CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

trail-sense-android-tests

Add Trail Sense UI automation tests using AutomationLibrary and ToolTestBase for tool workflows or androidTests.

66

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Passed

No findings from the security scan

SKILL.md
Quality
Evals
Security

Trail Sense Android Tests

Create UI automation tests for Trail Sense tools using AutomationLibrary and ToolTestBase.

Test Structure

package com.kylecorry.trail_sense.tools.<toolname>

import com.kylecorry.trail_sense.R
import com.kylecorry.trail_sense.test_utils.AutomationLibrary.*
import com.kylecorry.trail_sense.test_utils.TestUtils
import com.kylecorry.trail_sense.test_utils.TestUtils.back
import com.kylecorry.trail_sense.test_utils.TestUtils.clickListItemMenu
import com.kylecorry.trail_sense.test_utils.ToolTestBase
import com.kylecorry.trail_sense.test_utils.views.*
import com.kylecorry.trail_sense.tools.tools.infrastructure.Tools
import org.junit.Test

class Tool<Name>Test : ToolTestBase(Tools.<TOOL_ID>) {

    @Test
    fun verifyBasicFunctionality() {
        hasText(R.id.title, string(R.string.tool_title))

        canCreateItem()
        canEditItem()
        canDeleteItem()
        verifyQuickAction()
    }

    private fun canCreateItem() { /* ... */ }
    private fun canEditItem() { /* ... */ }
    private fun canDeleteItem() { /* ... */ }
    private fun verifyQuickAction() { /* ... */ }
}

Location: app/src/androidTest/java/com/kylecorry/trail_sense/tools/<toolname>/Tool<Name>Test.kt

Workflow

  1. Inspect the target tool and nearby Tool*Test.kt files. This step is complete when the tool id, package, primary screen title, and existing test patterns are known.
  2. Add or update the test class under the location above. This step is complete when the test covers the requested workflow and uses stable selectors from the strategy below.
  3. Run a focused emulator integration test when an emulator is available. This step is complete when the requested test passes or the blocking reason is reported.

Selection Strategy: Text Over IDs

Prefer text-based selection for most interactions. Use IDs only when necessary.

Use Text For

// Clicking tabs, buttons, menu items, options
click(string(R.string.distance))
click(string(R.string.delete))
click("High")
click("Test Group")

// Verifying text anywhere on screen
hasText(string(R.string.no_paths))
hasText("Test Path")

// Dialog inputs (by label/hint)
input(string(R.string.name), "My Item")
input(string(R.string.distance), "1.0")

// Checkbox state by label
isChecked(string(R.string.tide_clock))

Use IDs Only For

// Title bars (for verification)
hasText(R.id.paths_title, string(R.string.paths))
hasText(R.id.tide_title, "Tide 1")

// Add/play buttons (no text label)
click(R.id.add_btn)
click(R.id.play_btn)

// Specific input fields
input(R.id.searchbox, "query")
input(R.id.tide_name, "Tide 1")
input(R.id.utm, "42, -72")

// Result/data views
hasText(R.id.result, "3.2808 ft")
hasText(R.id.total_percent_packed, "50% packed")

// Charts and special views
isVisible(R.id.chart)
scrollToEnd(R.id.scroll_view)

// Toolbar buttons
click(toolbarButton(R.id.paths_title, Side.Right))

Andromeda List Item IDs

List items use Andromeda library IDs:

click(com.kylecorry.andromeda.views.R.id.title)
hasText(com.kylecorry.andromeda.views.R.id.title, "Item Name")
hasText(com.kylecorry.andromeda.views.R.id.description, "Details")
click(com.kylecorry.andromeda.views.R.id.checkbox)
click(com.kylecorry.andromeda.views.R.id.trailing_icon_btn)
click(com.kylecorry.andromeda.views.R.id.menu_btn)

