Canonicalize squeeze out of reshape and specialize back

Author: ricardoV94

pytensor/tensor/rewriting/shape.py

@@ -36,6 +36,7 @@
     register_useless,
     topo_constant_folding,
 )
+from pytensor.tensor.rewriting.elemwise import apply_local_dimshuffle_lift
 from pytensor.tensor.shape import (
     Reshape,
     Shape,
@@ -757,40 +758,36 @@ def apply(self, fgraph):
 pytensor.compile.mode.optdb.register("UnShapeOpt", UnShapeOptimizer(), position=10)
 
 
+@register_useless
 @register_canonicalize
 @node_rewriter([Reshape])
-def local_useless_dimshuffle_in_reshape(fgraph, node):
+def local_useless_expand_dims_in_reshape(fgraph, node):
     """
-    Removes useless DimShuffle operation inside Reshape:
-
-      reshape(vector.dimshuffle('x', 0), shp) => reshape(vector, shp)
-      reshape(matrix.dimshuffle('x', 0, 'x', 1), shp) => reshape(matrix, shp)
-      reshape(row.dimshuffle(1, 'x'), shp) => reshape(row, shp)
-      reshape(col.dimshuffle(0), shp) => reshape(col, shp)
+    Removes useless expand_dims `DimShuffle` operations inside Reshape:
+      reshape(expand_dims(vector, axis=0), shp) => reshape(vector, shp)
+      reshape(expand_dims(matrix, axis=(0, 2), shp) => reshape(matrix, shp)
 
+    Implicit (and useless) squeezes are kept in the graph, as they are
+    part of the canonical form of the graph.
     """
-    dimshuffled_x, new_shape = node.inputs
+    expanded_x, new_shape = node.inputs
 
     if not (
-        dimshuffled_x.owner is not None
-        and isinstance(dimshuffled_x.owner.op, DimShuffle)
+        expanded_x.owner is not None
+        and isinstance(expanded_x.owner.op, DimShuffle)
+        and expanded_x.owner.op.augment
     ):
         return False
 
-    [inp] = dimshuffled_x.owner.inputs
-    new_order = dimshuffled_x.owner.op.new_order
-    new_order_of_nonbroadcast = []
-    for i, s in zip(new_order, node.inputs[0].type.shape, strict=True):
-        if s != 1:
-            new_order_of_nonbroadcast.append(i)
-    no_change_in_order = all(
-        new_order_of_nonbroadcast[i] <= new_order_of_nonbroadcast[i + 1]
-        for i in range(len(new_order_of_nonbroadcast) - 1)
-    )
-    if no_change_in_order:
-        ret = inp.reshape(new_shape)
-        copy_stack_trace(node.outputs[0], ret)
-        return [ret]
+    [x] = expanded_x.owner.inputs
+
+    new_order = tuple(o for o in expanded_x.owner.op.new_order if o != "x")
+    if new_order != tuple(range(x.type.ndim)):
+        x = x.dimshuffle(new_order)
+
+    new_reshaped_x = x.reshape(new_shape)
+    copy_stack_trace(node.outputs[0], new_reshaped_x)
+    return [new_reshaped_x]
 
 
 @register_canonicalize("shape_unsafe")
@@ -920,10 +917,10 @@ def local_useless_reshape(fgraph, node):
 
     shape_feature = getattr(fgraph, "shape_feature", None)
 
-    # Match Reshape(x, [x.shape[0], ..., x.shape[-1]]), accounting for -1
-    # or cases where all but one dimension are provably preserved
+    # Match case where at least (n-1) entries correspond to the original shape:
+    # Reshape(x, [x.shape[0], ..., x.shape[-1]]), or Reshape(x, [x.shape[0], y, x.shape[2], ... x.shape[-1]])
+    # Where y can be -1 or anything with an unknown value, since the only valid reshape is still a no reshape.
     output_shape_is = _unpack_shape_vector(output_shape)
-
     nb_m1 = 0
     shape_match = [False] * inp.type.ndim
     for dim in range(inp.type.ndim):
@@ -935,48 +932,136 @@ def local_useless_reshape(fgraph, node):
             nb_m1 += 1
 
     if nb_m1 <= 1 and all(shape_match):
-        return [inp]
+        return [inp]  # This is provably correct
 
     # There is one missing match, but all other dimensions match
+    # Such as x.type.shape == (3, 5, None) and output_shape == (3, 5, y)
     if (nb_m1 == 0) and (shape_match.count(False) == 1):
-        return [inp]
+        return [inp]  # This could mask a shape error
 
     return False
 
 
-@register_canonicalize
+@register_canonicalize("shape_unsafe")
 @node_rewriter([Reshape])
 def local_reshape_to_dimshuffle(fgraph, node):
-    r"""Replace broadcastable dimensions in `Reshape` nodes with `DimShuffle`\s.
+    r"""Remove `Reshape` operations over length-1 (broadcastable) dimensions.
 
