Merge branch 'main' into tvf_impl

Author: paultiq

.githooks/post-checkout

@@ -1,6 +1,3 @@
-#!/bin/sh
-if [ "$3" = "1" ]; then
-    echo "Updating submodules..."
-    git submodule update --init --recursive
-fi
+#!/usr/bin/env bash
 
+git submodule update --init --recursive

.github/workflows/code_quality.yml

@@ -0,0 +1,45 @@
+name: Code Quality Checks
+on:
+  workflow_dispatch:
+    inputs:
+      git_ref:
+        type: string
+        description: Git ref of the DuckDB python package
+        required: false
+  workflow_call:
+    inputs:
+      git_ref:
+        type: string
+        description: Git ref of the DuckDB python package
+        required: false
+
+defaults:
+  run:
+    shell: bash
+
+jobs:
+  run_checks:
+    name: Run linting, formatting and static type checker
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+        with:
+          ref: ${{ inputs.git_ref }}
+          fetch-depth: 0
+          persist-credentials: false
+
+      - name: Install Astral UV
+        uses: astral-sh/setup-uv@v6
+        with:
+          version: "0.7.14"
+          python-version: 3.9
+
+      - name: pre-commit (cache)
+        uses: actions/cache@v4
+        with:
+          path: ~/.cache/pre-commit
+          key: pre-commit-${{ hashFiles('.pre-commit-config.yaml') }}
+
+      - name: pre-commit (--all-files)
+        run: |
+          uvx pre-commit run --show-diff-on-failure --color=always --all-files

.github/workflows/coverage.yml

@@ -151,4 +151,4 @@ jobs:
           echo "### C++ Coverage Summary" >> $GITHUB_STEP_SUMMARY
           echo '```' >> $GITHUB_STEP_SUMMARY
           echo "$SUMMARY_CPP" >> $GITHUB_STEP_SUMMARY
-          echo '```' >> $GITHUB_STEP_SUMMARY
+          echo '```' >> $GITHUB_STEP_SUMMARY

.github/workflows/on_pr.yml

@@ -23,9 +23,14 @@ jobs:
     name: Make sure submodule is in a sane state
     uses: ./.github/workflows/submodule_sanity.yml
 
+  code_quality:
+    name: Code-quality checks
+    needs: submodule_sanity_guard
+    uses: ./.github/workflows/code_quality.yml
+
   packaging_test:
     name: Build a minimal set of packages and run all tests on them
-    needs: submodule_sanity_guard
+    needs: code_quality
     # Skip packaging tests for draft PRs
     if: ${{ github.event_name != 'pull_request' || github.event.pull_request.draft == false }}
     uses: ./.github/workflows/packaging.yml
@@ -36,7 +41,7 @@ jobs:
 
   coverage_test:
     name: Run coverage tests
-    needs: submodule_sanity_guard
+    needs: code_quality
     # Only run coverage test for draft PRs
     if: ${{ github.event_name == 'pull_request' && github.event.pull_request.draft == true }}
     uses: ./.github/workflows/coverage.yml

.github/workflows/on_push.yml

@@ -18,8 +18,13 @@ concurrency:
   cancel-in-progress: true
 
 jobs:
+  code_quality:
+    name: Code-quality checks
+    uses: ./.github/workflows/code_quality.yml
+
   test:
     name: Run coverage tests
+    needs: code_quality
     uses: ./.github/workflows/coverage.yml
     with:
       git_ref: ${{ github.ref }}

.github/workflows/packaging_wheels.yml

@@ -30,7 +30,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python: [ cp39, cp310, cp311, cp312, cp313, cp314, cp314t ]
+        python: [ cp39, cp310, cp311, cp312, cp313, cp314 ]
         platform:
           - { os: windows-2025,     arch: amd64,      cibw_system: win }
           - { os: ubuntu-24.04,     arch: x86_64,     cibw_system: manylinux }
@@ -45,10 +45,6 @@ jobs:
           - { minimal: true, python: cp311 }
           - { minimal: true, python: cp312 }
           - { minimal: true, platform: { arch: universal2 } }
-          # Windows+cp314t disabled due to test failures in CI. 
-          # TODO: Diagnose why tests fail (access violations) in some configurations
-          - { python: cp314t, platform: { os: windows-2025 } }  
-
     runs-on: ${{ matrix.platform.os }}
     env:
       CIBW_TEST_SKIP: ${{ inputs.testsuite == 'none' && '*' || '*-macosx_universal2' }}
@@ -94,9 +90,7 @@ jobs:
         env:
           CIBW_ARCHS: ${{ matrix.platform.arch == 'amd64' && 'AMD64' || matrix.platform.arch }}
           CIBW_BUILD: ${{ matrix.python }}-${{ matrix.platform.cibw_system }}_${{ matrix.platform.arch }}
-          # PYTHON_GIL=1: Suppresses the RuntimeWarning that the GIL is enabled on free-threaded builds. 
-          # TODO: Remove PYTHON_GIL=1 when free-threaded is supported. 
-          CIBW_ENVIRONMENT: PYTHON_GIL=1
+
       - name: Upload wheel
         uses: actions/upload-artifact@v4
         with:

