation Steps:
  1. Use existing get_overdue_tasks or extend if needed.
  2. In app.py rendering loop, check overdue status and wrap task title with st.markdown(f"<span class='overdue'>...</span>", unsafe_allow_html=True).
  3. Add CSS for .overdue { color: red; font-weight: bold; } via st.markdown with <style>.
- Refactoring & Validation:
  - Extract overdue-check logic into is_task_overdue for test coverage.
  - Run tests and verify overdue styling appears.
Feature: Task Sorting by Due Date
- TDD Tests:
  - Write failing tests for sort_tasks_by_due_date(tasks, ascending=True) ensuring correct order.
  - Write UI tests for selecting sort order and rendering accordingly.
- Implementation Steps:
  1. Implement sort_tasks_by_due_date in tasks.py.
  2. Add a sort option in app.py sidebar (st.selectbox) for “Sort by Due Date: Ascending/Descending”.
  3. Apply sorting function before rendering tasks.
- Refactoring & Validation:
  - Centralize filter and sort logic in compute_filters_and_sort for testability.
  - Confirm tests pass and UI sorting works as expected.

Features Working

Behavior‑Driven Development (BDD)

Tooling: We chose pytest‑bdd so we could write Gherkin scenarios and step definitions in Python.
Feature File: tests/feature/add_task.feature describes 5 end‑to‑end behaviors:
1. Adding a task
2. Completing a task
3. Deleting a task
4. Filtering by category
5. Editing a task

Feature: Task management

  Scenario: Adding a new task
    Given I start with no tasks
    When I add a task with title "Buy milk" and description "Get 2% milk" and priority "High" and category "Personal" and due date "2025-05-01"
    Then the task list contains exactly 1 task
    And the task titled "Buy milk" has description "Get 2% milk"
    And the task titled "Buy milk" has priority "High"
    And the task titled "Buy milk" has category "Personal"
    And the task titled "Buy milk" has due date "2025-05-01"
    And the task titled "Buy milk" is not completed

  Scenario: Completing a task
    Given a task titled "Write report" with description "Draft v1" and priority "Medium" and category "Work" and due date "2025-06-01" exists
    When I mark the task "Write report" as complete
    Then the task "Write report" is marked completed

  Scenario: Deleting a task
    Given tasks titled "Old task" and "Keep task" exist
    When I delete the task "Old task"
    Then the task list does not contain "Old task"
    And the task list still contains "Keep task"

  Scenario: Filtering tasks by category
    Given tasks titled "A" (category "Work") and "B" (category "Personal") and "C" (category "Work") exist
    When I filter tasks by category "Work"
    Then only tasks titled "A" and "C" are visible

  Scenario: Editing a task
    Given I have an existing task titled "Draft" with description "Initial" and priority "Low" and category "School" and due date "2025-07-01" exists
    When I edit the task "Draft" changing title to "Final" and description to "Revised" and priority to "High" and category to "Work" and due date to "2025-07-15"
    Then the task list contains a task titled "Final"
    And that task has description "Revised"
    And that task has priority "High"
    And that task has category "Work"
    And that task has due date "2025-07-15"

Step Definitions: Implemented in tests/feature/steps/test_add_steps.py. We use fixtures to set up tmp_path + session_state, then call our core functions (handle_new_task, complete_task, delete_task, etc.).
Streamlit Integration:

  def show_bdd_button():
      if st.button("Run BDD Tests", key="bdd"):
          subprocess.run(["pytest","tests/feature","-q"])

Test Output

============================================== tests coverage ===============================================
_____________________________ coverage: platform darwin, python 3.13.1-final-0 ______________________________

Name              Stmts   Miss  Cover   Missing
-----------------------------------------------
src/__init__.py       0      0   100%
src/app.py          113     72    36%   25-32, 36-62, 89-91, 100, 103, 106-110, 138-165, 173-195, 203, 206-207, 210, 213, 216-217, 220, 349-350
src/tasks.py         59     20    66%   23-31, 71, 97, 110-111, 128-129, 140-141, 151-152, 159, 177, 180
-----------------------------------------------
TOTAL               172     92    47%
5 passed in 0.77s

What I’ve Learned Throughout This Process

Over the course of this assignment, I learned just how critical robust error handling and comprehensive test coverage are to software quality. Every bug— from the duplicate-ID issue to the double-click button glitch—taught me the importance of writing tests that not only confirm correct behavior but also anticipate edge cases.

Debugging felt like incredibly exhaustive detective work: each pytest failure message pointed me to a new hypothesis, and each fix inspired fresh tests to lock down the behavior. Setting up BDD scenarios with pytest‑bdd forced me to think in terms of user stories and workflows, not just individual functions, which was a powerful shift in perspective.

Balancing Streamlit’s reactive UI with pure‑logic functions drove home the value of separation of concerns—extracting business logic into src/tasks.py made unit and property‑based tests far more straightforward. I also saw how TDD accelerates design: writing failing tests for a feature before implementing it kept the scope focused and minimized regression.

Adding the “Run Selected Tests” and “Run BDD Tests” buttons in the app highlighted how developer ergonomics can be built into the product itself—empowering users (and myself) to validate the entire suite with a single click. Watching coverage climb from 25% to over 97% was immensely satisfying and underscored that true confidence in code comes from seeing it exercise every branch, every corner case.

Overall, this process taught me patience (sometimes the simplest typo can block a hundred lines of code), persistence, and pride in delivering a to‑do app that’s not just functional, but battle‑tested end to end.