Add prerequisite installation instructions to all examples

advanced/RW_NS.ipynb

@@ -1,5 +1,26 @@
 {
  "cells": [
+  {
+   "cell_type": "markdown",
+   "id": "7c7c10oq0bx",
+   "metadata": {},
+   "source": [
+    "## Prerequisites\n",
+    "\n",
+    "Install the required dependencies:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "id": "ibs4jedhvj",
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "!pip install git+https://github.com/handley-lab/blackjax\n",
+    "!pip install tqdm anesthetic numpy"
+   ]
+  },
   {
    "cell_type": "markdown",
    "id": "40a4be43",

basic/basic.ipynb

@@ -2,16 +2,20 @@
  "cells": [
   {
    "cell_type": "markdown",
+   "source": "# Walkthrough of Nested Sampling\n\nWe walk through the same components in the [Quickstart](quickstart.ipynb), to give more detail as to what the choices mean.\n\n## Problem definition\n\nWe show how to use the basic user interface for Nested Slice Sampling. We will set up a 20D problem involving a multivariate unit Gaussian $\\mathcal{N}(0,\\mathbf{I})$ prior, and a multivariate Gaussian likelihood with a randomized mean and covariance. The resulting posterior and normalizing constant is analytically known.",
    "metadata": {},
-   "source": [
-    "# Walkthrough of Nested Sampling\n",
-    "\n",
-    "We walk through the same components in the [Quickstart](quickstart.ipynb), to give more detail as to what the choices mean.\n",
-    "\n",
-    "## Problem definition\n",
-    "\n",
-    "We show how to use the basic user interface for Nested Slice Sampling. We will set up a 20D problem involving a multivariate unit Gaussian $\\mathcal{N}(0,\\mathbf{I})$ prior, and a multivariate Gaussian likelihood with a randomized mean and covariance. The resulting posterior and normalizing constant is analytically known."
-   ]
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "code",
+   "metadata": {},
+   "source": "!pip install git+https://github.com/handley-lab/blackjax\n!pip install anesthetic numpy tqdm matplotlib"
   },
   {
    "cell_type": "code",
@@ -463,4 +467,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 2
-}
+}

basic/line_fitting.ipynb

@@ -16,15 +16,13 @@
   },
   {
    "cell_type": "markdown",
-   "id": "e0j955op4hr",
-   "metadata": {},
-   "source": [
-    "### Installation"
-   ]
+   "id": "sk93iv14d1a",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",
-   "execution_count": 1,
+   "execution_count": null,
    "id": "wl6jvwv27m",
    "metadata": {
     "execution": {
@@ -34,20 +32,14 @@
      "shell.execute_reply": "2025-09-22T17:42:50.231409Z"
     }
    },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m A new release of pip is available: \u001b[0m\u001b[31;49m23.3.1\u001b[0m\u001b[39;49m -> \u001b[0m\u001b[32;49m25.2\u001b[0m\n",
-      "\u001b[1m[\u001b[0m\u001b[34;49mnotice\u001b[0m\u001b[1;39;49m]\u001b[0m\u001b[39;49m To update, run: \u001b[0m\u001b[32;49mpython3.11 -m pip install --upgrade pip\u001b[0m\n"
-     ]
-    }
-   ],
-   "source": [
-    "!pip install -q git+https://github.com/handley-lab/blackjax anesthetic tqdm"
-   ]
+   "outputs": [],
+   "source": "!pip install -q git+https://github.com/handley-lab/blackjax anesthetic tqdm matplotlib distrax optax flax"
+  },
+  {
+   "cell_type": "markdown",
+   "id": "z8k3wvxkqf",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",
@@ -958,4 +950,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 5
-}
+}

basic/quickstart.ipynb

@@ -2,12 +2,20 @@
  "cells": [
   {
    "cell_type": "markdown",
+   "source": "# Quickstart\n\nAssuming some level of knowledge with the blackjax style, the following is a minimum working example of a basic sampling problem",
    "metadata": {},
-   "source": [
-    "# Quickstart\n",
-    "\n",
-    "Assuming some level of knowledge with the blackjax style, the following is a minimum working example of a basic sampling problem "
-   ]
+   "execution_count": null,
+   "outputs": []
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "code",
+   "metadata": {},
+   "source": "!pip install git+https://github.com/handley-lab/blackjax\n!pip install anesthetic numpy tqdm"
   },
   {
    "cell_type": "code",
@@ -128,4 +136,4 @@
  },
  "nbformat": 4,
  "nbformat_minor": 2
-}
+}