-    The goal is to avoid using `Reshape` to add or remove broadcastable
-    dimensions, and to use `DimShuffle` instead, since `DimShuffle`\s can
-    cancel out and/or be removed later on.
+    It's always valid to squeeze an input before doing the same reshape operation.
+    Equivalently, it's always valid to remove `1` entries from the reshape shape
+    and replace them by an expand_dims after the rewritten reshape operation.
+
+    We chose to canonicalize the graph in this way as it allows isolating
+    operations that are unique to the reshaping operation (mixing dimensions)
+    from those that can be more legibly encoded by DimShuffle (squeeze and expand_dims).
+    This can allow further simplifications by other rewrites that target
+    DimShuffle but not Reshape, as well as facilitate the removal of useless reshape operations.
 
     For example:
-        - reshape(x, (1, n)) -> DimShuffle{x,0}(Reshape(x, (n,))
-        - reshape(x, (1, m, 1, n, 1, 1)) -> DimShuffle{x,0,x,1,x,x}(Reshape(x, (m, n)))
+        - reshape(col, (m, n)) -> reshape(squeeze(col, axis=1), (m, n))
+        - reshape(col, (1, m, n)) -> expand_dims(reshape(squeeze(col, axis=1), (m, n)), axis=0)
+        - reshape(x, (1, m, 1, n, 1, 1)) -> expand_dims(reshape(x, (m, n)), axis=(0, 2, 4, 5))
+
     """
     inp, output_shape = node.inputs
     [output] = node.outputs
 
-    unpacked_shape = _unpack_shape_vector(output_shape)
-    expand_axes = []
-    new_output_shape = []
-    for i, dim in enumerate(unpacked_shape):
-        if isinstance(dim, Constant) and dim.data == 1:
-            expand_axes.append(i)
-        else:
-            new_output_shape.append(dim)
+    # Remove any broadcastable dimensions from the input
+    squeeze_axes = [i for i, bcast in enumerate(inp.type.broadcastable) if bcast]
+
+    # Trivial case, all dimensions of input/output are known to be broadcastable:
+    # there's nothing to reshape
+    if all(inp.type.broadcastable) or all(output.type.broadcastable):
+        new_output_shape = []
+        expand_axes = tuple(range(output.type.ndim))
+
+    else:
+        unpacked_shape = _unpack_shape_vector(output_shape)
+        new_output_shape = []
+        expand_axes = []
+        for i, dim_length in enumerate(unpacked_shape):
+            if isinstance(dim_length, Constant) and (
+                dim_length.data == 1
+                # -1 can be an implicit expand_dims, but it's tricky to prove
+                # as we would need to check whether all other dimensions
+                # already explain the full size of the array.
+                # Example: np.zeros((2, 2, 2)).reshape((8, -1))
+                # We rely on the output static shape which will already have figured
+                # it out for some (but not all) cases
+                or (dim_length.data == -1 and output.type.shape[i] == 1)
+            ):
+                expand_axes.append(i)
+            else:
+                new_output_shape.append(dim_length)
+
+    if squeeze_axes or expand_axes:
+        new_out = inp.squeeze(squeeze_axes)
+
+        if new_output_shape:
+            new_out = new_out.reshape(new_output_shape)
+            copy_stack_trace(output, new_out)
+
+        new_out = expand_dims(new_out, expand_axes)
+
+        if not new_output_shape:
+            # Eagerly merge consecutive squeeze and expand_dims
+            new_out = apply_local_dimshuffle_lift(fgraph, new_out)
 
-    if len(new_output_shape) != output.type.ndim:
-        inner = inp.reshape(new_output_shape)
-        copy_stack_trace(output, inner)
-        new_out = expand_dims(inner, expand_axes)
         copy_stack_trace(output, new_out)
         return [new_out]
 
