Check dataset integrity

.github/hooks/.pre-commit-config.yaml

@@ -32,18 +32,27 @@ repos:
 
   - repo: local
     hooks:
-      - id: uvx-ruff-format
-        name: uvx ruff format
-        entry: bash -c 'uvx ruff format'
+      - id: uv-run-typecheker
+        name: uv run ty check
         language: system
+        entry: bash -c 'uv run ty check'
         types_or: [python, pyi]
         require_serial: true
 
   - repo: local
     hooks:
-      - id: uvx-ty-check
-        name: uvx ty check
-        entry: bash -c 'uvx ty check'
+      - id: uv-ruff-linter
+        name: uv run ruff check
         language: system
+        entry: bash -c 'uv run ruff check'
+        types_or: [python, pyi]
+        require_serial: true
+
+  - repo: local
+    hooks:
+      - id: uv-ruff-formatter
+        name: uv run ruff format
+        language: system
+        entry: bash -c 'uv run ruff format'
         types_or: [python, pyi]
         require_serial: true

.github/workflows/ci.yaml

@@ -65,6 +65,10 @@ jobs:
         run: |
           make test
 
+      - name: Run integration tests with pytest
+        run: |
+          make i-test
+
   test-arm:
     runs-on: ${{ matrix.os }}
     strategy:

.github/workflows/coverage.yaml

@@ -27,18 +27,18 @@ jobs:
         run: |
           make build
 
-      - name: Run tests with coverage
+      - name: Run unit tests with coverage
         run: |
-          uv run pytest --cov --cov-branch --cov-report=xml
+          uv run pytest -n auto --cov --cov-branch --cov-report=xml -m "not integration"
 
       - name: Upload coverage to Codecov
         uses: codecov/codecov-action@v5
         with:
           token: ${{ secrets.CODECOV_TOKEN }}
 
-      - name: Test with pytest
+      - name: Run unit tests for Codecov
         run: |
-          uv run pytest --cov --junitxml=junit.xml
+          uv run pytest -n auto --cov --junitxml=junit.xml -m "not integration"
 
       - name: Upload test results to Codecov
         uses: codecov/codecov-action@v5

.github/workflows/issues.yaml

@@ -6,8 +6,6 @@ on:
 
 jobs:
   validate-issue-labels:
-    environment:
-      name: management
     runs-on: ubuntu-latest
     permissions:
       issues: write

.github/workflows/pull_requests.yaml

@@ -6,8 +6,6 @@ on:
 
 jobs:
   label:
-    environment:
-      name: management
     runs-on: ubuntu-latest
     permissions:
       contents: read
@@ -20,8 +18,6 @@ jobs:
           sync-labels: true
 
   assignees:
-    environment:
-      name: management
     runs-on: ubuntu-latest
     permissions:
       contents: read

Makefile

@@ -1,5 +1,5 @@
-.PHONY: all build setup setup-tensorboard clean destroy \
-		test stest run \
+.PHONY: all release build setup setup-tensorboard clean destroy \
+		test stest i-test si-test run \
 		check format typecheck lint lint-fix lint-rule lint-rule-fix \
 		docs docs-build docs-serve \
 		loc help
@@ -14,6 +14,8 @@ DOCS_ADDR=127.0.0.1:8000
 
 all: clean setup check test
 
+release: clean setup check test i-test
+
 build: clean setup
 
 setup:
@@ -56,12 +58,20 @@ lint-rule-fix:
 	$(UV) run $(LINTER) check --select $(R) --fix
 
 test:
-	@echo '=== Tests ==='
-	$(UV) run $(PYTEST) --cov=$(PROJECT_NAME) --cov-report=term-missing
+	@echo '=== Running unit tests in parallel ==='
+	$(UV) run $(PYTEST) -n auto --cov=$(PROJECT_NAME) --cov-report=term-missing -m "not integration"
 
 stest:
-	@echo '=== Test for $(T) ==='
-	$(UV) run $(PYTEST) -s $(PROJECT_NAME)/tests/$(T)
+	@echo '=== Running single unit test for $(T) ==='
+	$(UV) run $(PYTEST) -n auto -s $(PROJECT_NAME)/tests/$(T)
+
+i-test:
+	@echo '=== Running integration tests in parallel ==='
+	$(UV) run $(PYTEST) -n auto -m "integration"
+
+si-test:
+	@echo '=== Running single integration test for $(T) ==='
+	$(UV) run $(PYTEST) -n auto -s $(PROJECT_NAME)/integration_tests/$(T) -m "integration"
 
