Address PR review feedback

- Restore pixi references in docs and warnings (previously removed without explanation)
- Restore helpful workflow comments in pypi.yml
- Move Pyodide build docs from dev_start_guide to install.rst (more appropriate location)
- Add user-facing WASM/Pyodide installation instructions in install.rst

Author: ericmjl

.github/workflows/pypi.yml

@@ -144,10 +144,12 @@ jobs:
   build_pyodide_wheel:
     name: Build Pyodide (WASM) wheel
     needs: check_changes
+    # Run if it's a release or if relevant files changed on main
     if: |
       needs.check_changes.outputs.should_run == 'true'
     runs-on: ubuntu-latest
     permissions:
+      # write id-token and attestations are required to attest build provenance
       id-token: write
       attestations: write
     steps:

.pre-commit-config.yaml

@@ -25,7 +25,7 @@ repos:
     rev: v1.0.0                                                                          
     hooks:                                                                               
     - id: sphinx-lint 
-      args: ["."]
+      args: ["-i", ".pixi", "."]
   - repo: https://github.com/astral-sh/ruff-pre-commit
     rev: v0.14.0
     hooks:

doc/dev_start_guide.rst

@@ -222,54 +222,6 @@ Afterward, you can go to `html/index.html` and navigate the changes in a browser
 **Do not commit the `html` directory. The documentation is built automatically.**
 For more documentation customizations such as different formats e.g., PDF, refer to the `Sphinx documentation <https://www.sphinx-doc.org/en/master/usage/builders/index.html>`_.
 
-Building a WebAssembly (Pyodide) wheel
---------------------------------------
-
-To build a wheel targeting WebAssembly for use with `Pyodide <https://pyodide.org/>`_ (e.g. for the browser or JupyterLite), use the Pyodide build tooling. This produces a wheel in ``dist/`` with a name like ``*-cpXXX-cpXXX-pyodide_*_wasm32.whl``.
-
-On Pyodide/wasm32, the **numba** dependency is omitted (via environment markers in ``pyproject.toml``), so the package can be installed with ``micropip.install(url)`` without pulling in numba.
-
-**Full workflow (copy-paste)**
-
-Use Python 3.11, 3.12, or 3.13 (Pyodide does not yet support 3.14). From the project root:
-
-**1. One-time: install Emscripten**
-
-.. code-block:: bash
-
-   # Clone and install the Emscripten SDK
-   git clone https://github.com/emscripten-core/emsdk.git /path/to/emsdk
-   cd /path/to/emsdk
-   # Use the version required by pyodide-build (check with: pyodide config get emscripten_version)
-   ./emsdk install 3.1.45
-   ./emsdk activate 3.1.45
-   source emsdk_env.sh
-
-In every new shell where you will run ``pyodide build``, run ``source /path/to/emsdk/emsdk_env.sh`` first (or add it to your profile).
-
-**2. One-time per machine: create a venv and install the Pyodide build extra**
-
-.. code-block:: bash
-
-   python3.12 -m venv .venv-pyodide
-   source .venv-pyodide/bin/activate   # On Windows: .venv-pyodide\Scripts\activate
-   pip install -e ".[pyodide]"
-
-This installs the project in editable form and the ``pyodide-build`` tool (declared in the ``pyodide`` optional dependency).
-
-**3. Each time you want to build the wasm wheel**
-
-.. code-block:: bash
-
-   source /path/to/emsdk/emsdk_env.sh
-   source .venv-pyodide/bin/activate
-   cd /path/to/pytensor
-   pyodide build
-
-The wheel will appear in ``dist/``. PyPI does not yet accept emscripten/wasm32 wheels; host the file elsewhere (e.g. GitHub Releases) and install in Pyodide with ``micropip.install(url)``.
-
-To see which Emscripten version your installed ``pyodide-build`` expects, run ``pyodide config get emscripten_version`` and use that in step 1 if it differs from ``3.1.45``. For more detail, see `Pyodide: building packages <https://pyodide.org/en/stable/development/building-packages-from-source.html>`_.
-
 Other tools that might help
 ===========================
 

doc/install.rst

@@ -26,3 +26,22 @@ The current development branch of PyTensor can be installed from GitHub using `p
 
 
 To use the Numba and JAX backend you will need to install these libraries in addition to PyTensor. Please refer to `Numba's installation instructions <https://numba.readthedocs.io/en/stable/user/installing.html>`__ and `JAX's installation instructions  <https://github.com/google/jax#installation>`__ respectively.
+
+
+Installing on Pyodide (WebAssembly)
+-----------------------------------
+
+PyTensor can be used in browser-based Python environments via `Pyodide <https://pyodide.org/>`__, such as `JupyterLite <https://jupyterlite.readthedocs.io/>`__.
+
+PyPI does not yet accept WebAssembly wheels, so you'll need to install from a GitHub Release:
+
+.. code-block:: python
+
+    import micropip
+    micropip.install("https://github.com/pymc-devs/pytensor/releases/download/vVERSION/pytensor-VERSION-py312-py312-pyodide_2024_0_wasm32.whl")
+
+Replace ``VERSION`` with the desired release version (e.g., ``2.26.4``).
+
+.. note::
+
+   The Numba backend is not available on Pyodide/wasm32. PyTensor will use other backends automatically.

doc/troubleshooting.rst

@@ -196,7 +196,7 @@ or running :func:`pytensor.link.c.cmodule.default_blas_ldflags`.
 Here are some different way to configure BLAS:
 
 0) Do nothing and use the default config.
-This will usually work great for installation via conda/mamba (conda-forge channel).
+This will usually work great for installation via conda/mamba/pixi (conda-forge channel).
 It will usually fail to link altogether for installation via pip.
 
 1) Disable the usage of BLAS and fall back on NumPy for dot products. To do

pytensor/link/c/cmodule.py

@@ -2985,7 +2985,7 @@ def default_blas_ldflags() -> str:
     _logger.debug("Failed to identify blas ldflags. Will leave them empty.")
     warnings.warn(
         "PyTensor could not link to a BLAS installation. Operations that might benefit from BLAS will be severely degraded.\n"
-        "This usually happens when PyTensor is installed via pip. We recommend it be installed via conda/mamba instead.\n"
+        "This usually happens when PyTensor is installed via pip. We recommend it be installed via conda/mamba/pixi instead.\n"
         "Alternatively, you can use an experimental backend such as Numba or JAX that perform their own BLAS optimizations, "
         "by setting `pytensor.config.mode == 'NUMBA'` or passing `mode='NUMBA'` when compiling a PyTensor function.\n"
         "For more options and details see https://pytensor.readthedocs.io/en/latest/troubleshooting.html#how-do-i-configure-test-my-blas-library",