 
+@register_specialize
+@node_rewriter([Reshape])
+def local_fuse_squeeze_reshape(fgraph, node):
+    r"""If there is a squeeze right before a reshape, merge them.
+
+    This undoes the effect of `local_reshape_to_dimshuffle` that is applied during canonicalization.
+    """
+    x, new_shape = node.inputs
+
+    if (
+        x.owner is not None
+        and isinstance(x.owner.op, DimShuffle)
+        and x.owner.op.is_squeeze
+    ):
+        # A reshape can always subsume a squeeze.
+        x = x.owner.inputs[0]
+        return [x.reshape(new_shape)]
+
+
+@register_specialize
+@node_rewriter([DimShuffle])
+def local_fuse_expand_dims_reshape(fgraph, node):
+    r"""If there is an expand_dims right after a reshape, merge them.
+
+    This undoes the effect of `local_reshape_to_dimshuffle` that is applied during canonicalization.
+    """
+    if not node.op.is_expand_dims:
+        return None
+
+    reshaped_x = node.inputs[0]
+
+    if not (reshaped_x.owner and isinstance(reshaped_x.owner.op, Reshape)):
+        return None
+
+    if len(fgraph.clients[reshaped_x]) > 1:
+        # The reshape is used elsewhere, don't fuse as it can sometimes require a copy.
+        # Example: `x = pt.matrix(); y = x.T.reshape(-1); out = y[: None] * y[None, :]`
+        return None
+
+    x, new_shape = reshaped_x.owner.inputs
+
+    # Add expand_dims to shape
+    new_shape = list(_unpack_shape_vector(new_shape))
+    for i in node.op.augment:
+        new_shape.insert(i, 1)
+
+    new_reshaped_x = x.reshape(new_shape)
+    copy_stack_trace(node.outputs[0], new_reshaped_x)
+    return [new_reshaped_x]
+
+
 @register_canonicalize
 @register_specialize
 @node_rewriter([Reshape])

tests/tensor/rewriting/test_basic.py

@@ -332,7 +332,6 @@ def test_basic_tile(self):
 
         mode = rewrite_mode.including(
             "local_dimshuffle_lift",
-            "local_useless_dimshuffle_in_reshape",
             "local_alloc_sink_dimshuffle",
         )
         f = function([x], [y], mode=mode)

tests/tensor/rewriting/test_elemwise.py

@@ -56,7 +56,10 @@
 from pytensor.tensor.math import round as pt_round
 from pytensor.tensor.math import sum as pt_sum
 from pytensor.tensor.rewriting.elemwise import FusionOptimizer, local_dimshuffle_lift
-from pytensor.tensor.rewriting.shape import local_useless_dimshuffle_in_reshape
+from pytensor.tensor.rewriting.shape import (
+    local_fuse_squeeze_reshape,
+    local_useless_expand_dims_in_reshape,
+)
 from pytensor.tensor.shape import reshape
 from pytensor.tensor.type import (
     TensorType,
@@ -182,7 +185,7 @@ def test_dimshuffle_lift_multi_out_elemwise(self):
         assert not local_dimshuffle_lift.transform(g, g.outputs[0].owner)
 
 
-def test_local_useless_dimshuffle_in_reshape():
+def test_local_useless_expand_dims_in_reshape():
     vec = TensorType(dtype="float64", shape=(None,))("vector")
     mat = TensorType(dtype="float64", shape=(None, None))("mat")
     row = TensorType(dtype="float64", shape=(1, None))("row")
@@ -204,7 +207,11 @@ def test_local_useless_dimshuffle_in_reshape():
         clone=False,
     )
     assert len(g.apply_nodes) == 4 * 3
-    useless_dimshuffle_in_reshape = out2in(local_useless_dimshuffle_in_reshape)
+    useless_dimshuffle_in_reshape = out2in(
+        local_useless_expand_dims_in_reshape,
+        # Useless squeeze in reshape is not a canonicalization anymore
+        local_fuse_squeeze_reshape,
+    )
     useless_dimshuffle_in_reshape.rewrite(g)
     assert equal_computations(
         g.outputs,
@@ -218,15 +225,12 @@ def test_local_useless_dimshuffle_in_reshape():
     # Check stacktrace was copied over correctly after rewrite was applied
     assert check_stack_trace(g, ops_to_check="all")
 
