Refactor package structure and update dependencies

.github/workflows/code_checks.yml

@@ -35,7 +35,7 @@ jobs:
         uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
-          version: "0.9.7"
+          version: "0.9.22"
           enable-cache: true
 
       - name: "Set up Python"
@@ -54,3 +54,5 @@ jobs:
         uses: pypa/gh-action-pip-audit@1220774d901786e6f652ae159f7b6bc8fea6d266
         with:
           virtual-environment: .venv/
+          ignore-vulns: |
+            GHSA-xm59-rqc7-hhvf

.github/workflows/publish.yml

@@ -22,7 +22,7 @@ jobs:
         uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
-          version: "0.9.7"
+          version: "0.9.22"
           enable-cache: true
 
       - name: "Set up Python"

.github/workflows/unit_tests.yml

@@ -43,7 +43,7 @@ jobs:
         uses: astral-sh/setup-uv@681c641aba71e4a1c380be3ab5e12ad51f415867
         with:
           # Install a specific version of uv.
-          version: "0.9.7"
+          version: "0.9.22"
           enable-cache: true
 
       - name: "Set up Python"

.pre-commit-config.yaml

@@ -1,6 +1,6 @@
 repos:
   - repo: https://github.com/pre-commit/pre-commit-hooks
-    rev: v5.0.0  # Use the ref you want to point at
+    rev: v6.0.0  # Use the ref you want to point at
     hooks:
     - id: trailing-whitespace
     - id: check-ast
@@ -11,7 +11,7 @@ repos:
     - id: end-of-file-fixer
     - id: mixed-line-ending
       args: [--fix=lf]
-    - id: check-byte-order-marker
+    - id: fix-byte-order-marker
     - id: check-merge-conflict
     - id: check-symlinks
     - id: detect-private-key
@@ -20,12 +20,12 @@ repos:
     - id: check-toml
 
   - repo: https://github.com/astral-sh/uv-pre-commit
-    rev: 0.6.12
+    rev: 0.9.22
     hooks:
       - id: uv-lock
 
   - repo: https://github.com/astral-sh/ruff-pre-commit
-    rev: 'v0.11.4'
+    rev: v0.14.10
     hooks:
     - id: ruff
       args: [--fix, --exit-non-zero-on-fix]
@@ -34,7 +34,7 @@ repos:
       types_or: [python, jupyter]
 
   - repo: https://github.com/pre-commit/mirrors-mypy
-    rev: v1.15.0
+    rev: v1.19.1
     hooks:
     - id: mypy
       entry: python3 -m mypy --config-file pyproject.toml
@@ -57,5 +57,5 @@ ci:
     autoupdate_branch: ''
     autoupdate_commit_msg: '[pre-commit.ci] pre-commit autoupdate'
     autoupdate_schedule: weekly
-    skip: [pytest,doctest,mypy]
+    skip: [mypy]
     submodules: false

aieng-eval-agents/aieng/topic/impl/__init__.py → aieng-eval-agents/aieng/agent_evals/impl/__init__.py

@@ -1,6 +1,6 @@
 """Example implementation for an example topic."""
 
-from aieng.topic.impl.example_impl import example_impl
+from aieng.agent_evals.impl.example_impl import example_impl
 
 
 __all__ = ["example_impl"]

aieng-eval-agents/pyproject.toml

@@ -1,116 +1,26 @@
-[build-system]
-requires = ["hatchling"]
-build-backend = "hatchling.build"
-
-[dependency-groups]
-dev = [
-    "codecov>=2.1.13",
-    "mypy>=1.14.1",
-    "nbqa>=1.9.1",
-    "pip-audit>=2.7.3",
-    "pre-commit>=4.1.0",
-    "pytest>=8.3.4",
-    "pytest-asyncio>=0.25.2",
-    "pytest-cov>=6.0.0",
-    "pytest-mock>=3.14.0",
-    "ruff>=0.9.2",
-]
-
 [project]
 name = "aieng-eval-agents"
 version = "0.1.0"
 description = "Agentic AI Evaluation reference implementation from Vector Institute AI Engineering"
 authors = [{name = "Vector AI Engineering", email = "ai_engineering@vectorinstitute.ai"}]
-requires-python = ">=3.11,<4.0"
+requires-python = ">=3.12,<4.0"
 readme = "README.md"
 license = "MIT"
-dependencies = []
+dependencies = [
+    "kagglehub>=0.3.13",
+]
+
+[dependency-groups]
+dev = [
+    "pytest>=8.3.5",
+]
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
 
