continue test fixes and respect perform_io_checks flag

bioimageio/spec/_internal/types/_file_source.py

@@ -17,6 +17,8 @@
 from pydantic_core import core_schema
 from typing_extensions import Annotated
 
+from bioimageio.spec._internal.validation_context import get_internal_validation_context
+
 
 class RelativePath:
     path: PurePosixPath
@@ -95,9 +97,9 @@ def _validate(cls, value: Union[pathlib.Path, str], info: ValidationInfo):
         if ret.path.is_absolute():
             raise ValueError(f"{value} is an absolute path.")
 
-        root = (info.context or {}).get("root")
-        if root is not None:
-            ret._check_exists(root)
+        context = get_internal_validation_context(info.context)
+        if context["perform_io_checks"]:
+            ret._check_exists(context["root"])
 
         return ret
 

tests/conftest.py

@@ -8,6 +8,8 @@
 
 EXAMPLE_SPECS = Path(__file__).parent / "../example_specs"
 
+UNET2D_ROOT = EXAMPLE_SPECS / "models/unet2d_nuclei_broad"
+
 
 @pytest.fixture(scope="session")
 def stardist04_data():
@@ -17,12 +19,13 @@ def stardist04_data():
 
 @pytest.fixture(scope="session")
 def unet2d_root() -> Path:
-    return EXAMPLE_SPECS / "models/unet2d_nuclei_broad"
+    """deprecated, use constant UNET2D_ROOT instead"""
+    return UNET2D_ROOT
 
 
 @pytest.fixture(scope="session")
-def unet2d_data(unet2d_root: Path):
-    with (unet2d_root / "rdf.yaml").open() as f:
+def unet2d_data():
+    with (UNET2D_ROOT / "rdf.yaml").open() as f:
         data = yaml.load(f)
 
     return MappingProxyType(data)

tests/test_description.py

@@ -1,6 +1,6 @@
 from pydantic import HttpUrl
 
-from bioimageio.spec._description import build_description, validate_format
+from bioimageio.spec._description import validate_format
 from bioimageio.spec._internal.types import BioimageioYamlContent
 from bioimageio.spec._internal.validation_context import ValidationContext
 
@@ -10,10 +10,10 @@ def test_forward_compatibility(unet2d_data: BioimageioYamlContent):
     v_future = "9999.0.0"
     data["format_version"] = v_future  # assume it is valid in a future format version
 
-    summary = build_description(
+    summary = validate_format(
         data, context=ValidationContext(root=HttpUrl("https://example.com/"), perform_io_checks=False)
-    ).validation_summaries[0]
-    assert summary.status == "passed", summary
+    )
+    assert summary.status == "passed", summary.errors
 
     # expect warning about treating future format version as latest
     ws = summary.warnings

tests/test_generic/test_v0_3.py

@@ -1,12 +1,16 @@
 from __future__ import annotations
 
 from pathlib import Path
-from typing import Any, Dict
+from typing import Any, Dict, Union
 
 import pytest
 
 from bioimageio.spec._internal.constants import WARNING
-from bioimageio.spec._internal.validation_context import ValidationContext, get_internal_validation_context
+from bioimageio.spec._internal.validation_context import (
+    InternalValidationContext,
+    ValidationContext,
+    get_internal_validation_context,
+)
 from bioimageio.spec.generic.v0_3 import Generic
 from tests.utils import check_node
 
@@ -42,7 +46,7 @@
     ],
 )
 def test_generic_valid(kwargs: Dict[str, Any]):
-    check_node(Generic, kwargs)
+    check_node(Generic, kwargs, context=ValidationContext(perform_io_checks=False))
 
 
 @pytest.mark.parametrize(
@@ -59,12 +63,12 @@ def test_generic_valid(kwargs: Dict[str, Any]):
                 license="BSD-2-Clause-FreeBSD",
                 cite=[dict(text="lala", url=EXAMPLE_DOT_COM)],
             ),
-            get_internal_validation_context(warning_level=WARNING),
+            get_internal_validation_context(warning_level=WARNING, perform_io_checks=False),
             id="deprecated license",
         ),
         (
             dict(format_version=Generic.implemented_format_version, version="0.1.0", type="my_type", name="their name"),
-            {},
+            ValidationContext(perform_io_checks=False),
         ),
         (
             dict(
@@ -74,9 +78,9 @@ def test_generic_valid(kwargs: Dict[str, Any]):
                 name="its name",
                 attachments={"files": [Path(__file__), "missing"], "something": 42},
             ),
-            {},
+            ValidationContext(perform_io_checks=False),
         ),
     ],
 )
-def test_generic_invalid(kwargs: Dict[str, Any], context: ValidationContext):
+def test_generic_invalid(kwargs: Dict[str, Any], context: Union[ValidationContext, InternalValidationContext]):
     check_node(Generic, kwargs, context=context, is_invalid=True)

tests/test_model/test_v0_4.py

@@ -1,4 +1,5 @@
 from datetime import datetime
+from pathlib import Path
 from typing import Any, Dict, Union
 
 import pytest
@@ -17,14 +18,14 @@
     OutputTensor,
     Postprocessing,
     Preprocessing,
-    RelativeFilePath,
     ScaleLinearKwargs,
     ScaleMeanVariance,
     ScaleRange,
     ScaleRangeKwargs,
     TensorName,
     Weights,
 )
+from tests.conftest import UNET2D_ROOT
 from tests.utils import check_node, check_type, unset
 