Common Patterns

Create/Edit/Delete Flow

private fun canCreateItem() {
    click(R.id.add_btn)
    click(string(R.string.new_item))
    input(string(R.string.name), "Test Item")
    clickOk()
    hasText("Test Item")
}

private fun canEditItem() {
    clickListItemMenu(string(R.string.edit))
    input("Test Item", "Test Item 2")
    clickOk()
    hasText("Test Item 2")
}

private fun canDeleteItem() {
    clickListItemMenu(string(R.string.delete))
    clickOk()
    not { hasText("Test Item 2", waitForTime = 0) }
}

Toolbar Menu

click(toolbarButton(R.id.title, Side.Right))
click(string(R.string.export))

Quick Actions

private fun verifyQuickAction() {
    TestUtils.openQuickActions()
    click(quickAction(Tools.QUICK_ACTION_ID))
    // Verify action occurred
    TestUtils.closeQuickActions()
}

Optional Elements

optional {
    hasText(string(R.string.disclaimer))
    clickOk()
}

Navigate Back

backUntil { isVisible(R.id.paths_title, waitForTime = 1000) }

Key Functions

FunctionUsage
click(text)Click by text (contains match)
click(text, exact = true)Click by exact text
click(R.id.x)Click by ID
hasText(text)Verify text on screen
hasText(R.id.x, text)Verify text in view
hasText(R.id.x, Regex(...))Verify regex pattern
input(R.id.x, text)Enter text by ID
input(label, text)Enter text by label
clickOk()Click OK in dialogs
clickListItemMenu(label)Click list item overflow menu
optional { }Ignore failures
not { }Assert action fails
string(R.string.x)Get string resource
scrollUntil { }Scroll until condition
backUntil { }Press back until condition

Deterministic Tests: Seed Data and Pin the Location

Emulator sensors and the clock are not predictable, so regex-only assertions (\d+ ft) can't catch wrong values. Prefer known answers:

  • Pin the location: ToolTestBase(Tools.X, Coordinate(42.03, -71.97))
  • Seed data before the tool opens by overriding seedData(), using TestData (addBeacon, addPathFromGpx, addPressureTrend, ...). Guard with assumeFalse(isStagingBuild) since staging builds can't be seeded.
  • To use different data per test, call relaunchTool { TestData.add...() } (the tool is closed while seeding, so background updates can't race it).
  • Compute expected values independently of the app (a published value, or a quick script from the GPX/coordinates), not by calling the code under test.
  • Use fixed dates (TestUtils.pickDate) instead of "today".

Helpers for checking values: hasDataPoint(value, label) (a stat tile) and hasTextsInOrder(id, texts) (list ordering). hasText(String) can't match across a leading newline, use a Regex for multi-line values.

Canvas content (the map) can't be found by UiAutomator. Convert a coordinate to a screen pixel with the view's projection inside scenario.onActivity and tap that point (see ToolMapTest).

Running Tests

./scripts/run-emulator-integration-tests.sh
./scripts/run-emulator-integration-tests.sh com.kylecorry.trail_sense.tools.notes.ToolNotesTest 180

The filter only supports a single class or method (a comma separated list runs just the first class), so run each class separately. The script defaults to a 1800 second timeout. Most individual integration tests should finish in 60 to 180 seconds, so pass a shorter timeout for focused runs when practical.

Source Files

For complete API details and additional functions, read:

  • app/src/androidTest/java/com/kylecorry/trail_sense/test_utils/AutomationLibrary.kt
  • app/src/androidTest/java/com/kylecorry/trail_sense/test_utils/TestUtils.kt
  • app/src/androidTest/java/com/kylecorry/trail_sense/test_utils/views/

For example tests, see existing tool tests in:

  • app/src/androidTest/java/com/kylecorry/trail_sense/tools/*/Tool*Test.kt
Repository

kylecorry31/Trail-Sense

Path

.agents/skills/trail-sense-android-tests/SKILL.md

Last updated
First committed

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.