Add docs

docs/_static/css/custom.css

@@ -15,3 +15,7 @@ div.wy-side-nav-search {
 .wy-table-responsive table td,.wy-table-responsive table th {
     white-space: normal;
 }
+
+.rst-content .code-block-caption {
+  text-align: left;
+}

docs/index.rst

@@ -84,6 +84,8 @@ Contents
   :maxdepth: 2
   :titlesonly:
 
+  install
+  treelite-migration
   api
   compiler_param
   tl2cgen-doxygen

docs/install.rst

@@ -0,0 +1,146 @@
+==================
+Installation Guide
+==================
+
+You may choose one of two methods to install TL2cgen on your system:
+
+.. contents::
+  :local:
+  :depth: 1
+
+.. _install-pip:
+
+Download binary releases from PyPI (Recommended)
+================================================
+This is probably the most convenient method. Simply type
+
+.. code-block:: console
+
+  pip install tl2cgen
+
+to install the TL2cgen package. The command will locate the binary release that is compatible with
+your current platform. Check the installation by running
+
+.. code-block:: python
+
+  import tl2cgen
+
+in an interactive Python session. This method is available for only Windows, MacOS, and Linux.
+For other operating systems, see the next section.
+
+.. note:: Windows users need to install Visual C++ Redistributable
+
+  TL2cgen requires DLLs from `Visual C++ Redistributable
+  <https://www.microsoft.com/en-us/download/details.aspx?id=48145>`_
+  in order to function, so make sure to install it. Exception: If
+  you have Visual Studio installed, you already have access to
+  necessary libraries and thus don't need to install Visual C++
+  Redistributable.
+
+.. note:: Installing OpenMP runtime on MacOS
+
+  TL2cgen requires the presence of OpenMP runtime. To install OpenMP runtime on a Mac OSX system,
+  run the following command:
+
+  .. code-block:: bash
+
+    brew install libomp
+
+
+.. _install-conda:
+
+Download binary releases from Conda
+===================================
+
+TL2cgen is also available on Conda.
+
+.. code-block:: console
+
+  conda install -c conda-forge tl2cgen
+
+to install the TL2cgen package. See https://anaconda.org/conda-forge/tl2cgen to check the
+available platforms.
+
+.. _install-source:
+
+Compile TL2cgen from the source
+===============================
+Installation consists of two steps:
+
+1. Build the native library from C++ code.
+2. Install the Python package.
+
+.. note:: Name of the native library
+
+   ================== =====================
+   Operating System   Shared library
+   ================== =====================
+   Windows            ``tl2cgen.dll``
+   Mac OS X           ``libtl2cgen.dylib``
+   Linux / other UNIX ``libtl2cgen.so``
+   ================== =====================
+
+To get started, clone TL2cgen repo from GitHub.
+
+.. code-block:: bash
+
+  git clone https://github.com/dmlc/tl2cgen.git
+  cd tl2cgen
+
+The next step is to build the native library.
+
+1-1. Compiling the native library on Linux and MacOS
+----------------------------------------------------
+Here, we use CMake to generate a Makefile:
+
+.. code-block:: bash
+
+  mkdir build
+  cd build
+  cmake ..
+
+Once CMake finished running, simply invoke GNU Make to obtain the native
+libraries.
+
+.. code-block:: bash
+
+  make
+
+The compiled library will be under the ``build/`` directory.
+
+.. note:: Compiling TL2cgen with multithreading on MacOS
+
+  TL2cgen requires the presence of OpenMP runtime. To install OpenMP runtime on a MacOS system,
+  run the following command:
+
+  .. code-block:: bash
+
+    brew install libomp
+
+1-2. Compiling native libraries on Windows
+------------------------------------------
+We can use CMake to generate a Visual Studio project. The following snippet assumes that Visual
+Studio 2022 is installed. Adjust the version depending on the copy that's installed on your system.
+
+.. code-block:: dosbatch
+
+  mkdir build
+  cd build
+  cmake .. -G"Visual Studio 17 2022" -A x64
+
+.. note:: Visual Studio 2019 or newer is required
+
+  TL2cgen uses the C++17 standard. Ensure that you have Visual Studio version 2019 or newer.
+
+Once CMake finished running, open the generated solution file (``tl2cgen.sln``) in Visual Studio.
+From the top menu, select **Build > Build Solution**.
+
+2. Installing Python package
+----------------------------
+The Python package is located at the ``python`` subdirectory. Run Pip to install the Python
+package. The Python package will re-use the native library built in Step 1.
+
+.. code-block:: bash
+
+  cd python
+  pip install .  # will re-use libtl2cgen.so