integrations/c.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# C\n\nThis example demonstrates how to use BlackJAX nested sampling with C implementations of likelihood and prior functions. The C code is compiled to a shared library and accessed via Python's `ctypes` library, with JAX's `pure_callback` providing the bridge.\n\n## Prerequisites\n\nInstall the required Python packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install numpy tqdm\n```\n\nYou'll also need a C compiler (gcc) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the C implementation\n\nFirst, create a file `model.c` with your likelihood and prior functions:"
+   "source": "# C\n\nThis example demonstrates how to use BlackJAX nested sampling with C implementations of likelihood and prior functions. The C code is compiled to a shared library and accessed via Python's `ctypes` library, with JAX's `pure_callback` providing the bridge.\n\nYou'll also need a C compiler (gcc) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the C implementation\n\nFirst, create a file `model.c` with your likelihood and prior functions:"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "markdown",

integrations/cpp.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# C++\n\nThis example demonstrates how to use BlackJAX nested sampling with C++ implementations of likelihood and prior functions. The C++ code is compiled using pybind11 to create a Python module, with JAX's `pure_callback` providing the bridge.\n\n## Prerequisites\n\nInstall the required Python packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install pybind11 numpy tqdm\n```\n\nYou'll also need a C++ compiler (g++) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the C++ implementation\n\nFirst, create a file `model.cpp` with your likelihood and prior functions using pybind11:"
+   "source": "# C++\n\nThis example demonstrates how to use BlackJAX nested sampling with C++ implementations of likelihood and prior functions. The C++ code is compiled using pybind11 to create a Python module, with JAX's `pure_callback` providing the bridge.\n\nYou'll also need a C++ compiler (g++) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the C++ implementation\n\nFirst, create a file `model.cpp` with your likelihood and prior functions using pybind11:"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "markdown",

integrations/cupy.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# CuPy\n\nThis example demonstrates how to use BlackJAX nested sampling with CuPy, the GPU-accelerated drop-in replacement for NumPy. CuPy enables GPU computation for likelihood and prior functions while maintaining NumPy-like syntax.\n\n## Prerequisites\n\nInstall the required packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install cupy-cuda12x numpy tqdm  # Or cupy-cuda11x for CUDA 11.x\n```\n\nNote: CuPy requires CUDA. Choose the appropriate CuPy package for your CUDA version.\n\n## Run nested sampling with CuPy functions"
+   "source": "# CuPy\n\nThis example demonstrates how to use BlackJAX nested sampling with CuPy, the GPU-accelerated drop-in replacement for NumPy. CuPy enables GPU computation for likelihood and prior functions while maintaining NumPy-like syntax.\n\nNote: CuPy requires CUDA. Choose the appropriate CuPy package for your CUDA version.\n\n## Run nested sampling with CuPy functions"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",

integrations/fortran.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# Fortran\n\nThis example demonstrates how to use BlackJAX nested sampling with Fortran implementations of likelihood and prior functions. The Fortran code is compiled to a shared library and accessed via Python's `ctypes` library, with JAX's `pure_callback` providing the bridge.\n\n## Prerequisites\n\nInstall the required Python packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install numpy tqdm\n```\n\nYou'll also need a Fortran compiler (gfortran) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the Fortran implementation\n\nFirst, create a file `model.f90` with your likelihood and prior functions:"
+   "source": "# Fortran\n\nThis example demonstrates how to use BlackJAX nested sampling with Fortran implementations of likelihood and prior functions. The Fortran code is compiled to a shared library and accessed via Python's `ctypes` library, with JAX's `pure_callback` providing the bridge.\n\nYou'll also need a Fortran compiler (gfortran) installed on your system.\n\n## Setup Instructions\n\n### 1. Create the Fortran implementation\n\nFirst, create a file `model.f90` with your likelihood and prior functions:"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "markdown",