 
@@ -47,16 +48,12 @@ def test_linked_model_invalid(kwargs: Dict[str, Any]):
     "kwargs,expected",
     [
         (
-            dict(opset_version=8, source="https://example.com", sha256="s" * 64),
-            dict(opset_version=8, source="https://example.com/", sha256="s" * 64),
+            dict(source=UNET2D_ROOT / "weights.onnx", sha256="s" * 64),
+            dict(source=UNET2D_ROOT / "weights.onnx", sha256="s" * 64),
         ),
+        (dict(opset_version=5, source=UNET2D_ROOT / "weights.onnx", sha256="s" * 64), ValidationError),
         (
-            dict(source="https://example.com", sha256="s" * 64),
-            dict(source="https://example.com/", sha256="s" * 64),
-        ),
-        (dict(opset_version=5, source="https://example.com", sha256="s" * 64), ValidationError),
-        (
-            dict(source="https://example.com", sha256="s"),
+            dict(source=UNET2D_ROOT / "weights.onnx", sha256="s"),
             ValidationError,
         ),
     ],
@@ -237,9 +234,9 @@ def test_output_tensor(kwargs: Dict[str, Any]):
 
 
 @pytest.fixture
-def model_data():
+def model_data(unet2d_root: Path):
     return Model(
-        documentation=RelativeFilePath("docs.md"),
+        documentation=unet2d_root / "README.md",
         license="MIT",
         git_repo="https://github.com/bioimage-io/python-bioimage-io",
         format_version="0.4.9",
@@ -274,9 +271,9 @@ def model_data():
         ],
         name="Model",
         tags=[],
-        weights=Weights(onnx=OnnxWeights(source=RelativeFilePath("weights.onnx"))),
-        test_inputs=[RelativeFilePath("test_ipt.npy")],
-        test_outputs=[RelativeFilePath("test_out.npy")],
+        weights=Weights(onnx=OnnxWeights(source=unet2d_root / "weights.onnx")),
+        test_inputs=[unet2d_root / "test_ipt.npy"],
+        test_outputs=[unet2d_root / "test_out.npy"],
         type="model",
     ).model_dump()
 
@@ -304,17 +301,13 @@ def model_data():
 )
 def test_model(model_data: Dict[str, Any], update: Dict[str, Any]):
     model_data.update(update)
-    summary = validate_format(
-        model_data, context=ValidationContext(root=HttpUrl("https://example.com/"), perform_io_checks=False)
-    )
+    summary = validate_format(model_data, context=ValidationContext(perform_io_checks=False))
     assert summary.status == "passed", summary.format()
 
 
 def test_warn_long_name(model_data: Dict[str, Any]):
     model_data["name"] = "veeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeeery loooooooooooooooong name"
-    summary = validate_format(
-        model_data, context=ValidationContext(root=HttpUrl("https://example.com/"), perform_io_checks=False)
-    )
+    summary = validate_format(model_data, context=ValidationContext(perform_io_checks=False))
     assert summary.status == "passed", summary.format()
     assert summary.warnings[0].loc == ("name",), summary.format()
     assert summary.warnings[0].msg in "Name longer than 64 characters."

tests/test_model/test_v0_5.py

@@ -1,8 +1,8 @@
 from datetime import datetime
+from pathlib import Path
 from typing import Any, Dict, Union
 
 import pytest
-from pydantic import HttpUrl
 
 from bioimageio.spec._description import validate_format
 from bioimageio.spec._internal.validation_context import ValidationContext
@@ -24,6 +24,7 @@
     TensorId,
     Weights,
 )
+from tests.conftest import UNET2D_ROOT
 from tests.utils import check_node, check_type
 