docs/treelite-migration.rst

@@ -0,0 +1,117 @@
+=================================================
+Migration Guide: How to migrate from Treelite 3.x
+=================================================
+
+Starting from 4.0 version, Treelite no longer supports compiling tree models into
+C code. That part of Treelite has been migrated to TL2cgen.
+
+In this guide, we will show you how to migrate your code to use TL2cgen instead of
+Treelite version 3.x.
+
+.. contents::
+  :local:
+
+Model loading
+=============
+The logic for loading tree model into memory remains identical. Keep using Treelite
+for model loading.
+
+.. code-block:: python
+  :caption: Load XGBoost model from files
+
+  model = treelite.Model.load("xgboost_model.bin", model_format="xgboost")
+  # Or
+  model = treelite.Model.load("xgboost_model.json", model_format="xgboost_json")
+  # Or
+  model = treelite.Model.from_xgboost(bst)  # bst is xgboost.Booster type
+
+.. code-block:: python
+  :caption: Load LightGBM model
+
+  model = treelite.Model.load("lightgbm_model.txt", model_format="lightgbm")
+  # Or
+  model = treelite.Model.from_lightgbm(bst)  # bst is lightgbm.Booster type
+
+.. code-block:: python
+  :caption: Load scikit-learn model
+
+  model = treelite.sklearn.import_model(clf)  # clf is a scikit-learn model object
+
+
+.. code-block:: python
+  :caption: Load a tree model using the model builder interface
+
+  builder = treelite.ModelBuilder(num_feature=3)
+
+  tree = treelite.ModelBuilder.Tree()
+  tree[0].set_numerical_test_node(
+    feature_id=0,
+    opname="<",
+    threshold=5.0,
+    default_left=True,
+    left_child_key=1,
+    right_child_key=2
+  )
+  tree[1].set_leaf_node(0.6)
+  tree[2].set_leaf_node(-0.4)
+  tree[0].set_root()
+
+  builder.append(tree)
+  model = builder.commit()  # Obtain treelite.Model object
+
+Generating C code
+=================
+
+For generating C code from tree models, replace :py:meth:`treelite.Model.compile`
+with :py:meth:`tl2cgen.generate_c_code`
+
+.. code-block:: python
+
+  # Before: Treelite 3.x
+  model.compile(dirpath="./code_dir", params={})
+
+  # After: TL2cgen. Note the model object being passed as the first argument
+  tl2cgen.generate_c_code(model, dirpath="./code_dir", params={})
+
+Exporting libraries
+===================
+Replace :py:meth:`treelite.Model.export_lib` with :py:meth:`tl2cgen.export_lib`:
+
+.. code-block:: python
+
+  # Before: Treelite 3.x
+  model.export_lib(toolchain="msvc", libpath="./mymodel.dll", params={})
+
+  # After: TL2cgen. The model object is passed as the first argument
+  tl2cgen.export_lib(model, toolchain="msvc", libpath="./mymodel.dll", params={})
+
+:py:meth:`treelite.Model.export_srcpkg` is replaced with :py:meth:`tl2cgen.export_srcpkg`.
+Note that the parameter ``platform`` was removed in :py:meth:`tl2cgen.export_srcpkg`.
+
+.. code-block:: python
+
+  # Before: Treelite 3.x
+  model.export_srcpkg(platform="unix", toolchain="gcc", pkgpath="./mymodel_pkg.zip",
+                      libname="mymodel.so", params={})
+
+  # After: TL2cgen. The model object is passed as the first argument
+  # 'platform' parameter is removed.
+  tl2cgen.export_srcpkg(model, toolchain="gcc", pkgpath="./mymodel_pkg.zip",
+                        libname="mymodel.so", params={})
+
+Annotating branches
+===================
+Replace :py:class:`treelite.Annotator` with :py:meth:`tl2cgen.annotate_branch`.
+Instead of calling two methods :py:meth:`treelite.Annotator.annotate_branch` and
+:py:meth:`treelite.Annotator.save`, you only need to call one,
+:py:meth:`tl2cgen.annotate_branch`:
+
+.. code-block:: python
+
+  # Before: Treelite 3.x
+  annotator = treelite.Annotator()
+  annotator.annotate_branch(model, dmat)
+  annotator.save(path="mymodel-annotation.json")
+
+  # After: TL2cgen. Only one method call is needed.
+  tl2cgen.annotate_branch(model, dmat, path="mymodel-annotation.json")