Python API typing, lint, config/makefile

.github/workflows/ci-workflow.yml

@@ -24,7 +24,7 @@ concurrency:
 jobs:
   gcc-build-test:
     name: gcc build & test
-    needs: [clang-formatting-check, sanity-checks]
+    needs: [clang-formatting-check, sanity-checks, python-lint-check]
     runs-on: kuzu-self-hosted-testing
     env:
       NUM_THREADS: 32
@@ -95,10 +95,10 @@ jobs:
 
       - name: Start Docker container
         run: |
-         docker run -d --name kuzu-x86 \
-                    -v $PWD:/kuzu -w /kuzu \
-                    -e NUM_THREADS=2 -e GEN=ninja -e CC=gcc -e CXX=g++ \
-                    i386/debian:latest tail -f /dev/null
+          docker run -d --name kuzu-x86 \
+                     -v $PWD:/kuzu -w /kuzu \
+                     -e NUM_THREADS=2 -e GEN=ninja -e CC=gcc -e CXX=g++ \
+                     i386/debian:latest tail -f /dev/null
 
       - name: Install dependencies
         run: |
@@ -161,7 +161,7 @@ jobs:
 
   clang-build-test:
     name: clang build and test
-    needs: [clang-formatting-check, sanity-checks]
+    needs: [clang-formatting-check, sanity-checks, python-lint-check]
     runs-on: kuzu-self-hosted-testing
     env:
       NUM_THREADS: 32
@@ -209,7 +209,7 @@ jobs:
 
   msvc-build-test:
     name: msvc build & test
-    needs: [clang-formatting-check, sanity-checks]
+    needs: [clang-formatting-check, sanity-checks, python-lint-check]
     runs-on: self-hosted-windows
     env:
       # Shorten build path as much as possible
@@ -343,6 +343,23 @@ jobs:
       - name: Check extension format
         run: python3 scripts/run-clang-format.py --clang-format-executable /usr/bin/clang-format-11 -r extension/
 
+  python-lint-check:
+    name: python lint check
+    runs-on: macos-14
+    steps:
+      - uses: actions/checkout@v3
+
+      - name: Setup Python
+        uses: actions/setup-python@v2
+        with:
+          python-version: "3.11"
+
+      - name: Run Python lint
+        working-directory: tools/python_api
+        run: |
+          make requirements
+          ./.venv/bin/ruff check src_py test --verbose
+
   rustfmt-check:
     name: rustfmt check
     runs-on: ubuntu-22.04
@@ -399,7 +416,7 @@ jobs:
 
   macos-build-test:
     name: apple clang build & test
-    needs: [clang-formatting-check, sanity-checks, rustfmt-check]
+    needs: [clang-formatting-check, sanity-checks, rustfmt-check, python-lint-check]
     runs-on: self-hosted-mac-x64
     env:
       NUM_THREADS: 32

.gitignore

@@ -17,6 +17,7 @@ compile_commands.json
 ### Python
 # Byte-compiled / optimized / DLL files
 __pycache__/
+.venv/
 *.py[cod]
 *$py.class
 cmake-build-debug/

scripts/pip-package/package_tar.py

@@ -9,21 +9,23 @@
 
 base_dir = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 
+
 def _get_kuzu_version():
-    cmake_file = os.path.abspath(os.path.join(base_dir, '..', 'CMakeLists.txt'))
+    cmake_file = os.path.abspath(os.path.join(base_dir, "..", "CMakeLists.txt"))
     with open(cmake_file) as f:
         for line in f:
-            if line.startswith('project(Kuzu VERSION'):
-                raw_version = line.split(' ')[2].strip()
-                version_nums = raw_version.split('.')
+            if line.startswith("project(Kuzu VERSION"):
+                raw_version = line.split(" ")[2].strip()
+                version_nums = raw_version.split(".")
                 if len(version_nums) <= 3:
                     return raw_version
                 else:
                     dev_suffix = version_nums[3]
-                    version = '.'.join(version_nums[:3])
+                    version = ".".join(version_nums[:3])
                     version += ".dev%s" % dev_suffix
                     return version
 