 
@@ -32,20 +33,20 @@
     [
         dict(
             id="t1",
-            test_tensor="https://example.com/test.npy",
+            test_tensor=UNET2D_ROOT / "test_input.npy",
             data={"values": ["cat", "dog", "parrot"]},
         ),
         dict(
             id="t2",
-            test_tensor="https://example.com/test.npy",
+            test_tensor=UNET2D_ROOT / "test_input.npy",
             data=[
                 {"values": ["cat", "dog", "parrot"]},
                 {"values": ["mouse", "zebra", "elephant"]},
             ],
         ),
         dict(
             id="t3",
-            test_tensor="https://example.com/test.npy",
+            test_tensor=UNET2D_ROOT / "test_input.npy",
             data=[
                 {"values": [1, 2, 3]},
                 {"type": "uint8"},
@@ -54,7 +55,7 @@
         pytest.param(
             dict(
                 id="t4",
-                test_tensor="https://example.com/test.npy",
+                test_tensor=UNET2D_ROOT / "test_input.npy",
                 data=[
                     {"values": ["mouse", "zebra", "elephant"]},
                     {"type": "uint8"},
@@ -74,7 +75,7 @@ def test_tensor_base(kwargs: Dict[str, Any]):
         pytest.param(
             dict(
                 id="t5",
-                test_tensor="https://example.com/test.npy",
+                test_tensor=UNET2D_ROOT / "test_input.npy",
                 data=[
                     {"values": ["cat", "dog", "parrot"]},
                     {"values": [1.1, 2.2, 3.3]},
@@ -85,7 +86,7 @@ def test_tensor_base(kwargs: Dict[str, Any]):
         pytest.param(
             dict(
                 id="t7",
-                test_tensor="https://example.com/test.npy",
+                test_tensor=UNET2D_ROOT / "test.npy",
                 data=[
                     {"values": ["mouse", "zebra", "elephant"]},
                     {"type": "int8"},
@@ -117,7 +118,7 @@ def test_tensor_base_invalid(kwargs: Dict[str, Any]):
                     "kwargs": {"max_percentile": 99, "min_percentile": 5, "mode": "per_sample", "axes": ("x", "y")},
                 }
             ],
-            "test_tensor": "https://example.com/test.npy",
+            "test_tensor": UNET2D_ROOT / "test_input.npy",
         },
     ],
 )
@@ -131,7 +132,7 @@ def test_input_tensor(kwargs: Dict[str, Any]):
         pytest.param(
             dict(
                 id="input_2",
-                test_tensor="https://example.com/test.npy",
+                test_tensor=UNET2D_ROOT / "test.npy",
                 data=[
                     {"values": ["cat", "dog", "parrot"]},
                     {"values": ["mouse", "zebra", "elephant"]},
@@ -169,9 +170,9 @@ def test_input_axis(kwargs: Union[Dict[str, Any], SpaceInputAxis]):
 
 
 @pytest.fixture
-def model_data():
+def model_data(unet2d_root: Path):
     data = Model(
-        documentation=HttpUrl("https://example.com/docs.md"),
+        documentation=unet2d_root / "README.md",
         license="MIT",
         git_repo="https://github.com/bioimage-io/python-bioimage-io",
         format_version="0.5.0",
@@ -196,7 +197,7 @@ def model_data():
                     SpaceInputAxis(id=AxisId("y"), size=20),
                     ChannelAxis(size=3),
                 ],
-                test_tensor=HttpUrl("https://example.com/test_ipt.npy"),
+                test_tensor=unet2d_root / "test_ipt.npy",
             ),
         ],
         outputs=[
@@ -208,17 +209,14 @@ def model_data():
                     SpaceOutputAxis(id=AxisId("y"), size=25),
                     ChannelAxis(size=6),
                 ],
-                test_tensor=HttpUrl("https://example.com/test_out.npy"),
+                test_tensor=unet2d_root / "test_out.npy",
             ),
         ],
         name="Model",
         tags=[],
-        weights=Weights(onnx=OnnxWeights(source=HttpUrl("https://example.com/weights.onnx"), opset_version=15)),
+        weights=Weights(onnx=OnnxWeights(source=unet2d_root / "weights.onnx", opset_version=15)),
         type="model",
     ).model_dump()
-    # make data more editable
-    data["outputs"] = list(data["outputs"])
-    data["outputs"][0]["axes"] = list(data["outputs"][0]["axes"])
     return data
 
 
@@ -227,19 +225,19 @@ def model_data():
     [
         pytest.param(dict(name="µ-unicode-model/name!"), id="unicode name"),
         dict(run_mode={"name": "special_run_mode", "kwargs": dict(marathon=True)}),
-        dict(weights={"torchscript": {"source": "https://example.com/weights", "pytorch_version": 1.15}}),
-        dict(weights={"keras_hdf5": {"source": "https://example.com/weights", "tensorflow_version": 1.10}}),
-        dict(weights={"tensorflow_js": {"source": "https://example.com/weights", "tensorflow_version": 1.10}}),
+        dict(weights={"torchscript": {"source": UNET2D_ROOT / "weights.onnx", "pytorch_version": 1.15}}),
+        dict(weights={"keras_hdf5": {"source": UNET2D_ROOT / "weights.onnx", "tensorflow_version": 1.10}}),
+        dict(weights={"tensorflow_js": {"source": UNET2D_ROOT / "weights.onnx", "tensorflow_version": 1.10}}),
         dict(
             weights={
-                "tensorflow_saved_model_bundle": {"source": "https://example.com/weights", "tensorflow_version": 1.10}
+                "tensorflow_saved_model_bundle": {"source": UNET2D_ROOT / "weights.onnx", "tensorflow_version": 1.10}
             }
         ),
-        dict(weights={"onnx": {"source": "https://example.com/weights", "opset_version": 15}}),
+        dict(weights={"onnx": {"source": UNET2D_ROOT / "weights.onnx", "opset_version": 15}}),
         dict(
             weights={
                 "pytorch_state_dict": {
-                    "source": "https://example.com/weights",
+                    "source": UNET2D_ROOT / "weights.onnx",
                     "pytorch_version": "1.15",
                     "architecture": {
                         "callable": "https://example.com/file.py:Model",