 [tool.hatch.build.targets.sdist]
 include = ["aieng/"]
 
 [tool.hatch.build.targets.wheel]
 include = ["aieng/"]
-
-[tool.mypy]
-ignore_missing_imports = true
-install_types = true
-pretty = true
-namespace_packages = true
-explicit_package_bases = true
-non_interactive = true
-warn_unused_configs = true
-allow_any_generics = false
-allow_subclassing_any = false
-allow_untyped_calls = false
-allow_untyped_defs = false
-allow_incomplete_defs = false
-check_untyped_defs = true
-allow_untyped_decorators = false
-warn_redundant_casts = true
-warn_unused_ignores = true
-warn_return_any = true
-implicit_reexport = false
-strict_equality = true
-extra_checks = true
-
-[tool.ruff]
-include = ["*.py", "pyproject.toml", "*.ipynb"]
-line-length = 88
-
-[tool.ruff.format]
-quote-style = "double"
-indent-style = "space"
-docstring-code-format = true
-
-[tool.ruff.lint]
-select = [
-    "A", # flake8-builtins
-    "B", # flake8-bugbear
-    "COM", # flake8-commas
-    "C4", # flake8-comprehensions
-    "RET", # flake8-return
-    "SIM", # flake8-simplify
-    "ICN", # flake8-import-conventions
-    "Q", # flake8-quotes
-    "RSE", # flake8-raise
-    "D", # pydocstyle
-    "E", # pycodestyle
-    "F", # pyflakes
-    "I", # isort
-    "W", # pycodestyle
-    "N", # pep8-naming
-    "ERA", # eradicate
-    "PL", # pylint
-]
-fixable = ["A", "B", "COM", "C4", "RET", "SIM", "ICN", "Q", "RSE", "D", "E", "F", "I", "W", "N", "ERA", "PL"]
-ignore = [
-    "B905", # `zip()` without an explicit `strict=` parameter
-    "E501", # line too long
-    "D203", # 1 blank line required before class docstring
-    "D213", # Multi-line docstring summary should start at the second line
-    "PLR2004", # Replace magic number with named constant
-    "PLR0913", # Too many arguments
-    "COM812", # Missing trailing comma
-]
-
-# Ignore import violations in all `__init__.py` files.
-[tool.ruff.lint.per-file-ignores]
-"__init__.py" = ["E402", "F401", "F403", "F811"]
-
-[tool.ruff.lint.pep8-naming]
-ignore-names = ["X*", "setUp"]
-
-[tool.ruff.lint.isort]
-lines-after-imports = 2
-
-[tool.ruff.lint.pydocstyle]
-convention = "numpy"
-
-[tool.ruff.lint.pycodestyle]
-max-doc-length = 88
-
-[tool.pytest.ini_options]
-markers = [
-    "integration_test: marks tests as integration tests",
-]

aieng-eval-agents/tests/aieng/topic/impl/test_example_impl.py → aieng-eval-agents/tests/aieng/agent_evals/impl/test_example_impl.py

@@ -1,6 +1,6 @@
 """Test example implementation."""
 
-from aieng.topic.impl import example_impl
+from aieng.agent_evals.impl import example_impl
 
 
 def test_example_impl():

