Add slides draft

property_based.py

@@ -0,0 +1,6 @@
+from hypothesis import given
+from hypothesis.strategies import floats, integers
+def divide(x, y): return x / y
+
+@given(integers(), integers())
+def test_divide(x, y): assert divide(x, y) < x

slides.md

@@ -0,0 +1,81 @@
+---
+title: Advanced Python Testing
+author: Abhishek Dasgupta
+date: 3 September 2024
+---
+
+# Advanced Python Testing
+
+Tired of waiting for network requests, writing millions of edge cases
+and copy pasting outputs for testing? This walkthrough is for you.
+
+Examples are in Python, but there are similar libraries available in
+other languages.
+
+https://github.com/abhidg/advanced-python-testing
+
+# Mocking
+
+- Test a code interface without requiring the underlying code to be run
+- Useful when testing the actual interface is
+  - time-consuming
+  - requires network access
+  - not possible in an automated way
+
+# Mocking in Python
+
+# Property based testing
+
+- Checks that a particular property is satisfied by the code
+- Normally one writes example based tests
+
+```python
+import pytest
+def fahrenheit_to_celsius(temp): return 5 * (temp - 32) / 9
+
+@pytest.mark.parametrize("f,c", [(212, 100), (32, 0)])
+def test_fahrenheit_to_celsius(f,c):
+    assert fahrenheit_to_celsius(f) == c
+```
+
+- Cons of example based testing is not being exhaustive enough
+- What is `fahrenheit_to_celsius(NaN)`?
+
+# Property based testing using `hypothesis`
+
+- Allows one to express properties without providing examples
+- Generates random examples and edge cases to test your code
+- Number of examples, and search strategies can be customised
+
+# Property based testing using `hypothesis`
+
+```python
+from hypothesis import given
+from hypothesis.strategies import floats, integers
+def divide(x, y): return x / y
+
+@given(integers(), integers())
+def test_divide(x, y): assert divide(x, y) < x
+```
+
+You can also find counterexamples - assert the inverse of a property
+that you know to be true to be only in very few cases, and let
+`hypothesis` search through the parameter space to find a failing case.
+
+# Snapshot based testing
+
+- Captures a _snapshot_ of a function or command and tests that it does not change
+- Useful for testing outputs, data transformations
+- Can be emulated manually (save output to disk, read and test that it is the same), but snapshot testing libraries make this easier
+
+# Snapshot based testing using `syrupy`
+
+- `syrupy` takes snapshots
+- easy to use
+
+# Walkthrough: finding the sun!
+
+- **Problem**: Find a block of N hours when it will be sunny and clear, given a location.
+- We will try to find a block of sun of two hours in Newcastle over the next week
+- OpenMeteo has a nice free API
+- Steps: fetch data, process and find the sun block