integrations/julia.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# Julia\n\nThis example demonstrates how to use BlackJAX nested sampling with Julia implementations of likelihood and prior functions. Julia code is called via JSON-RPC in a separate process to avoid threading conflicts.\n\n## Prerequisites\n\nInstall the required Python packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install numpy tqdm\n```\n\nInstall Julia and the required Julia packages:\n```julia\nusing Pkg\nPkg.add(\"LinearAlgebra\")\nPkg.add(\"Distributions\")\nPkg.add(\"JSON\")\n```\n\n## Setup Instructions\n\n### 1. Create the model file\n\nCreate a file `model.jl` with your likelihood and prior functions:\n\n```julia\nusing LinearAlgebra\nusing Distributions\n\nfunction loglikelihood(theta)\n    theta = convert(Matrix{Float64}, theta)\n    dist = MvNormal(ones(5), 0.01 * I(5))\n    return [logpdf(dist, theta[i, :]) for i in 1:size(theta, 1)]\nend\n\nfunction logprior(theta)\n    theta = convert(Matrix{Float64}, theta)\n    dist = MvNormal(zeros(5), I(5))\n    return [logpdf(dist, theta[i, :]) for i in 1:size(theta, 1)]\nend\n```\n\n### 2. Create the RPC server\n\nCreate a file `julia_server.jl` to handle RPC communication:\n\n```julia\nusing JSON\nusing Base64\n\n# Include the model functions\ninclude(\"model.jl\")\n\n# Simple request/response loop\nwhile true\n    try\n        line = readline()\n        if isempty(line)\n            break\n        end\n        \n        request = JSON.parse(line)\n        func_name = request[\"function\"]\n        \n        # Decode base64 numpy array\n        theta_bytes = base64decode(request[\"data\"])\n        theta = reinterpret(Float64, theta_bytes)\n        rows = request[\"rows\"]\n        cols = request[\"cols\"]\n        theta = reshape(theta, cols, rows)'  # Transpose for column-major\n        \n        # Call function\n        if func_name == \"loglikelihood\"\n            result = loglikelihood(theta)\n        elseif func_name == \"logprior\"\n            result = logprior(theta)\n        else\n            error(\"Unknown function: $func_name\")\n        end\n        \n        # Send response\n        response = Dict(\"result\" => result)\n        println(JSON.json(response))\n        flush(stdout)\n    catch e\n        println(JSON.json(Dict(\"error\" => string(e))))\n        flush(stdout)\n    end\nend\n```\n\n### 3. Run nested sampling with Julia functions"
+   "source": "# Julia\n\nThis example demonstrates how to use BlackJAX nested sampling with Julia implementations of likelihood and prior functions. Julia code is called via JSON-RPC in a separate process to avoid threading conflicts.\n\n## Setup Instructions\n\n### 1. Create the model file\n\nCreate a file `model.jl` with your likelihood and prior functions:\n\n```julia\nusing LinearAlgebra\nusing Distributions\n\nfunction loglikelihood(theta)\n    theta = convert(Matrix{Float64}, theta)\n    dist = MvNormal(ones(5), 0.01 * I(5))\n    return [logpdf(dist, theta[i, :]) for i in 1:size(theta, 1)]\nend\n\nfunction logprior(theta)\n    theta = convert(Matrix{Float64}, theta)\n    dist = MvNormal(zeros(5), I(5))\n    return [logpdf(dist, theta[i, :]) for i in 1:size(theta, 1)]\nend\n```\n\n### 2. Create the RPC server\n\nCreate a file `julia_server.jl` to handle RPC communication:\n\n```julia\nusing JSON\nusing Base64\n\n# Include the model functions\ninclude(\"model.jl\")\n\n# Simple request/response loop\nwhile true\n    try\n        line = readline()\n        if isempty(line)\n            break\n        end\n        \n        request = JSON.parse(line)\n        func_name = request[\"function\"]\n        \n        # Decode base64 numpy array\n        theta_bytes = base64decode(request[\"data\"])\n        theta = reinterpret(Float64, theta_bytes)\n        rows = request[\"rows\"]\n        cols = request[\"cols\"]\n        theta = reshape(theta, cols, rows)'  # Transpose for column-major\n        \n        # Call function\n        if func_name == \"loglikelihood\"\n            result = loglikelihood(theta)\n        elseif func_name == \"logprior\"\n            result = logprior(theta)\n        else\n            error(\"Unknown function: $func_name\")\n        end\n        \n        # Send response\n        response = Dict(\"result\" => result)\n        println(JSON.json(response))\n        flush(stdout)\n    catch e\n        println(JSON.json(Dict(\"error\" => string(e))))\n        flush(stdout)\n    end\nend\n```\n\n### 3. Run nested sampling with Julia functions"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",

