pytensor 2.35.0__tar.gz → 2.35.1__tar.gz

{pytensor-2.35.0/pytensor.egg-info → pytensor-2.35.1}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: pytensor
-Version: 2.35.0
+Version: 2.35.1
 Summary: Optimizing compiler for evaluating mathematical expressions on CPUs and GPUs.
 Author-email: pymc-devs <pymc.devs@gmail.com>
 License-Expression: BSD-3-Clause

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/_version.py

@@ -8,11 +8,11 @@ import json
 
 version_json = '''
 {
- "date": "2025-10-14T09:22:05+0200",
+ "date": "2025-10-20T17:04:33+0200",
  "dirty": false,
  "error": null,
- "full-revisionid": "f772066a9691c64430e100581ab1c2398e43451e",
- "version": "2.35.0"
+ "full-revisionid": "5547eb086c6f281621fb81f2ef93e041ee05aece",
+ "version": "2.35.1"
 }
 '''  # END VERSION_JSON
 

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/configdefaults.py

@@ -1099,12 +1099,6 @@ def add_scan_configvars():
 
 
 def add_numba_configvars():
-    config.add(
-        "numba__vectorize_target",
-        ("Default target for numba.vectorize."),
-        EnumStr("cpu", ["parallel", "cuda"], mutable=True),
-        in_c_key=False,
-    )
     config.add(
         "numba__fastmath",
         ("If True, use Numba's fastmath mode."),

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/configparser.py

@@ -157,7 +157,6 @@ class PyTensorConfigParser:
     scan__allow_gc: bool
     scan__allow_output_prealloc: bool
     # add_numba_configvars
-    numba__vectorize_target: str
     numba__fastmath: bool
     numba__cache: bool
     # add_caching_dir_configvars

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/link/c/basic.py

@@ -1392,7 +1392,8 @@ class CLinker(Linker):
 
             # It is important that a variable (i)
             # yield a 'position' that reflects its role in code_gen()
-            if isinstance(i, AtomicVariable):  # orphans
+            inp_sig = isig = fgraph_inputs_dict.get(i, False)  # inputs
+            if isinstance(i, AtomicVariable):  # orphans or constant inputs
                 if id(i) not in constant_ids:
                     isig = (i.signature(), topological_pos, i_idx)
                     # If the PyTensor constant provides a strong hash
@@ -1412,11 +1413,7 @@ class CLinker(Linker):
                     constant_ids[id(i)] = isig
                 else:
                     isig = constant_ids[id(i)]
-                # print 'SIGNATURE', i.signature()
-                # return i.signature()
-            elif i in fgraph_inputs_dict:  # inputs
-                isig = fgraph_inputs_dict[i]
-            else:
+            elif inp_sig is None:
                 if i.owner is None:
                     assert all(all(out is not None for out in o.outputs) for o in order)
                     assert all(input.owner is None for input in fgraph.inputs)
@@ -1432,7 +1429,7 @@ class CLinker(Linker):
                     )
                 else:
                     isig = (op_pos[i.owner], i.owner.outputs.index(i))  # temps
-            return (isig, i in no_recycling)
+            return (inp_sig, isig, i in no_recycling)
 
         version = []
         for node_pos, node in enumerate(order):

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/link/c/cmodule.py

@@ -2890,7 +2890,7 @@ def default_blas_ldflags() -> str:
     if rpath is not None:
         maybe_add_to_os_environ_pathlist("PATH", rpath)
     try:
-        # 1. Try to use MKL with INTEL OpenMP threading
+        # 1a. Try to use MKL with INTEL OpenMP threading
         _logger.debug("Checking MKL flags with intel threading")
         return _check_libs(
             all_libs,
@@ -2906,6 +2906,24 @@ def default_blas_ldflags() -> str:
         )
     except Exception as e:
         _logger.debug(e)
+    try:
+        # 1b. Try to use MKL with INTEL OpenMP threading with renamed iomp5 library
+        # Ref: <https://www.intel.com/content/www/us/en/docs/onemkl/developer-guide-windows/2024-1/selecting-libraries-to-link-with.html>
+        _logger.debug("Checking MKL flags with intel threading, iomp5md")
+        return _check_libs(
+            all_libs,
+            required_libs=[
+                "mkl_core",
+                "mkl_rt",
+                "mkl_intel_thread",
+                "iomp5md",
+                "pthread",
+            ],
+            extra_compile_flags=[f"-Wl,-rpath,{rpath}"] if rpath is not None else [],
+            cxx_library_dirs=cxx_library_dirs,
+        )
+    except Exception as e:
+        _logger.debug(e)
     try:
         # 2. Try to use MKL with GNU OpenMP threading
         _logger.debug("Checking MKL flags with GNU OpenMP threading")

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/link/jax/dispatch/slinalg.py

@@ -10,6 +10,7 @@ from pytensor.tensor.slinalg import (
     Cholesky,
     CholeskySolve,
     Eigvalsh,
+    Expm,
     LUFactor,
     PivotToPermutations,
     Solve,
@@ -176,6 +177,15 @@ def jax_funcify_QR(op, **kwargs):
     mode = op.mode
 
     def qr(x, mode=mode):
-        return jax.scipy.linalg.qr(x, mode=mode)
+        res = jax.scipy.linalg.qr(x, mode=mode)
+        return res[0] if len(res) == 1 else res
 
     return qr
+
+
+@jax_funcify.register(Expm)
+def jax_funcify_Expm(op, **kwargs):
+    def expm(x):
+        return jax.scipy.linalg.expm(x)
+
+    return expm

{pytensor-2.35.0 → pytensor-2.35.1}/pytensor/link/numba/dispatch/__init__.py

@@ -9,8 +9,10 @@ import pytensor.link.numba.dispatch.nlinalg
 import pytensor.link.numba.dispatch.random
 import pytensor.link.numba.dispatch.scan
 import pytensor.link.numba.dispatch.scalar
+import pytensor.link.numba.dispatch.shape
 import pytensor.link.numba.dispatch.signal
 import pytensor.link.numba.dispatch.slinalg
+import pytensor.link.numba.dispatch.sort
 import pytensor.link.numba.dispatch.sparse
 import pytensor.link.numba.dispatch.subtensor
 import pytensor.link.numba.dispatch.tensor_basic