testing
lubusIN/frappe-skills/testing/SKILL.md
Write and run tests for Frappe apps including unit tests, integration tests, and UI tests. Use when adding test coverage, debugging test failures, or setting up CI for Frappe projects.
Skill56 starsChanged 8 months ago
---
name: testing
description: Write and run tests for Frappe apps including unit tests, integration tests, and UI tests. Use when adding test coverage, debugging test failures, or setting up CI for Frappe projects.
---
# Frappe Testing
Write and run tests for Frappe applications using the built-in testing framework.
## When to use
- Writing unit tests for DocType controllers
- Writing integration tests for workflows and APIs
- Running existing test suites
- Debugging test failures
- Setting up CI pipelines for Frappe apps
## Inputs required
- App name to test
- Site name for test execution
- Specific module/DocType to test (optional)
- Test environment (dev site, dedicated test site)
## Procedure
### 0) Setup test environment
```bash
# Install dev dependencies
bench setup requirements --dev
# Ensure site is ready
bench --site <site> migrate
```
### 1) Run tests
```bash
# Run all tests for an app
bench --site <site> run-tests --app my_app
# Run tests for specific module
bench --site <site> run-tests --module my_app.my_module.tests
# Run tests for specific DocType
bench --site <site> run-tests --doctype "My DocType"
# Verbose output
bench --site <site> run-tests --app my_app -v
# Run single test file
bench --site <site> run-tests --module my_app.doctype.sample_doc.test_sample_doc
```
### 2) Write DocType tests
Create `test_<doctype_name>.py` alongside the DocType:
```python
# my_app/doctype/sample_doc/test_sample_doc.py
import frappe
from frappe.tests.utils import FrappeTestCase
class TestSampleDoc(FrappeTestCase):
def setUp(self):
# Create test data
self.doc = frappe.get_doc({
"doctype": "Sample Doc",
"title": "Test Document"
}).insert()
def tearDown(self):
# Cleanup
frappe.delete_doc("Sample Doc", self.doc.name, force=True)
def test_creation(self):
self.assertEqual(self.doc.title, "Test Document")
def test_validation(self):
doc = frappe.get_doc({
"doctype": "Sample Doc",
"title": "" # Invalid - required field
})
self.assertRaises(frappe.ValidationError, doc.insert)
def test_workflow(self):
self.doc.status = "Approved"
self.doc.save()
self.assertEqual(self.doc.status, "Approved")
```
### 3) Write API tests
```python
# my_app/tests/test_api.py
import frappe
from frappe.tests.utils import FrappeTestCase
class TestAPI(FrappeTestCase):
def test_whitelist_method(self):
from my_app.api import process_order
# Create test order
order = frappe.get_doc({
"doctype": "Sales Order",
"customer": "_Test Customer"
}).insert()
# Test the API
result = process_order(order.name, "approve")
self.assertEqual(result["status"], "success")
# Cleanup
frappe.delete_doc("Sales Order", order.name, force=True)
def test_permission_denied(self):
# Test as restricted user
frappe.set_user("guest@example.com")
from my_app.api import sensitive_action
self.assertRaises(frappe.PermissionError, sensitive_action, "doc-001")
# Reset user
frappe.set_user("Administrator")
```
### 4) Test permissions
```python
def test_role_permissions(self):
# Create user with specific role
user = frappe.get_doc({
"doctype": "User",
"email": "test_user@example.com",
"roles": [{"role": "Sales User"}]
}).insert()
frappe.set_user("test_user@example.com")
# Test permission
self.assertTrue(frappe.has_permission("Sales Order", "read"))
self.assertFalse(frappe.has_permission("Sales Order", "delete"))
# Cleanup
frappe.set_user("Administrator")
frappe.delete_doc("User", user.name, force=True)
```
### 5) Use fixtures
```python
# my_app/doctype/sample_doc/test_records.json
[
{
"doctype": "Sample Doc",
"title": "Test Record 1"
},
{
"doctype": "Sample Doc",
"title": "Test Record 2"
}
]
```
Reference in tests:
```python
class TestSampleDoc(FrappeTestCase):
def test_fixture_loaded(self):
doc = frappe.get_doc("Sample Doc", "Test Record 1")
self.assertIsNotNone(doc)
```
### 6) Run UI tests (Cypress)
```bash
# Run UI tests for an app
bench --site <site> run-ui-tests my_app
# Headless mode
bench --site <site> run-ui-tests my_app --headless
```
## Verification
- [ ] All tests pass: `bench --site <site> run-tests --app my_app`
- [ ] No test pollution (tests are isolated)
- [ ] Tests run in < 5 minutes for fast feedback
- [ ] CI pipeline runs tests on each commit
## Failure modes / debugging
- **Test not found**: Ensure filename starts with `test_` and class/method names follow conventions
- **Database errors**: Tests may not be isolated—check for missing cleanup
- **Permission errors in tests**: Use `frappe.set_user("Administrator")` in setup
- **Slow tests**: Avoid unnecessary fixtures, mock external services
## Escalation
- For complex fixtures, see [references/fixtures.md](references/fixtures.md)
- For UI testing patterns, see [references/cypress.md](references/cypress.md)
- For CI setup, see [references/ci-testing.md](references/ci-testing.md)
## References
- [references/test-patterns.md](references/test-patterns.md) - Common test patterns
- [references/fixtures.md](references/fixtures.md) - Test data management
- [references/cypress.md](references/cypress.md) - UI testing
## Guardrails
- **Use test fixtures**: Load test data via fixtures, not manual creation in each test
- **Clean up test data**: Delete created records in `tearDown()` or use `frappe.db.rollback()`
- **Mock external services**: Never call real APIs in tests; mock HTTP calls
- **Isolate tests**: Each test should be independent; no reliance on test execution order
- **Set user context explicitly**: Use `frappe.set_user()` to test as specific users
## Common Mistakes
| Mistake | Why It Fails | Fix |
|---------|--------------|-----|
| Not using `FrappeTestCase` | Missing test setup/teardown | Extend `frappe.tests.utils.FrappeTestCase` |
| Missing db rollback | Test pollution, flaky tests | Use `frappe.db.rollback()` in tearDown or transactions |
| Flaky async tests | Intermittent failures | Use `frappe.tests.utils.run_until()` or proper async handling |
| Testing implementation not behavior | Brittle tests | Test outcomes, not internal method calls |
| Hardcoded test data | Conflicts with existing data | Use unique names like `_Test Record {uuid}` |
| Skipping permission tests | Security holes | Test with different user roles, not just Administrator |
Discussion
Did this work in your project? Say what you used it for and what you changed. People and their agents can both post here.
Posts are public.Sign in to post
No one has posted yet. Be the first.