integrations/numba.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# Numba\n\nThis example demonstrates how to use BlackJAX nested sampling with Numba JIT-compiled functions. Numba compiles Python functions to machine code at runtime, providing significant speedups for numerical computations while keeping the code in pure Python.\n\n## Prerequisites\n\nInstall the required packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install numba numpy tqdm\n```\n\n## Run nested sampling with Numba functions"
+   "source": "# Numba\n\nThis example demonstrates how to use BlackJAX nested sampling with Numba JIT-compiled functions. Numba compiles Python functions to machine code at runtime, providing significant speedups for numerical computations while keeping the code in pure Python.\n\n## Run nested sampling with Numba functions"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",

integrations/numpy.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# NumPy/SciPy\n\nThis example demonstrates how to use BlackJAX nested sampling with NumPy/SciPy distributions. SciPy provides a wide range of statistical distributions that can be used directly with BlackJAX through the wrapping pattern.\n\n## Prerequisites\n\nInstall the required packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install scipy numpy tqdm\n```"
+   "source": "# NumPy/SciPy\n\nThis example demonstrates how to use BlackJAX nested sampling with NumPy/SciPy distributions. SciPy provides a wide range of statistical distributions that can be used directly with BlackJAX through the wrapping pattern."
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",

integrations/numpy_scipy.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# NumPy/SciPy\n\nThis example demonstrates how to use BlackJAX nested sampling with NumPy and SciPy functions. This approach is useful when you have existing likelihood functions written in NumPy/SciPy that you want to use with BlackJAX.\n\n## Prerequisites\n\nInstall the required packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install scipy numpy tqdm\n```\n\n## Run nested sampling with NumPy/SciPy functions"
+   "source": "# NumPy/SciPy\n\nThis example demonstrates how to use BlackJAX nested sampling with NumPy and SciPy functions. This approach is useful when you have existing likelihood functions written in NumPy/SciPy that you want to use with BlackJAX.\n\n## Run nested sampling with NumPy/SciPy functions"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
   },
   {
    "cell_type": "code",

integrations/pytorch.ipynb

@@ -3,7 +3,17 @@
   {
    "cell_type": "markdown",
    "metadata": {},
-   "source": "# PyTorch\n\nThis example demonstrates how to use BlackJAX nested sampling with PyTorch distributions and likelihood functions. The key is wrapping PyTorch functions to be compatible with JAX using `jax.pure_callback`.\n\n## Prerequisites\n\nInstall the required packages:\n```bash\npip install git+https://github.com/handley-lab/blackjax\npip install torch numpy tqdm\n```\n\n## Run nested sampling with PyTorch functions"
+   "source": "# PyTorch\n\nThis example demonstrates how to use BlackJAX nested sampling with PyTorch distributions and likelihood functions. The key is wrapping PyTorch functions to be compatible with JAX using `jax.pure_callback`.\n\n## Run nested sampling with PyTorch functions"
+  },
+  {
+   "cell_type": "markdown",
+   "source": "## Prerequisites\n\nInstall the required dependencies:",
+   "metadata": {}
+  },
+  {
+   "cell_type": "code",
+   "source": "!pip install git+https://github.com/handley-lab/blackjax\n!pip install torch numpy tqdm",
+   "metadata": {}
   },
   {
    "cell_type": "code",