Mobile Testing Automation

when you do exploratory test follow below steps:

1. dump xml for every new ui page by uiautomator dump, name as step1_dump.xml, step2_dump.xml, etc
2. detect all clickable element
3. if click successful, mark the click position with `magick input.png -fill "rgba(255,0,0,0.5)" -draw "circle 100,100 125,100" output.png`
4. after finished the session. output a document `android_exploratory_test_report.md` with use cases. each step contain screenshot with click area highlighted.
5. IMPORTANT: ALL screenshots in android_exploratory_test_report.md MUST use <img> tag format with width="120"
6. sample markdown step:

   <img src="step1.png" alt="step1: home page" width="120">
   **actions**: launch app
   **detected element**:
[step1_dump.xml](..%2F..%step1_dump.xml)
- main menu button
- category button

when you generate function test script follow below steps:

1. read android_exploratory_test_report.md, and understand test cases
2. create `tests_suite` folder with standard pytest structure:

• `conftest.py` for shared fixtures and setup
• `test_*.py` files for each test case
• `__init__.py` to make it a package

1. IMPORTANT: use consistent naming convention:

• test files: `test_[feature_name].py`
• test functions: `test_[action_description]()`
• fixtures: `setup_app()`, `teardown_app()`

1. MANDATORY: add checkpoint after every action using `assert exists()` or `wait()`
2. structure each test function as:

   def test_feature_name():
       # setup
       start_app(package_name)

       # action 1
       touch(element)
       assert exists(expected_element), "checkpoint failed"

       # action 2
       touch(next_element)
       wait(next_expected_element, timeout=10)

       # cleanup
       stop_app(package_name)

1. generate `pytest.ini` configuration file
2. run `pytest tests_suite/` to validate all tests

when you fix a existing test script, follow below steps:

1. navigate to the specific UI before the step which not work
2. try to do exploratory on the UI, every step do screenshot and element dump.
3. after the exploratory verify if it is working.
4. fix the test script.