+
 if __name__ == "__main__":
     if len(sys.argv) == 2:
         file_name = sys.argv[1]
@@ -53,8 +55,22 @@ def _get_kuzu_version():
         os.makedirs(os.path.join(tempdir, "kuzu"))
         for path in ["setup.py", "setup.cfg", "MANIFEST.in"]:
             shutil.copy2(path, os.path.join(tempdir, path))
-        shutil.copy2("../../LICENSE", os.path.join(tempdir, "LICENSE.txt"))
+        shutil.copy2("../../LICENSE", os.path.join(tempdir, "LICENSE"))
         shutil.copy2("../../README.md", os.path.join(tempdir, "README.md"))
+
+        shutil.copy2(
+            "../../tools/python_api/pyproject.toml",
+            os.path.join(tempdir, "pyproject.toml"),
+        )
+        # Update the version in pyproject.toml
+        with open(os.path.join(tempdir, "pyproject.toml"), "r") as f:
+            lines = f.readlines()
+        with open(os.path.join(tempdir, "pyproject.toml"), "w") as f:
+            for line in lines:
+                if line.startswith("version ="):
+                    f.write('version = "%s"\n' % _get_kuzu_version())
+                else:
+                    f.write(line)
         shutil.copy2("README.md", os.path.join(tempdir, "README_PYTHON_BUILD.md"))
 
         subprocess.check_call([sys.executable, "setup.py", "egg_info"], cwd=tempdir)

tools/python_api/Makefile

@@ -0,0 +1,42 @@
+.DEFAULT_GOAL := help
+
+PYTHONPATH=
+SHELL=/bin/bash
+VENV=.venv
+
+ifeq ($(OS),Windows_NT)
+	VENV_BIN=$(VENV)/Scripts
+else
+	VENV_BIN=$(VENV)/bin
+endif
+
+.venv:  ## Set up a Python virtual environment and install dev packages
+	python3 -m venv $(VENV)
+	$(MAKE) requirements
+
+.PHONY: requirements
+requirements: .venv  ## Install/update Python dev packages
+	@unset CONDA_PREFIX \
+	&& $(VENV_BIN)/python -m pip install -U uv \
+	&& $(VENV_BIN)/uv pip install --upgrade -r requirements_dev.txt \
+
+.PHONY: lint
+lint:  ## Apply autoformatting and linting rules
+	$(VENV_BIN)/ruff check src_py test
+	$(VENV_BIN)/ruff format src_py test
+	-$(VENV_BIN)/mypy src_py test
+
+.PHONY: build
+build:  ## Compile kuzu (and install in 'build') for Python
+	$(MAKE) -C ../../ python
+	cp src_py/*.py build/kuzu/
+
+.PHONY: test
+test:  ## Run the Python unit tests
+	cp src_py/*.py build/kuzu/ && cd build
+	$(VENV_BIN)/pytest test
+
+.PHONY: help
+help:  ## Display this help information
+	@echo -e "\033[1mAvailable commands:\033[0m"
+	@grep -E '^[a-z.A-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | awk 'BEGIN {FS = ":.*?## "}; {printf "  \033[36m%-18s\033[0m %s\n", $$1, $$2}' | sort

tools/python_api/pyproject.toml

@@ -0,0 +1,97 @@
+[project]
+name = "kuzu"
+description = "Highly scalable, extremely fast, easy-to-use embeddable graph database"
+readme = "README.md"
+license = { file = "LICENSE" }
+keywords = ["graph", "database"]
+version = "0.0.1"
+
+[project.urls]
+Homepage = "https://kuzudb.com/"
+Documentation = "https://docs.kuzudb.com/"
+Repository = "https://github.com/kuzudb/kuzu"
+Changelog = "https://github.com/kuzudb/kuzu/releases"
+
+[tool.mypy]
+files = ["src_py", "test"]
+strict = true
+enable_error_code = [
+  "redundant-expr",
+  "truthy-bool",
+  "ignore-without-code",
+]
+disable_error_code = [
+  "import-not-found",
+  "no-any-return",
+]
+
+[[tool.mypy.overrides]]
+module = [
+  "_kuzu.*",
+  "networkx.*",
+  "pandas.*",
+  "pyarrow.*",
+  "pytz.*",
+  "torch_geometric.*",
+]
+ignore_missing_imports = true
+
+[tool.ruff]
+line-length = 119
+preview = true
+fix = true
+
+[tool.ruff.lint]
+select = [
+  "E", # pycodestyle
+  "W", # pycodestyle
+  "F", # Pyflakes
+  "B", # flake8-bugbear
+  "C4", # flake8-comprehensions
+  "D", # flake8-docstrings
+  "D213", # Augment NumPy docstring convention: Multi-line docstring summary should start at the second line
+  "D417", # Augment NumPy docstring convention: Missing argument descriptions
+  "I", # isort
+  "SIM", # flake8-simplify
+  "TCH", # flake8-type-checking
+  "TID", # flake8-tidy-imports
+  "UP", # pyupgrade
+  "PT", # flake8-pytest-style
+  "RUF", # Ruff-specific rules
+  "PTH", # flake8-use-pathlib
+  "FA", # flake8-future-annotations
+  "PIE", # flake8-pie
+  "TD", # flake8-todos
+  "TRY", # tryceratops
+  "EM", # flake8-errmsg
+  "FBT001", # flake8-boolean-trap
+]
+ignore = [
+  "D401", # First line should be in imperative mood
+  "D105", # Missing docstring in magic method
+  "D100", # Missing docstring in public module
+  "D104", # Missing docstring in public package
+  "D205", # 1 blank line required between summary line and description
+  "TD002", # Missing author in TODO
+  "TD003", # Missing issue link on the line following TODO
+]
+
+[tool.ruff.lint.per-file-ignores]
+"test/**/*.py" = ["D100", "D102", "D103", "E501", "F841", "TCH002"]
+"src_py/torch_geo*.py" = ["E501", "FBT001"]
+
+[tool.ruff.lint.pycodestyle]
+max-doc-length = 119
+
+[tool.ruff.lint.pydocstyle]
+convention = "numpy"
+
+[tool.ruff.lint.flake8-type-checking]
+strict = true
+
+[tool.ruff.format]
+docstring-code-format = true
+
+[build-system]
+requires = ["setuptools", "wheel"]
+build-backend = "setuptools.build_meta"