pyproject.toml

@@ -3,65 +3,58 @@ name = "agentic-ai-evaluation-bootcamp-202602"
 version = "0.1.0"
 description = "Vector Institute Agentic AI Evaluation Bootcamp 202602"
 readme = "README.md"
-authors = [ {name = "Vector AI Engineering", email = "ai_engineering@vectorinstitute.ai"}]
+authors = [ {name = "Vector AI Engineering", email = "ai_engineering@vectorinstitute.ai"} ]
 license = "Apache-2.0"
-requires-python = ">=3.12"
+requires-python = ">=3.12,<4.0"
 dependencies = [
-    "aiohttp>=3.12.14",
+    "aieng-eval-agents>=0.1.0",
+    "aiohttp>=3.13.3",
     "beautifulsoup4>=4.13.4",
     "datasets>=3.6.0",
-    "e2b-code-interpreter>=1.5.2",
-    "gradio>=5.37.0",
-    "langfuse>=3.1.3",
+    "e2b-code-interpreter>=2.4.1",
+    "gradio>=6.0.2",
+    "langfuse>=3.10.3",
     "lxml>=6.0.0",
     "nest-asyncio>=1.6.0",
     "numpy<2.3.0",
-    "openai>=1.93.1",
-    "openai-agents>=0.1.0",
-    "plotly>=6.2.0",
-    "pydantic>=2.11.7",
-    "pydantic-ai-slim[logfire]>=0.3.7",
-    "pytest-asyncio>=0.25.2",
+    "openai>=2.8.1",
+    "openai-agents>=0.6.1",
+    "plotly>=6.5.0",
+    "pydantic>=2.12.4",
+    "pydantic-ai-slim[logfire]>=1.26.0",
     "scikit-learn>=1.7.0",
-    "weaviate-client>=4.15.4",
-    "urllib3>=2.6.0",
+    "weaviate-client>=4.18.3",
+    "urllib3>=2.6.3",
 ]
 
-[build-system]
-requires = ["hatchling"]
-build-backend = "hatchling.build"
-
-[tool.hatch.build.targets.wheel]
-packages = ["src"]
-
 [dependency-groups]
 dev = [
-    "aieng-platform-onboard>=0.3.6",
-    "mypy>=1.14.1",
+    "aieng-platform-onboard>=0.4.0",
+    "mypy>=1.19.0",
     "codecov>=2.1.13",
-    "ipykernel>=6.29.5",
-    "ipython>=9.4.0",
+    "ipykernel>=7.1.0",
+    "ipython>=9.8.0",
     "ipywidgets>=8.1.7",
     "jupyter>=1.1.1",
     "jupyterlab>=4.4.8",
     "nbqa>=1.9.1",
-    "pip-audit>=2.7.3",
-    "pre-commit>=4.1.0",
+    "pip-audit>=2.9.0",
+    "pre-commit>=4.2.0",
     "pytest>=8.3.4",
-    "pytest-asyncio>=0.25.2",
-    "pytest-cov>=6.0.0",
+    "pytest-asyncio>=0.26.0",
+    "pytest-cov>=6.1.1",
     "pytest-mock>=3.14.0",
-    "ruff>=0.12.2",
-    "transformers>=4.54.1",
+    "ruff>=0.14.7",
+    "transformers>=4.57.3",
 ]
 docs = [
     "jinja2>=3.1.6", # Pinning version to address vulnerability GHSA-cpwx-vrp4-4pq7
     "mkdocs>=1.6.0",
     "mkdocs-material>=9.6.15",
-    "mkdocstrings>=0.24.1",
-    "mkdocstrings-python>=1.16.12",
-    "ipykernel>=6.29.5",
-    "ipython>=9.4.0",
+    "mkdocstrings>=1.0.0",
+    "mkdocstrings-python>=2.0.1",
+    "ipykernel>=7.1.0",
+    "ipython>=9.8.0",
 ]
 web-search = [
     "google-cloud-firestore>=2.21.0",
@@ -74,6 +67,14 @@ web-search = [
 [tool.uv]
 default-groups = ["dev", "docs"]
 
+[tool.uv.workspace]
+members = [
+  "aieng-eval-agents",
+]
+
+[tool.uv.sources]
+aieng-eval-agents = { workspace = true }
+
 [tool.ruff]
 include = ["*.py", "pyproject.toml", "*.ipynb"]
 line-length = 88
@@ -142,5 +143,5 @@ markers = [
 
 [tool.coverage]
     [tool.coverage.run]
-    source=["src/"]
-    omit=["tests/*", "*__init__.py"]
+    source=["aieng-eval-agents/aieng"]
+    omit=["aieng-eval-agents/tests/*", "*__init__.py", "scripts/*"]