 # If the first argument is run...
 ifeq ($(firstword $(MAKECMDGOALS)),run)
@@ -104,6 +114,7 @@ help:
 	@echo "Usage: make [target]"
 	@echo "Targets:"
 	@echo "  all                     - Clean, setup, lint, typecheck, test"
+	@echo "  release                 - Clean, setup, lint, typecheck, test, i-test"
 	@echo "  build                   - Clean and setup"
 	@echo "  setup                   - Install dependencies"
 	@echo "  setup-tensorboard       - Install optional TensorBoard dependency"
@@ -115,8 +126,10 @@ help:
 	@echo "  lint-fix                - Run linting and fix issues"
 	@echo "  lint-rule R=<rule>      - Run linting for a specific rule (e.g., R=E501)"
 	@echo "  lint-rule-fix R=<rule>  - Run linting for a specific rule and fix issues"
-	@echo "  test                    - Run all tests"
-	@echo "  stest T=<test_name>     - Run a single test"
+	@echo "  test                    - Run all unit tests in parallel"
+	@echo "  stest T=<test_name>     - Run a single unit test"
+	@echo "  i-test                  - Run integration tests in parallel"
+	@echo "  si-test T=<test_name>   - Run a single integration test"
 	@echo "  run <file.py>           - Run a single file"
 	@echo "  docs                    - Build and serve documentation"
 	@echo "  docs-build              - Build documentation without serving"

docs/agents/references/testing.md

@@ -10,6 +10,8 @@
 - Prefer Makefile targets:
   - `make test`
   - `make stest T=<path-within-hyperbench/tests>`
+  - `make i-test`
+  - `make si-test T=<path-within-hyperbench/integration_tests>`
 - Use `uv run pytest` only for targeted one-off runs.
 - Run `make test` to see coverage reports and identify untested lines.
 

docs/development/makefile.md

@@ -7,12 +7,13 @@ You can always run `make help` to see the latest list.
 > Notes
 >
 > - These are the explicit CLI commands that each `make` target runs.
-> - `make all`, `make build`, `make check`, `make docs` are composites (they run multiple steps).
+> - `make all`, `make release`, `make build`, `make check`, `make docs` are composites (they run multiple steps).
 
 | Target | Description |
 |---|---|
 | `make help` | Print available targets |
 | `make all` | Clean, setup, check, test |
+| `make release` | Clean, setup, check, test, i-test |
 | `make build` | Clean and setup |
 | `make setup` | Install dependencies (via `uv`) and install HyperBench in editable mode |
 | `make setup-tensorboard` | Install optional TensorBoard extra |
@@ -26,6 +27,8 @@ You can always run `make help` to see the latest list.
 | `make docstring-check` | Check docstring formatting |
 | `make test` | Run all tests (with coverage) |
 | `make stest T=<test_path>` | Run a single test file/folder under `hyperbench/tests/` |
+| `make i-test` | Run all integration tests |
+| `make si-test T=<test_path>` | Run a single integration test file/folder under `hyperbench/integration_tests/` |
 | `make run <file.py>` | Run a single Python file (for example: `make run examples/gcn.py`) |
 | `make docs` | Build and serve documentation |
 | `make docs-build` | Build documentation without serving |
@@ -197,6 +200,7 @@ You can always run `make help` to see the latest list.
 Utility targets that run multiple steps in sequence.
 
 - `make all`: Clean, setup, check, test
+- `make release`: Clean, setup, check, test, i-test
 - `make build`: Clean and setup
 - `make check`: Run lint + format + typecheck
 - `make docs`: Build and serve documentation

examples/hgnn.py

@@ -48,9 +48,6 @@
         print(f"Val dataset:\n {val_dataset.hdata}\n")
         print(f"Test dataset:\n {test_dataset.hdata}\n")
 
-    # Save train hyperedge index before adding negatives (for CommonNeighbors)
-    train_hyperedge_index = train_dataset.hdata.hyperedge_index
-
     # Add negative samples to all splits
     for name, ds in [("Train", train_dataset), ("Val", val_dataset), ("Test", test_dataset)]:
         num_negative_samples = (

examples/hgnnp.py

@@ -48,8 +48,6 @@
         print(f"Val dataset:\n {val_dataset.hdata}\n")
         print(f"Test dataset:\n {test_dataset.hdata}\n")
 
-    train_hyperedge_index = train_dataset.hdata.hyperedge_index
-
     for name, ds in [("Train", train_dataset), ("Val", val_dataset), ("Test", test_dataset)]:
         num_negative_samples = (
             ds.hdata.num_hyperedges

hyperbench/__init__.py

@@ -7,3 +7,8 @@
 
 warnings.filterwarnings("ignore", message="Sparse CSR tensor support is in beta state")
 warnings.filterwarnings("ignore", message="Sparse invariant checks are implicitly disabled")
+
+warnings.filterwarnings(
+    "ignore",
+    message="ignore:Failing to pass a value to the 'type_params' parameter of 'typing._eval_type' is deprecated.*",
+)

hyperbench/data/__init__.py

@@ -26,6 +26,8 @@
     ThreadsMathsxDataset,
     TwitterDataset,
     VegasBarsReviewsDataset,
+    list_datasets,
+    get_dataset_by_name,
 )
 
 from .loader import DataLoader
@@ -116,4 +118,6 @@
     "VilLainEnricher",
     "VilLainHyperedgeAttrsEnricher",
     "create_sampler_from_strategy",
+    "get_dataset_by_name",
+    "list_datasets",
 ]

hyperbench/data/hif.py

@@ -373,7 +373,7 @@ def load_by_name(
 
     @classmethod
     def __extract_hif(cls, json_file: str) -> HIFHypergraph:
-        with open(json_file) as f:
+        with open(json_file, encoding="utf-8") as f:
             hiftext = json.load(f)
         if not validate_hif_json(json_file):
             raise ValueError(f"Dataset from file '{json_file}' is not HIF-compliant.")