paramrf 0.26.1__tar.gz → 0.26.4__tar.gz

{paramrf-0.26.1/paramrf.egg-info → paramrf-0.26.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: paramrf
-Version: 0.26.1
+Version: 0.26.4
 Summary: Parametric radio frequency modeling
 Author-email: Gary Allen <gvcallen@gmail.com>
 Project-URL: homepage, https://github.com/gvcallen/paramrf

{paramrf-0.26.1 → paramrf-0.26.4}/docs/api/index.rst

@@ -76,4 +76,4 @@ Utilities
    pmrf.unwrap
    pmrf.unwrap_self
    pmrf.as_fixed
-   pmrf.as_free
+   pmrf.as_variable

{paramrf-0.26.1 → paramrf-0.26.4/paramrf.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: paramrf
-Version: 0.26.1
+Version: 0.26.4
 Summary: Parametric radio frequency modeling
 Author-email: Gary Allen <gvcallen@gmail.com>
 Project-URL: homepage, https://github.com/gvcallen/paramrf

{paramrf-0.26.1 → paramrf-0.26.4}/paramrf.egg-info/SOURCES.txt

@@ -60,7 +60,6 @@ pmrf/noise_models.py
 pmrf/parameters.py
 pmrf/problem.py
 pmrf/serialization.py
-pmrf/topology.py
 pmrf/types.py
 pmrf/fitting/__init__.py
 pmrf/fitting/minimize.py
@@ -112,21 +111,25 @@ pmrf/optimize/solvers/optimistix.py
 pmrf/optimize/solvers/scipy.py
 pmrf/rf/__init__.py
 pmrf/rf/conversions.py
+pmrf/rf/mna.py
 pmrf/simulate/__init__.py
 pmrf/simulate/base.py
 pmrf/simulate/cascade.py
 pmrf/simulate/reduce.py
 pmrf/simulate/result.py
 pmrf/simulate/terminate.py
+pmrf/simulate/topology.py
 pmrf/simulate/solvers/__init__.py
 pmrf/simulate/solvers/hallbjorner.py
 pmrf/simulate/solvers/kron.py
 pmrf/simulate/solvers/linear_fractional_terminator.py
 pmrf/simulate/solvers/mobius_terminator.py
+pmrf/simulate/solvers/modified_kron.py
 pmrf/simulate/solvers/redheffer.py
 pmrf/simulate/solvers/transfer_cascader.py
 pmrf/utils/__init__.py
 pmrf/utils/array.py
+pmrf/utils/debug.py
 pmrf/utils/network.py
 pmrf/utils/optix.py
 pmrf/utils/random.py
@@ -156,5 +159,5 @@ tests/test_models/test_interconnected.py
 tests/test_models/test_lines.py
 tests/test_models/test_lumped.py
 tests/test_models/test_nodal.py
-tests/test_models/test_topological.py
+tests/test_models/test_sections.py
 tests/test_models/test_transformed.py

{paramrf-0.26.1 → paramrf-0.26.4}/pmrf/__init__.py

@@ -27,7 +27,7 @@ except PackageNotFoundError:
 # Re-exports
 from pmrf.models import Model as Model
 from pmrf.frequency import Frequency as Frequency
-from pmrf.topology import Topology as Topology
+from pmrf.simulate.topology import Topology as Topology
 from pmrf.problem import Problem as Problem
 
 #: The canonical type hint for a float, or a numpy or JAX array.

{paramrf-0.26.1 → paramrf-0.26.4}/pmrf/models/base.py

@@ -3,6 +3,7 @@ Base class for RF models.
 """
 
 from typing import Any, Callable, Self, TypeVar, Union
+from functools import cached_property
 import dataclasses
 
 import jax
@@ -12,12 +13,12 @@ import equinox as eqx
 import skrf
 import parax as prx
 
+from pmrf.rf.mna import MNAStamp
 from pmrf.utils.optix import focus, Lens
 from pmrf.parameters import Param
 from pmrf.frequency import Frequency
 from pmrf.rf import (
-    s2s, a2s, s2a, s2y, y2s, s2z, z2s, 
-    y2z, z2y, a2y, y2a, a2z, z2a, 
+    a2s, s2a, s2y, y2s, s2z, z2s, y2z, z2y, a2y, y2a, a2z, z2a, s2mna, y2mna, z2mna, a2mna,
 )
 from pmrf.math import CONVERSION_LOOKUP
 from pmrf.utils.type import is_overridden
@@ -25,7 +26,7 @@ from pmrf.utils import field, unwrap, unwrap_self
 
 T = TypeVar('T')
 
-PRIMARY_DOMAINS = ('s', 'a', 'y', 'z')
+PRIMARY_DOMAINS = ('s', 'a', 'y', 'z', 'mna')
 PRIMARY_METHODS = PRIMARY_DOMAINS + ('build', 'primary_matrix')
 HUB_Z0 = 50.0 + 0.0j
 
@@ -64,6 +65,7 @@ class Model(eqx.Module):
     :meth:`a`                         ABCD parameter matrix.
     :meth:`z`                         Impedance (Z) parameter matrix.
     :meth:`y`                         Admittance (Y) parameter matrix.
+    :meth:`mna`                       Modified Nodal Analysis (MNA) stamp matrices.
     :meth:`build`                     Build the model. Can be overridden for advanced models.
     :meth:`primary_matrix`            Return the primary matrix. Can be overridden for dynamic dispatch.
     :attr:`primary_domain`            The domain of the primary matrix as a string (e.g. ``"s"``, ``"a"``).
@@ -181,7 +183,7 @@ class Model(eqx.Module):
 
     # ---- Introspection properties --------------------------------------------------------
     
-    @property
+    @cached_property
     def number_of_ports(self) -> int:
         """Number of ports.
 
