don't install everything on lint step

Author: PastelStorm

.github/workflows/ci.yml

@@ -23,8 +23,8 @@ jobs:
         cache-dependency-glob: "uv.lock"
     - name: Set up Python ${{ env.PYTHON_VERSION }}
       run: uv python install ${{ env.PYTHON_VERSION }}
-    - name: Install dependencies
-      run: uv sync --extra test --frozen
+    - name: Install lint dependencies
+      run: uv sync --only-group lint --frozen
     - name: Lint
       run: make check
 
@@ -52,7 +52,7 @@ jobs:
     - name: Install dependencies and run core tests
       run: |
         sudo apt-get update && sudo apt-get install --yes poppler-utils libreoffice
-        uv sync --extra test --frozen
+        uv sync --group test --frozen
         make install-pandoc
         make install-nltk-models
         sudo add-apt-repository -y ppa:alex-p/tesseract-ocr5
@@ -114,6 +114,6 @@ jobs:
         df -h
     - name: Test Dockerfile
       run: |
-        uv sync --extra test --frozen
+        uv sync --group test --frozen
         make docker-build
         make docker-test

.github/workflows/docker-publish.yml

@@ -82,7 +82,7 @@ jobs:
     - name: Set up Python ${{ env.PYTHON_VERSION }}
       run: uv python install ${{ env.PYTHON_VERSION }}
     - name: Install test dependencies
-      run: uv sync --extra test --frozen
+      run: uv sync --group test --frozen
     - name: Test image
       run: |
         export DOCKER_IMAGE="$DOCKER_BUILD_REPOSITORY:${{ matrix.arch }}-$SHORT_SHA"

CHANGELOG.md

@@ -8,6 +8,11 @@
 * Use `.python-version` file as single source of truth for Python version across all CI workflows
 * Re-enable arm64 Docker image builds using a dedicated ARM runner (`opensource-linux-arm64-4core`), restoring multiarch support for both amd64 and arm64
 * Switch all CI workflows to faster self-hosted runners (`opensource-linux-8core`)
+* Split lint tools into a lightweight dependency group so the CI lint step no longer installs heavy runtime dependencies
+* Add explicit dependencies for `backoff`, `pandas`, `psutil`, `pypdf`, and `requests` (previously only transitive via `unstructured[all-docs]`)
+* Pre-download NLTK models before parallel test runs to prevent race conditions
+* Parallelize Docker smoke tests with pytest-xdist
+* Remove unused `ARCH` variable from Makefile
 
 ## 0.0.93
 * Refactored the Dockerfile to use the chainguard/wolfi-base image instead of the unstructured/base-image. This is to align with the recent change in the unstructured repo where the same change was made.

Makefile

@@ -25,7 +25,7 @@ install-base-packages:
 
 .PHONY: install-test
 install-test:
-	uv sync --extra test --frozen
+	uv sync --group test --frozen
 
 .PHONY: install-nltk-models
 install-nltk-models:

pyproject.toml

@@ -14,11 +14,14 @@ dependencies = [
     "requests >=2.32.0, <3.0.0",
 ]
 
-[project.optional-dependencies]
-test = [
+[dependency-groups]
+lint = [
     "ruff >=0.15.0, <1.0.0",
     "mypy >=1.19.1, <2.0.0",
     "types-requests >=2.32.4, <3.0.0",
+]
+test = [
+    {include-group = "lint"},
     "pytest >=9.0.2, <10.0.0",
     "pytest-cov >=7.0.0, <8.0.0",
     "pytest-mock >=3.15.1, <4.0.0",