build setup (#33)

* build setup

* fixes

.dockerignore

@@ -0,0 +1,12 @@
+.git/
+.github/
+.idea/
+.pytest_cache/
+.venv/
+build/
+dist/
+cython_debug/
+htmlcov/
+*.egg-info/
+.coverage
+.DS_Store

.gitignore

@@ -145,4 +145,7 @@ cython_debug/
 
 # ignore cython autogenerated files
 skranger/ensemble/ranger.cpp
-skranger/ensemble/ranger.html
+skranger/ensemble/ranger.html
+
+# src copy to deal with symlinks
+skranger/ensemble/ranger/

Dockerfile

@@ -0,0 +1,39 @@
+# Use the python slim image
+FROM python:3.8.3-slim
+
+# Add build deps for python packages
+# curl to install vendored poetry
+# g++ to build sksurv
+RUN apt-get update && \
+    apt-get install curl g++ -y && \
+    apt-get clean
+
+# Set the working directory to app
+WORKDIR /app
+
+# Set the poetry version explicitly
+ENV POETRY_VERSION=1.0.5
+# Unbuffer the logger so we always get logs
+ENV PYTHONUNBUFFERED=1
+# Update the path for poetry python
+ENV PATH=/root/.poetry/bin:/root/.local/bin:$PATH
+
+# Install vendored poetry
+RUN curl -sSL https://raw.githubusercontent.com/python-poetry/poetry/master/get-poetry.py | python
+
+# Add dep configs
+ADD ./pyproject.toml .
+ADD ./poetry.lock .
+
+# Install packages
+# Disable virtualenv creation so we just use the already-installed python
+RUN poetry config virtualenvs.create false && \
+    poetry run pip install -U pip && \
+    poetry install && \
+    rm -r ~/.cache/pip
+
+# Add everything
+ADD . .
+
+# Set the entrypoint to poetry
+ENTRYPOINT ["poetry"]

Makefile

@@ -3,6 +3,22 @@ build:
 	poetry run python build.py clean
 	poetry run python build.py build_ext --inplace --force
 
+.PHONY: copy
+copy:
+	poetry run python buildpre.py
+
+.PHONY: dist
+dist: copy
+	poetry build
+
+.PHONY: docker
+docker:
+	docker build -t skranger .
+
+.PHONY: linux
+linux: copy docker
+	docker run --rm -v $(shell pwd)/dist:/app/dist:rw skranger build
+
 .PHONY: docs
 docs:
 	cd docs && \
@@ -15,7 +31,20 @@ fmt:
 	poetry run isort -y
 	poetry run black .
 
+.PHONY: sdist
+sdist: copy
+	poetry build -f sdist
+
+.PHONY: setup
+setup:
+	git submodule init
+	git submodule update
+	asdf install
+	poetry install --no-root
+	poetry run python buildpre.py
+	poetry install
+
 .PHONY: test
 test:
 	poetry run pytest --cov=skranger --cov-report=html tests/
-	open htmlcov/index.html
+	open htmlcov/index.html

README.rst

@@ -126,3 +126,18 @@ License
 -------
 
 ``skranger`` is licensed under `GPLv3 <https://github.com/crflynn/skranger/blob/master/LICENSE.txt>`__.
+
+Development
+-----------
+
+To develop locally, it is recommended to have ``asdf``, ``make`` and a C++ compiler already installed. After cloning, run ``make setup``. This will setup the ranger submodule, install python and poetry from ``.tool-versions``, install dependencies using poetry, copy the ranger source code into skranger, and then build and install skranger in the local virtualenv.
+
+To format code, run ``make fmt``. This will run isort and black against the .py files.
+
+To run tests and inspect coverage, run ``make test``.
+
+To rebuild in place after making changes, run ``make build``.
+
+To create python package artifacts, run ``make dist``.
+
+To build and view documentation, run ``make docs``.

build.py

@@ -7,15 +7,16 @@
 
 # skranger project directory
 top = os.path.dirname(os.path.abspath(__file__))
-# the cpp source code
-ranger_src = os.path.join("ranger", "src")
 
 # include skranger, ranger, and numpy headers
+# requires running buildpre.py to find src in this location
 include_dirs = [
     top,
-    os.path.join(top, "skranger"),
     os.path.join(top, "skranger", "ensemble"),
-    os.path.join(top, ranger_src),
+    os.path.join(top, "skranger", "ensemble", "ranger", "src"),
+    os.path.join(top, "skranger", "ensemble", "ranger", "src", "Forest"),
+    os.path.join(top, "skranger", "ensemble", "ranger", "src", "Tree"),
+    os.path.join(top, "skranger", "ensemble", "ranger", "src", "utility"),
     np.get_include(),
 ]
 
@@ -58,7 +59,7 @@ def create_extension(module_name):
 
 setup(
     ext_modules=cythonize(
-        ext_modules, gdb_debug=False, force=True, annotate=True, compiler_directives={"language_level": "3"}
+        ext_modules, gdb_debug=False, force=True, annotate=False, compiler_directives={"language_level": "3"}
     )
 )
 

buildpre.py

@@ -0,0 +1,18 @@
+import os
+import shutil
+
+top = os.path.dirname(os.path.abspath(__file__))
+
+
+def copy_ranger_source():
+    """Copy the ranger cpp source, following symlinks."""
+    src = os.path.join(top, "ranger", "cpp_version")
+    dst = os.path.join(top, "skranger", "ensemble", "ranger")
+    try:
+        shutil.rmtree(dst)
+    except FileNotFoundError:
+        pass
+    shutil.copytree(src, dst, symlinks=False)
+
+
+copy_ranger_source()

pyproject.toml

@@ -41,7 +41,18 @@ readme = "README.rst"
 homepage = "https://github.com/crflynn/skranger"
 repository = "https://github.com/crflynn/skranger"
 documentation = "https://github.com/crflynn/skranger"
+include = ["skranger/ensemble/ranger/**/*"]
 build = "build.py"
+keywords = ["random", "forest", "ranger", "machine", "learning"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Science/Research",
+    "License :: OSI Approved :: GNU General Public License v3 (GPLv3)",
+    "Programming Language :: C++",
+    "Programming Language :: Python",
+    "Programming Language :: Python :: 3",
+    "Topic :: Scientific/Engineering",
+]
 
 [tool.poetry.dependencies]
 python = "^3.6.1"
@@ -61,6 +72,6 @@ pandas = "^1.0.3"
 pytest-cov = "^2.9.0"
 
 [build-system]
-requires = ["poetry>=0.12"]
+requires = ["poetry>=0.12", "cython==3.0a5", "numpy>=1.18.3"]
 build-backend = "poetry.masonry.api"