Development

.gitignore

@@ -6,9 +6,9 @@ archive/*
 .vscode/*
 .env
 app/*
-circulus_workspace.code-workspace
+circulax_workspace.code-workspace
 docs/examples/
 docs/references/
 *code-workspace
 site/*
-circulus/_version.py
+circulax/_version.py

ReadMe.md

@@ -1,11 +1,11 @@
-# **Circulus**
+# **circulax**
 
 <img src="docs/images/logo_white.svg" alt="logo" width="500">
 
 ## **A Differentiable, Functional Circuit Simulator based on JAX**
-Circulus is a differentiable circuit simulation framework built on [JAX](https://docs.jax.dev/en/latest/notebooks/thinking_in_jax.html), [Optimistix](https://github.com/patrick-kidger/optimistix) and [Diffrax](https://docs.kidger.site/diffrax/). It treats circuit netlists as systems of Ordinary Differential Equations (ODEs), leveraging Diffrax's suite of numerical solvers for transient analysis.
+circulax is a differentiable circuit simulation framework built on [JAX](https://docs.jax.dev/en/latest/notebooks/thinking_in_jax.html), [Optimistix](https://github.com/patrick-kidger/optimistix) and [Diffrax](https://docs.kidger.site/diffrax/). It treats circuit netlists as systems of Ordinary Differential Equations (ODEs), leveraging Diffrax's suite of numerical solvers for transient analysis.
 
-By using JAX as its backend, Circulus provides:
+By using JAX as its backend, circulax provides:
 
 **Native Differentiation**: Full support for forward and reverse-mode automatic differentiation through the solver, enabling gradient-based parameter optimization and inverse design.
 
@@ -15,9 +15,9 @@ By using JAX as its backend, Circulus provides:
 
 **Modular Architecture**: A functional approach to simulation that integrates directly into machine learning and scientific computing workflows.
 
-Standard tools (SPICE, Spectre, Ngspice) rely on established matrix stamping methods and CPU-bound sparse solvers. Circulus leverages the JAX ecosystem to offer specific advantages in optimization and hardware utilization:
+Standard tools (SPICE, Spectre, Ngspice) rely on established matrix stamping methods and CPU-bound sparse solvers. circulax leverages the JAX ecosystem to offer specific advantages in optimization and hardware utilization:
 
-| Feature | Legacy(SPICE) | Circulus |
+| Feature | Legacy(SPICE) | circulax |
 | ----------- | ----------- | ----------- |
 | Model Definition    | Hardcoded C++ / Verilog-A | Simple python functions |
 | Derivatives   | Hardcoded (C) or Compiler-Generated (Verilog-A) | Automatic Differentiation (AD)|
@@ -27,13 +27,13 @@ Standard tools (SPICE, Spectre, Ngspice) rely on established matrix stamping met
 
 
 ## **Simulator setup**
-Circulus strictly separates Physics, Topology, and Analysis, enabling the interchange of solvers or models without netlist modification.
+circulax strictly separates Physics, Topology, and Analysis, enabling the interchange of solvers or models without netlist modification.
 
 ### **Physics Layer**
 Components are defined as simple Python functions wrapped with the ```@component``` decorator. This functional interface abstracts away the boilerplate, allowing users to define physics using simple voltage/current/field/flux relationships.
 
 ```python
-from circulus.base_component import component, Signals, States
+from circulax.base_component import component, Signals, States
 import jax.numpy as jnp
 
 @component(ports=("p1", "p2"))
@@ -84,7 +84,7 @@ The solver is a generic DAE engine linking Diffrax (Time-stepping) and Optimisti
 ##  **Installation**
 
