Stop defining min as negative of max

My best guess of why this was done, was that historically Max was implemented before Min. Now that we have Min, and we always rewrite into it by the end, there seems to be no good reason not to start with it

Author: ricardoV94

pytensor/tensor/math.py

@@ -609,19 +609,11 @@ def min(x, axis=None, keepdims=False):
         will broadcast correctly against the original tensor.
 
     """
-    x = as_tensor_variable(x)
-    str_x_type = str(x.dtype)
-    if str_x_type.startswith("float") or str_x_type in int_dtypes:
-        return -max(-x, axis=axis, keepdims=keepdims)
-    elif str_x_type in uint_dtypes:
-        itype = np.iinfo(x.dtype)
-        max_val = np.array(itype.max, dtype=itype.dtype)
-        return max_val - max(max_val - x, axis=axis, keepdims=keepdims)
-    elif str_x_type == "bool":
-        return ~max(~x, axis=axis, keepdims=keepdims)
-    else:
-        # Be careful about unsigned integers, complex
-        raise NotImplementedError()
+    out = Min(axis=axis)(x)
+
+    if keepdims:
+        out = makeKeepDims(x, out, axis)
+    return out
 
 
 def argmin(x, axis=None, keepdims=False):

pytensor/tensor/rewriting/uncanonicalize.py

@@ -31,45 +31,14 @@
 
 """
 
-from pytensor import scalar as ps
-from pytensor.graph.rewriting.basic import copy_stack_trace, node_rewriter
+from pytensor.graph.rewriting.basic import node_rewriter
 from pytensor.tensor.basic import Alloc, alloc, constant
-from pytensor.tensor.elemwise import CAReduce, DimShuffle
-from pytensor.tensor.math import Min, neg
+from pytensor.tensor.elemwise import DimShuffle
 from pytensor.tensor.rewriting.basic import register_uncanonicalize
 from pytensor.tensor.shape import Reshape, reshape
 from pytensor.tensor.subtensor import Subtensor
 
 
-@register_uncanonicalize
-@node_rewriter([neg])
-def local_max_to_min(fgraph, node):
-    """
-    Change -(max(-x)) to min.
-
-    This is tested in tensor/tests/test_basic.py:test_min_max.
-
-    Notes
-    -----
-    We don't need an opt that will do the reverse as by default
-    the interface put only Max into the graph.
-
-    """
-    if node.op == neg and node.inputs[0].owner:
-        max = node.inputs[0]
-        if (
-            max.owner
-            and isinstance(max.owner.op, CAReduce)
-            and max.owner.op.scalar_op == ps.scalar_maximum
-        ):
-            neg_node = max.owner.inputs[0]
-            if neg_node.owner and neg_node.owner.op == neg:
-                new = Min(max.owner.op.axis)(neg_node.owner.inputs[0])
-                return [copy_stack_trace(node.outputs[0], new)]
-
-    return False
-
-
 @register_uncanonicalize
 @node_rewriter([Alloc])
 def local_alloc_dimshuffle(fgraph, node):