@@ -189,11 +191,14 @@ class Model(eqx.Module):
         -------
         int
         """
+        if is_overridden(type(self), Model, 'build'):
+            return self.build().number_of_ports
+
         freq = Frequency(1, 2, 2)
         eval = jax.eval_shape(lambda: self.s(freq))
         return eval.shape[1]
 
-    @property
+    @cached_property
     def nports(self) -> int:
         """Alias of :attr:`number_of_ports`."""
         return self.number_of_ports
@@ -490,11 +495,11 @@ class Model(eqx.Module):
         kwargs = {'z0': HUB_Z0} if primary_domain == 's' else {}
         val = self.primary_matrix(freq, **kwargs)
 
-        # Return or Convert
+        # Return direct
         if primary_domain == 'a':
             return val
         
-        # Convert via S parameters (Hub strategy)
+        # Convert with priority s, z, y
         if primary_domain == 's':
             return s2a(val, z0=HUB_Z0)
         elif primary_domain == 'z':
@@ -529,11 +534,11 @@ class Model(eqx.Module):
         kwargs = {'z0': HUB_Z0} if primary_domain == 's' else {}
         val = self.primary_matrix(freq, **kwargs)
 
-        # Return or Convert
+        # Return direct
         if primary_domain == 'z':
             return val
 
-        # Convert via S parameters (Hub strategy)
+        # Convert with priority s, a, y
         if primary_domain == 's':
             return s2z(val, z0=HUB_Z0)
         elif primary_domain == 'a':
@@ -568,11 +573,11 @@ class Model(eqx.Module):
         kwargs = {'z0': HUB_Z0} if primary_domain == 's' else {}
         val = self.primary_matrix(freq, **kwargs)
 
-        # Return or Convert
+        # Return direct
         if primary_domain == 'y':
             return val
 
-        # Convert via S parameters (Hub strategy)
+        # Convert with priority s, a, z
         if primary_domain == 's':
             return s2y(val, HUB_Z0)
         elif primary_domain == 'a':
@@ -581,6 +586,43 @@ class Model(eqx.Module):
             return z2y(val)
         else:
             raise NotImplementedError(f"Conversion from '{primary_domain}' to 'y' is not implemented.")
+        
+    @eqx.filter_jit
+    @unwrap_self
+    def mna(self, freq: Frequency) -> MNAStamp:
+        """
+        (experimental) Modified Nodal Analysis (MNA) stamp.
+
+        Can be overridden in sub-classes.
+        
+        If the model does not explicitly define an MNA stamp, this automatically 
+        delegates to the appropriate conversion utility (`s2mna`, `z2mna`, etc.). 
+        Explicitly defined Y-matrices are prioritized to maximize matrix sparsity, 
+        while other domains fall back to auxiliary variables to guarantee stability.
+        """
+        if is_overridden(type(self), Model, 'build'):
+            return self.build().mna(freq)
+
+        primary_domain = self.primary_domain
+        
+        if primary_domain == 'mna':
+            return self.primary_matrix(freq)
+            
+        # We prioritize y, z and a for sparsity, assuming the caller
+        # has created a numerically stable implementation.
+        if primary_domain == 'y' or is_overridden(type(self), Model, 'y'):
+            return y2mna(self.y(freq))
+        
+        if primary_domain == 'z' or is_overridden(type(self), Model, 'z'):
+            return z2mna(self.z(freq))
+        
+        if primary_domain == 'a' or is_overridden(type(self), Model, 'a'):
+            return a2mna(self.a(freq))
+        
+        if primary_domain == 's' or is_overridden(type(self), Model, 's'):
+            return s2mna(self.s(freq, z0=HUB_Z0), z0=HUB_Z0)
+            
+        return y2mna(self.y(freq))
 
     # ---- Magic methods and copying --------------------------------------------------
 
@@ -669,9 +711,11 @@ class Model(eqx.Module):
         methods like `.get()` and `.set()` to retrieve values
         or an updated model.
         
-        Warning: Note that updates are "surgical", so the value
-        is replaced as-is i.e. without any automatic converters.
-        When replacing parameters, ensure to pass in a fully
+        WARNING: All updates made by this method are "surgical".
+        In order words, values are replaced *as-is* without any converters
+        or verification applied (a new instance is still returned).
+        Any invariants must therefore be enforced or checked manually.
+        For example, when replacing parameters, ensure to pass in a fully
         constructed parameter and not a float.
         
         Examples

{paramrf-0.26.1 → paramrf-0.26.4}/pmrf/models/components/ideal.py

@@ -9,7 +9,7 @@ import equinox as eqx
 
 from pmrf.models import Model
 from pmrf.frequency import Frequency
-from pmrf.utils import field
+from pmrf.utils import field, error_if
 from pmrf.rf import renormalize_s
 from pmrf.types import ArrayLike, Param
 from pmrf.parameters import param
@@ -37,25 +37,6 @@ class Load(Model):
             repeat(freq.npoints, 0)
         return s
 
-    def y(self, freq: Frequency) -> jnp.ndarray:
-        gamma_arr = jnp.asarray(self.gamma)
-        
-        is_invalid = jnp.any(gamma_arr == -1.0)
-        
-        gamma_safe = eqx.error_if(
-            gamma_arr, 
-            is_invalid, 
-            "Y-matrix is singular and undefined for ideal short (-1.0) loads."
-        )
-        
-        y_val = (1.0 - gamma_safe) / (1.0 + gamma_safe)
-        
-        y = jnp.asarray(y_val).reshape(-1, 1, 1) * \
-            jnp.eye(self.nports, dtype=jnp.complex128).reshape((-1, self.nports, self.nports)).\
-            repeat(freq.npoints, 0)
-            
-        return y
-    
 
 class Short(Model):
     """