.pre-commit-config.yaml

@@ -0,0 +1,50 @@
+repos:
+  - repo: https://github.com/pre-commit/pre-commit-hooks
+    rev: v4.4.0
+    hooks:
+      - id: check-yaml
+        args: ["--allow-multiple-documents"]
+      - id: check-toml
+      - id: check-added-large-files
+      - id: detect-private-key
+      - id: check-merge-conflict
+      - id: forbid-new-submodules
+
+  - repo: https://github.com/astral-sh/ruff-pre-commit
+    # Ruff version.
+    rev: v0.13.3
+    hooks:
+      # Run the linter.
+      - id: ruff-check
+      # Run the formatter.
+      - id: ruff-format
+
+  - repo: https://github.com/pre-commit/mirrors-clang-format
+    rev: v21.1.2    # pick the version of clang-format you want
+    hooks:
+      - id: clang-format
+        files: \.(c|cpp|cc|h|hpp|cxx|hxx)$
+
+  - repo: https://github.com/cheshirekow/cmake-format-precommit
+    rev: v0.6.13
+    hooks:
+      - id: cmake-format
+
+  - repo: https://github.com/pre-commit/mirrors-mypy
+    rev: v1.18.2
+    hooks:
+      - id: mypy
+        entry: mypy -v
+        files: ^(duckdb/|_duckdb-stubs/)
+        exclude: ^duckdb/(experimental|query_graph)/
+        additional_dependencies: [ numpy, polars ]
+
+  - repo: local
+    hooks:
+      - id: post-checkout-submodules
+        name: Update submodule post-checkout
+        entry: .githooks/post-checkout
+        language: script
+        stages: [ post-checkout ]
+        pass_filenames: false
+        always_run: true

CHANGELOG.md

@@ -0,0 +1,19 @@
+# Changelog
+
+## v1.4.1
+**DuckDB Core**: v1.4.1
+
+### Bug Fixes
+- **ADBC Driver**: Fixed ADBC driver implementation (#81)
+- **SQLAlchemy compatibility**: Added `__hash__` method overload (#61)
+- **Error Handling**: Reset PyErr before throwing Python exceptions (#69)
+- **Polars Lazyframes**: Fixed Polars expression pushdown (#102)
+
+### Code Quality Improvements & Developer Experience
+- **MyPy Support**: MyPy is functional again and better integrated with the dev workflow
+- **Stubs**: Re-created and manually curated stubs for the binary extension
+- **Type Shadowing**: Deprecated `typing` and `functional` modules
+- **Linting & Formatting**: Comprehensive code quality improvements with Ruff
+- **Type Annotations**: Added missing overloads and improved type coverage
+- **Pre-commit Integration**: Added ruff, clang-format, cmake-format and mypy configs
+- **CI/CD**: Added code quality workflow

CMakeLists.txt

@@ -48,50 +48,84 @@ duckdb_add_library(duckdb_target)
 
 # Bundle in INTERFACE library
 add_library(_duckdb_dependencies INTERFACE)
-target_link_libraries(_duckdb_dependencies INTERFACE
-        pybind11::pybind11
-        duckdb_target
-)
+target_link_libraries(_duckdb_dependencies INTERFACE pybind11::pybind11
+                                                     duckdb_target)
 # Also add include directory
-target_include_directories(_duckdb_dependencies INTERFACE
-        $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/src/duckdb_py/include>
+target_include_directories(
+  _duckdb_dependencies
+  INTERFACE $<BUILD_INTERFACE:${CMAKE_CURRENT_SOURCE_DIR}/src/duckdb_py/include>
 )
 