tools/python_api/requirements_dev.txt

@@ -1,8 +1,16 @@
-pybind11>=2.6.0
-pytest
+# required for tests
+networkx~=3.0
+numpy
 pandas
 polars
-networkx~=3.0.0
-numpy
-pyarrow==14.0.1
+pyarrow>=14
+pybind11>=2.6.0
+pytest
+torch
 torch-geometric~=2.3.0
+setuptools~=69.0
+
+# required for lint/formatting
+ruff==0.3.2
+mypy==1.9.0
+uv==0.1.17

tools/python_api/src_cpp/py_connection.cpp

@@ -165,7 +165,7 @@ std::unordered_map<std::string, std::unique_ptr<Value>> transformPythonParameter
         if (!py::isinstance<py::str>(key)) {
             // TODO(Chang): remove ROLLBACK once we can guarantee database is deleted after conn
             conn->query("ROLLBACK");
-            throw std::runtime_error("Parameter name must be of type string but get " +
+            throw std::runtime_error("Parameter name must be of type string but got " +
                                      py::str(key.get_type()).cast<std::string>());
         }
         auto name = key.cast<std::string>();

tools/python_api/src_py/__init__.py

@@ -1,5 +1,5 @@
 """
-# Kùzu Python API bindings
+# Kùzu Python API bindings.
 
 This package provides a Python API for Kùzu graph database management system.
 
@@ -12,7 +12,7 @@
 ```python
 import kuzu
 
-db = kuzu.Database('./test')
+db = kuzu.Database("./test")
 conn = kuzu.Connection(db)
 
 # Define the schema
@@ -28,7 +28,7 @@
 conn.execute('COPY LivesIn FROM "lives-in.csv"')
 
 # Query the data
-results = conn.execute('MATCH (u:User) RETURN u.name, u.age;')
+results = conn.execute("MATCH (u:User) RETURN u.name, u.age;")
 while results.has_next():
     print(results.get_next())
 ```
@@ -37,29 +37,44 @@
 
 """
 
-import sys
+from __future__ import annotations
+
 import os
+import sys
 
 # Set RTLD_GLOBAL and RTLD_LAZY flags on Linux to fix the issue with loading
 # extensions
 if sys.platform == "linux":
     original_dlopen_flags = sys.getdlopenflags()
     sys.setdlopenflags(os.RTLD_GLOBAL | os.RTLD_LAZY)
 
-from .database import *
-from .connection import *
-from .query_result import *
-from .types import *
+from .connection import Connection
+from .database import Database
+from .prepared_statement import PreparedStatement
+from .query_result import QueryResult
+from .types import Type
 
-def __getattr__(name):
+
+def __getattr__(name: str) -> str | int:
     if name == "version":
         return Database.get_version()
     elif name == "storage_version":
         return Database.get_storage_version()
     else:
-        raise AttributeError(f"module {__name__!r} has no attribute {name!r}")
-
+        msg = f"module {__name__!r} has no attribute {name!r}"
+        raise AttributeError(msg)
+
 
 # Restore the original dlopen flags
 if sys.platform == "linux":
     sys.setdlopenflags(original_dlopen_flags)
+
+__all__ = [
+    "Connection",
+    "Database",
+    "PreparedStatement",
+    "QueryResult",
+    "Type",
+    "storage_version",
+    "version",
+]