 ```sh
-pip install circulus
+pip install circulax
 ```
 
 ## **Simulation Example**
@@ -93,10 +93,10 @@ pip install circulus
 import jax
 import diffrax
 
-from circulus.components import Resistor, Capacitor, Inductor, VoltageSource
-from circulus.compiler import compile_netlist
-from circulus.solvers.transient import VectorizedTransientSolver
-from circulus.solvers.strategies import DenseSolver
+from circulax.components import Resistor, Capacitor, Inductor, VoltageSource
+from circulax.compiler import compile_netlist
+from circulax.solvers.transient import VectorizedTransientSolver
+from circulax.solvers.strategies import DenseSolver
 import jax.numpy as jnp
 
 import matplotlib.pyplot as plt
@@ -195,4 +195,4 @@ plt.show()
 
 ## **License**
 
-Copyright © 2026, Chris Daunt, [Apache-2.0 License](https://github.com/cdaunt/circulus/blob/master/LICENSE)
+Copyright © 2026, Chris Daunt, [Apache-2.0 License](https://github.com/cdaunt/circulax/blob/master/LICENSE)

circulax/__init__.py

@@ -0,0 +1,5 @@
+"""circulax: A differentiable, JAX based circuit simulator."""
+from circulax.compiler import compile_netlist
+from circulax.netlist import circulaxNetlist as Netlist
+from circulax.netlist import build_net_map, netlist
+from circulax.solvers import analyze_circuit, setup_transient

circulus/compiler.py → circulax/compiler.py

@@ -9,8 +9,8 @@
 import jax
 import jax.numpy as jnp
 
-from circulus.components.base_component import PhysicsReturn, Signals
-from circulus.netlist import build_net_map
+from circulax.components.base_component import PhysicsReturn, Signals
+from circulax.netlist import build_net_map
 
 
 def ensure_time_signature(model_func: callable) -> callable:
@@ -194,7 +194,7 @@ def compile_netlist(netlist: dict, models_map: dict) -> tuple[dict, int, dict]:
             constructor). A ``"GND"`` instance with ``component="ground"`` is
             recognised and skipped.
         models_map: Mapping from model name strings to
-            :class:`~circulus.components.base_component.CircuitComponent`
+            :class:`~circulax.components.base_component.CircuitComponent`
             subclasses, e.g. ``{"Resistor": Resistor, "Capacitor": Capacitor}``.
 
     Returns:

circulus/components/electronic.py → circulax/components/electronic.py

@@ -3,7 +3,7 @@
 import jax.nn as jnn
 import jax.numpy as jnp
 
-from circulus.components.base_component import (
+from circulax.components.base_component import (
     PhysicsReturn,
     Signals,
     States,

circulus/components/photonic.py → circulax/components/photonic.py

@@ -6,8 +6,8 @@
 import jax.nn as jnn
 import jax.numpy as jnp
 
-from circulus.components.base_component import PhysicsReturn, Signals, States, component, source
-from circulus.s_transforms import s_to_y
+from circulax.components.base_component import PhysicsReturn, Signals, States, component, source
+from circulax.s_transforms import s_to_y
 
 # ===========================================================================
 # Passive Optical Components (S-Matrix based)

circulus/netlist.py → circulax/netlist.py

@@ -1,6 +1,6 @@
-"""Circulus netlists.
+"""circulax netlists.
 
-SAX netlists will be used as much as possible in circulus;
+SAX netlists will be used as much as possible in circulax;
 however, connections for node based simulators need to be handled slightly differently.
 """
 
@@ -26,7 +26,7 @@
     InstancePort, InstancePort | tuple[InstancePort, ...]
 ]
 
-CirculusNetlist = Annotated[
+circulaxNetlist = Annotated[
     TypedDict(
         "Netlist",
         {
@@ -52,11 +52,11 @@
     settings: Global circuit settings.
 """
 
-Netlist = CirculusNetlist
+Netlist = circulaxNetlist
 
-# Monkeypatch sax.Netlist to be CirculusNetlist so that all functions using sax.Netlist
+# Monkeypatch sax.Netlist to be circulaxNetlist so that all functions using sax.Netlist
 # May need to make this explicit in the future
-sax_netlist.Netlist = CirculusNetlist  # type: ignore[assignment]
+sax_netlist.Netlist = circulaxNetlist  # type: ignore[assignment]
 
 
 def build_net_map(netlist: dict) -> tuple[dict[str, int], int]:

circulus/s_transforms.py → circulax/s_transforms.py

@@ -1,7 +1,7 @@
 """Utilities for converting between S-parameters.
 
 Utilities for converting between S-parameter and admittance representations,