 # ────────────────────────────────────────────
 # Descend into the real DuckDB‑Python sources
 # ────────────────────────────────────────────
 add_subdirectory(src/duckdb_py)
 
-pybind11_add_module(_duckdb
-        $<TARGET_OBJECTS:python_src>
-        $<TARGET_OBJECTS:python_arrow>
-        $<TARGET_OBJECTS:python_common>
-        $<TARGET_OBJECTS:python_functional>
-        $<TARGET_OBJECTS:python_jupyter>
-        $<TARGET_OBJECTS:python_native>
-        $<TARGET_OBJECTS:python_numpy>
-        $<TARGET_OBJECTS:python_pandas>
-        $<TARGET_OBJECTS:python_pybind11>
-        $<TARGET_OBJECTS:python_connection>
-        $<TARGET_OBJECTS:python_expression>
-        $<TARGET_OBJECTS:python_relation>
-        $<TARGET_OBJECTS:python_type>
-)
+pybind11_add_module(
+  _duckdb
+  $<TARGET_OBJECTS:python_src>
+  $<TARGET_OBJECTS:python_arrow>
+  $<TARGET_OBJECTS:python_common>
+  $<TARGET_OBJECTS:python_functional>
+  $<TARGET_OBJECTS:python_jupyter>
+  $<TARGET_OBJECTS:python_native>
+  $<TARGET_OBJECTS:python_numpy>
+  $<TARGET_OBJECTS:python_pandas>
+  $<TARGET_OBJECTS:python_pybind11>
+  $<TARGET_OBJECTS:python_connection>
+  $<TARGET_OBJECTS:python_expression>
+  $<TARGET_OBJECTS:python_relation>
+  $<TARGET_OBJECTS:python_type>)
 # add _duckdb_dependencies
 target_link_libraries(_duckdb PRIVATE _duckdb_dependencies)
 
+# ────────────────────────────────────────────
+# Controlling symbol export
+#
+# We want to export exactly two symbols: - PyInit__duckdb: this allows CPython
+# to load the module - duckdb_adbc_init: the DuckDB ADBC driver
+#
+# The export of symbols on OSX and Linux is controlled by: - Visibility
+# annotations in the code (for this lib we use the PYBIND11_EXPORT macro) -
+# Telling the linker which symbols we want exported, which we do below
+#
+# For Windows, we rely on just the visbility annotations.
+# ────────────────────────────────────────────
+set_target_properties(
+  _duckdb
+  PROPERTIES CXX_VISIBILITY_PRESET hidden
+             C_VISIBILITY_PRESET hidden
+             VISIBILITY_INLINES_HIDDEN ON)
+
+if(APPLE)
+  target_link_options(
+    _duckdb PRIVATE "LINKER:-exported_symbol,_duckdb_adbc_init"
+    "LINKER:-exported_symbol,_PyInit__duckdb")
+elseif(UNIX AND NOT APPLE)
+  target_link_options(
+    _duckdb PRIVATE "LINKER:--export-dynamic-symbol=duckdb_adbc_init"
+    "LINKER:--export-dynamic-symbol=PyInit__duckdb")
+elseif(WIN32)
+  target_link_options(_duckdb PRIVATE "/EXPORT:duckdb_adbc_init"
+                      "/EXPORT:PyInit__duckdb")
+endif()
+
 # ────────────────────────────────────────────
 # Put the object file in the correct place
 # ────────────────────────────────────────────
 
-# If we're not building through scikit-build-core then we have to set a different dest dir
+# If we're not building through scikit-build-core then we have to set a
+# different dest dir
 include(GNUInstallDirs)
 if(DEFINED SKBUILD_PLATLIB_DIR)
   set(_DUCKDB_PY_INSTALL_DIR "${SKBUILD_PLATLIB_DIR}")
 elseif(DEFINED Python_SITEARCH)
   set(_DUCKDB_PY_INSTALL_DIR "${Python_SITEARCH}")
 else()
-  message(WARNING "Could not determine Python install dir. Falling back to CMAKE_INSTALL_LIBDIR.")
+  message(
+    WARNING
+      "Could not determine Python install dir. Falling back to CMAKE_INSTALL_LIBDIR."
+  )
   set(_DUCKDB_PY_INSTALL_DIR "${CMAKE_INSTALL_LIBDIR}")
 endif()