-    # Check that the rewrite does not get applied when the order
-    # of dimensions has changed.
+    # Check that the rewrite does not mess meaningful transpositions before the reshape
     reshape_dimshuffle_mat2 = reshape(mat.dimshuffle("x", 1, "x", 0), mat.shape)
     h = FunctionGraph([mat], [reshape_dimshuffle_mat2], clone=False)
     assert len(h.apply_nodes) == 3
     useless_dimshuffle_in_reshape.rewrite(h)
-    assert equal_computations(
-        h.outputs, [reshape(mat.dimshuffle("x", 1, "x", 0), mat.shape)]
-    )
+    assert equal_computations(h.outputs, [reshape(mat.dimshuffle(1, 0), mat.shape)])
 
 
 class TestFusion:

tests/tensor/rewriting/test_shape.py

@@ -6,7 +6,7 @@
 import pytensor.tensor as pt
 from pytensor import shared
 from pytensor.compile.function import function
-from pytensor.compile.mode import get_default_mode, get_mode
+from pytensor.compile.mode import Mode, get_default_mode, get_mode
 from pytensor.compile.ops import deep_copy_op
 from pytensor.configdefaults import config
 from pytensor.graph.basic import Apply, Variable, equal_computations
@@ -426,6 +426,60 @@ def test_basic(self):
 
         assert check_stack_trace(g, ops_to_check=(DimShuffle, Reshape))
 
+    def test_expand_dims(self):
+        x = pt.scalar()
+        # This reshape does an implicit expand_dims
+        out = x.reshape((1, -1))
+        assert isinstance(out.owner.op, Reshape)
+        new_out = rewrite_graph(out, include=("canonicalize",))
+        assert equal_computations([new_out], [pt.expand_dims(x, (0, 1))])
+
+    def test_squeeze_of_alloc(self):
+        # This shows up in the graph of repeat
+        x = pt.vector("x", shape=(9,))
+        bcast_x = pt.alloc(x, 1, 12, x.shape[0])
+
+        # This reshape does an implicit squeeze
+        out = bcast_x.reshape((12, x.shape[0]))
+
+        new_out = rewrite_graph(out, include=("canonicalize", "ShapeOpt"))
+        assert equal_computations([new_out], [pt.alloc(x, 12, 9)], strict_dtype=False)
+
+
+def test_expand_dims_squeeze_reshape_fusion():
+    x = pt.tensor("x", shape=(1, 9))
+    reshape_x = x.squeeze(0).reshape((3, 3))[..., None]
+
+    assert isinstance(reshape_x.owner.op, DimShuffle)
+    assert isinstance(reshape_x.owner.inputs[0].owner.op, Reshape)
+    assert isinstance(reshape_x.owner.inputs[0].owner.inputs[0].owner.op, DimShuffle)
+
+    out = rewrite_graph(reshape_x, include=("specialize",))
+
+    # In this case we cannot get rid of the reshape, squeeze or expand_dims,
+    # so we fuse them all in one reshape
+    assert equal_computations([out], [x.reshape((3, 3, 1))])
+
+
+def test_implicit_broadcasting_via_repeat():
+    x = pt.vector("x", shape=(3,), dtype=int)
+    y = pt.vector("y", shape=(9,), dtype=int)
+    out = x[None, :].repeat(9, axis=0) <= y[:, None].repeat(3, axis=1)
+    # There are two Reshapes in the graph
+    assert isinstance(out.owner.inputs[0].owner.op, Reshape)
+    assert isinstance(out.owner.inputs[1].owner.op, Reshape)
+
+    new_out = rewrite_graph(out, include=("canonicalize", "specialize"))
+    assert equal_computations([new_out], [x[None] <= y[:, None]])
+
+    no_rewrite_mode = Mode(linker="py", optimizer=None)
+    x_test = np.arange(3) + 1
+    y_test = np.arange(9)
+    np.testing.assert_allclose(
+        new_out.eval({x: x_test, y: y_test}, mode=no_rewrite_mode),
+        out.eval({x: x_test, y: y_test}, mode=no_rewrite_mode),
+    )
+
 
 def test_local_reshape_lift():
     x = tensor4()