OwTSNE: Don't clear computed values at invalidation, set flags instead

Author: pavlin-policar

Orange/widgets/unsupervised/owtsne.py

@@ -230,6 +230,24 @@ def update_coordinates(self):
             self.view_box.setAspectLocked(True, 1)
 
 
+class invalidated:
+    pca_projection = affinities = tsne_embedding = False
+
+    def __set__(self, instance, value):
+        # `self._invalidate = True` should invalidate everything
+        self.pca_projection = self.affinities = self.tsne_embedding = value
+
+    def __bool__(self):
+        # If any of the values are invalidated, this should return true
+        return self.pca_projection or self.affinities or self.tsne_embedding
+
+    def __str__(self):
+        return "%s(%s)" % (self.__class__.__name__, ", ".join(
+            "=".join([k, str(getattr(self, k))])
+            for k in ["pca_projection", "affinities", "tsne_embedding"]
+        ))
+
+
 class OWtSNE(OWDataProjectionWidget, ConcurrentWidgetMixin):
     name = "t-SNE"
     description = "Two-dimensional data projection with t-SNE."
@@ -250,6 +268,11 @@ class OWtSNE(OWDataProjectionWidget, ConcurrentWidgetMixin):
 
     left_side_scrolling = True
 
+    # Use `invalidated` descriptor so we don't break the usage of
+    # `_invalidated` in `OWDataProjectionWidget`, but still allow finer control
+    # over which parts of the embedding to invalidate
+    _invalidated = invalidated()
+
     class Information(OWDataProjectionWidget.Information):
         modified = Msg("The parameter settings have been changed. Press "
                        "\"Start\" to rerun with the new settings.")
@@ -323,21 +346,19 @@ def _multiscale_changed(self):
         self._invalidate_affinities()
 
     def _invalidate_pca_projection(self):
-        self.pca_projection = None
-        self.initialization = None
+        self._invalidated.pca_projection = True
         self._invalidate_affinities()
 
     def _invalidate_affinities(self):
-        self.affinities = None
+        self._invalidated.affinities = True
         self._invalidate_tsne_embedding()
 
     def _invalidate_tsne_embedding(self):
-        self.iterations_done = 0
-        self.tsne_embedding = None
-        self._invalidate_output()
+        self._invalidated.tsne_embedding = True
+        self._stop_running_task()
         self._set_modified(True)
 
-    def _invalidate_output(self):
+    def _stop_running_task(self):
         self.cancel()
         self.run_button.setText("Start")
 
@@ -397,8 +418,15 @@ def _toggle_run(self):
         else:
             self.run()
 
-    def set_data(self, data: Table):
-        super().set_data(data)
+    def handleNewSignals(self):
+        # We don't bother with the granular invalidation flags because
+        # `super().handleNewSignals` will just set all of them to False or will
+        # do nothing. However, it's important we remember its state because we
+        # won't call `run` if needed. `run` also relies on the state of
+        # `_invalidated` to properly set the intermediate values to None
+        prev_invalidated = bool(self._invalidated)
+        super().handleNewSignals()
+        self._invalidated = prev_invalidated
 
         if self._invalidated:
             self.run()
@@ -439,7 +467,17 @@ def enable_controls(self):
         self.controls.perplexity.setDisabled(self.multiscale)
 
     def run(self):
+        # Reset invalidated values as indicated by the flags
+        if self._invalidated.pca_projection:
+            self.pca_projection = None
+        if self._invalidated.affinities:
+            self.affinities = None
+        if self._invalidated.tsne_embedding:
+            self.iterations_done = 0
+            self.tsne_embedding = None
+
         self._set_modified(False)
+        self._invalidated = False
 
         # When the data is invalid, it is set to `None` and an error is set,
         # therefore it would be erroneous to clear the error here

Orange/widgets/unsupervised/tests/test_owtsne.py

@@ -250,6 +250,58 @@ def test_global_structure_info_msg_persists_if_data_is_reloaded(self):
             "The information message was cleared after data was reloaded"
         )
 
+    def test_invalidation_flow(self):
+        w = self.widget
+        # Setup widget: send data to input with global structure "off", then
+        # set global structure "on" (after the embedding is computed)
+        w.controls.multiscale.setChecked(False)
+        self.send_signal(w.Inputs.data, self.data)
+        self.wait_until_stop_blocking()
+        self.assertFalse(self.widget.Information.modified.is_shown())
+        # All the embedding components should computed
+        self.assertIsNotNone(w.pca_projection)
+        self.assertIsNotNone(w.affinities)
+        self.assertIsNotNone(w.tsne_embedding)
+        # All the invalidation flags should be set to false
+        self.assertFalse(w._invalidated.pca_projection)
+        self.assertFalse(w._invalidated.affinities)
+        self.assertFalse(w._invalidated.tsne_embedding)
+
+        # Trigger invalidation
+        w.controls.multiscale.setChecked(True)
+        self.assertTrue(self.widget.Information.modified.is_shown())
+        # Setting `multiscale` to true should set the invalidate flags for
+        # the affinities and embedding, but not the pca_projection
+        self.assertFalse(w._invalidated.pca_projection)
+        self.assertTrue(w._invalidated.affinities)
+        self.assertTrue(w._invalidated.tsne_embedding)
+
+        # The flags should now be set, but the embedding should still be
+        # available when selecting a subset of data and such
+        self.assertIsNotNone(w.pca_projection)
+        self.assertIsNotNone(w.affinities)
+        self.assertIsNotNone(w.tsne_embedding)
+
+        # We should still be able to send a data subset to the input and have
+        # the points be highlighted
+        self.send_signal(w.Inputs.data_subset, self.data[:10])
+        self.wait_until_stop_blocking()
+        subset = [brush.color().name() == "#46befa" for brush in
+                  w.graph.scatterplot_item.data["brush"][:10]]
+        other = [brush.color().name() == "#000000" for brush in
+                 w.graph.scatterplot_item.data["brush"][10:]]
+        self.assertTrue(all(subset))
+        self.assertTrue(all(other))
+
+        # Clear the data subset
+        self.send_signal(w.Inputs.data_subset, None)
+
+        # Run the optimization
+        self.widget.run_button.clicked.emit()
+        self.wait_until_stop_blocking()
+        # All of the inavalidation flags should have been cleared
+        self.assertFalse(w._invalidated)
+
 
 class TestTSNERunner(unittest.TestCase):
     @classmethod