-and for wrapping SAX model functions as Circulus components.
+and for wrapping SAX model functions as circulax components.
 """
 
 import inspect
@@ -10,7 +10,7 @@
 import jax.numpy as jnp
 from sax import get_ports, sdense
 
-from circulus.components.base_component import Signals, States, component
+from circulax.components.base_component import Signals, States, component
 
 
 @jax.jit
@@ -35,11 +35,11 @@ def s_to_y(S: jax.Array, z0: float = 1.0) -> jax.Array:
 
 
 def sax_component(fn: callable) -> callable:
-    """Decorator to convert a SAX model function into a Circulus component.
+    """Decorator to convert a SAX model function into a circulax component.
 
     Inspects ``fn`` at decoration time to discover its port interface via a
     dry run, then wraps its S-matrix output in an admittance-based physics
-    function compatible with the Circulus nodal solver.
+    function compatible with the circulax nodal solver.
 
     The conversion proceeds in three stages:
 
@@ -51,8 +51,8 @@ def sax_component(fn: callable) -> callable:
        it to an admittance matrix via :func:`s_to_y`, and returns
        ``I = Y @ V`` as a port current dict.
     3. **Component registration** — the wrapper is passed to
-       :func:`~circulus.components.base_component.component` with the
-       discovered ports, producing a :class:`~circulus.components.base_component.CircuitComponent`
+       :func:`~circulax.components.base_component.component` with the
+       discovered ports, producing a :class:`~circulax.components.base_component.CircuitComponent`
        subclass.
 
     Args:
@@ -62,7 +62,7 @@ def sax_component(fn: callable) -> callable:
             ``1.0`` during the dry run.
 
     Returns:
-        A :class:`~circulus.components.base_component.CircuitComponent`
+        A :class:`~circulax.components.base_component.CircuitComponent`
         subclass named after ``fn``.
 
     Raises:

circulus/solvers/assembly.py → circulax/solvers/assembly.py

@@ -7,7 +7,7 @@
 - **Full assembly** (:func:`assemble_system_real`, :func:`assemble_system_complex`)
   — evaluates both the residual and the forward-mode Jacobian via
   ``jax.jacfwd``. Used once per timestep to assemble and factor the frozen
-  Jacobian in :class:`~circulus.solver.FactorizedTransientSolver`.
+  Jacobian in :class:`~circulax.solver.FactorizedTransientSolver`.
 
 - **Residual only** (:func:`assemble_residual_only_real`,
   :func:`assemble_residual_only_complex`) — evaluates only the primal

circulus/solvers/linear.py → circulax/solvers/linear.py

@@ -27,7 +27,7 @@
 import numpy as np
 import optimistix as optx
 
-from circulus.solvers.assembly import assemble_system_complex, assemble_system_real
+from circulax.solvers.assembly import assemble_system_complex, assemble_system_real
 
 # Check if split solver available — KLUHandleManager was added in a later version of klujax.
 split_solver_available = True

circulus/solvers/transient.py → circulax/solvers/transient.py

@@ -11,13 +11,13 @@
 from jax.typing import ArrayLike
 #from klujax import free_numeric
 
-from circulus.solvers.assembly import (
+from circulax.solvers.assembly import (
     assemble_residual_only_complex,
     assemble_residual_only_real,
     assemble_system_complex,
     assemble_system_real,
 )
-from circulus.solvers.linear import CircuitLinearSolver
+from circulax.solvers.linear import CircuitLinearSolver
 
 
 def _compute_history(component_groups, y_c, t, num_vars) -> ArrayLike:

circulus/utils.py → circulax/utils.py

@@ -1,12 +1,12 @@
-"""Circulus utilities."""
+"""circulax utilities."""
 
 from typing import TYPE_CHECKING
 
 import equinox as eqx
 import jax.numpy as jnp
 
 if TYPE_CHECKING:
-    from circulus.compiler import ComponentGroup
+    from circulax.compiler import ComponentGroup
 
 
 def update_params_dict(

docs/gen_doc_stubs.py

@@ -5,7 +5,7 @@
 
 this_dir = Path(__file__).parent
 
-src_root = (this_dir / "../circulus").resolve()
+src_root = (this_dir / "../circulax").resolve()
 
 for path in src_root.rglob("*.py"):
     rel_path = path.relative_to(src_root)
@@ -20,7 +20,7 @@
     else:
         doc_path = rel_path.with_suffix(".md")
 
-    full_parts = ("circulus",) + parts
+    full_parts = ("circulax",) + parts
     identifier = ".".join(full_parts)
 
     output_filename = Path("references") / doc_path