feat(doctest): SSIM metric doctest

.github/workflows/docs.yml

@@ -61,3 +61,28 @@ jobs:
       - name: make linkcheck
         working-directory: ./docs/
         run: make linkcheck --jobs 2 SPHINXOPTS="--color -W"
+
+  doctest:
+    if: github.event_name == 'pull_request' || github.event_name == 'push'
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v2
+      - uses: actions/setup-python@v2
+        with:
+          python-version: 3.7
+
+      - run: sudo npm install katex@0.13.0 -g
+      - uses: actions/cache@v2
+        with:
+          path: ~/.cache/pip
+          key: pip-${{ hashFiles('requirements-dev.txt') }}-${{ hashFiles('docs/requirements.txt') }}
+
+      - name: Install docs deps
+        run: bash .github/workflows/install_docs_deps.sh
+
+      - name: make doctest
+        working-directory: ./docs/
+        run: |
+          make html SPHINXOPTS="--color -W"
+          make doctest
+          make coverage

docs/source/conf.py

@@ -339,3 +339,16 @@ def run(self):
     ("py:class", "torch.optim.lr_scheduler._LRScheduler"),
     ("py:class", "torch.utils.data.dataloader.DataLoader"),
 ]
+
+# doctest config
+doctest_global_setup = """
+import torch
+from torch import nn, optim
+
+from ignite.engine import *
+from ignite.handlers import *
+from ignite.metrics import *
+from ignite.utils import *
+
+manual_seed(666)
+"""

ignite/metrics/ssim.py

@@ -36,14 +36,22 @@ class SSIM(Metric):
         ``y_pred`` and ``y`` can be un-normalized or normalized image tensors. Depending on that, the user might need
         to adjust ``data_range``. ``y_pred`` and ``y`` should have the same shape.
 
-        .. code-block:: python
+        .. testcode::
 
             def process_function(engine, batch):
-                # ...
+                y_pred, y = batch
                 return y_pred, y
             engine = Engine(process_function)
             metric = SSIM(data_range=1.0)
-            metric.attach(engine, "ssim")
+            metric.attach(engine, 'ssim')
+            preds = torch.rand([4, 3, 16, 16])
+            target = preds * 0.75
+            state = engine.run([[preds, target]])
+            print(state.metrics['ssim'])
+
+        .. testoutput::
+
+            0.9218971...
 
     .. versionadded:: 0.4.2
     """