fix(yolo): correct data.yaml paths for all YOLO formats (#240)

roboflow/core/version.py

@@ -35,7 +35,7 @@
 from roboflow.util.annotations import amend_data_yaml
 from roboflow.util.general import extract_zip, write_line
 from roboflow.util.model_processor import package_custom_weights_interactive, validate_model_type_for_project
-from roboflow.util.versions import get_model_format, get_wrong_dependencies_versions
+from roboflow.util.versions import get_model_format
 
 if TYPE_CHECKING:
     import numpy as np
@@ -644,33 +644,62 @@ def __get_format_identifier(self, format):
         return friendly_formats.get(format, format)
 
     def __reformat_yaml(self, location: str, format: str):
-        """
-        Certain formats seem to require reformatting the downloaded YAML.
+        """Reformat the downloaded ``data.yaml`` for specific YOLO formats.
+
+        The ZIP downloaded from Roboflow contains a ``data.yaml`` whose
+        ``train`` / ``val`` / ``test`` paths are prefixed with ``../`` which
+        is incorrect — the split directories are siblings of ``data.yaml``,
+        not in the parent directory.  This method rewrites those paths so
+        that training works out-of-the-box.
+
+        * Modern formats (``yolov8``, ``yolov9``) use relative paths
+          (``train/images``, ``valid/images``, ``test/images``) as expected
+          by current ultralytics releases.
+        * Legacy formats (``yolov5pytorch``, ``yolov7pytorch``,
+          ``mt-yolov6``) use absolute paths constructed from *location*.
+
+        The ``test`` key is only written when the version actually has a
+        test split (``self.splits["test"] > 0``).
 
         Args:
-            location (str): filepath of the data directory that contains the yaml file
-            format (str): the format identifier string
-        """  # noqa: E501 // docs
+            location: Filepath of the data directory that contains the yaml file.
+            format: The format identifier string.
+        """
         data_path = os.path.join(location, "data.yaml")
 
+        def _strip_rel_prefix(path: str) -> str:
+            """Remove all leading ``../`` and ``./`` prefixes from *path*."""
+            while True:
+                if path.startswith("../"):
+                    path = path[3:]
+                elif path.startswith("./"):
+                    path = path[2:]
+                else:
+                    return path
+
         def data_yaml_callback(content: dict) -> dict:
-            if format == "mt-yolov6":
-                content["train"] = location + content["train"].lstrip(".")
-                content["val"] = location + content["val"].lstrip(".")
-                content["test"] = location + content["test"].lstrip(".")
-            if format in ["yolov5pytorch", "yolov7pytorch"]:
-                content["train"] = location + content["train"].lstrip("..")
-                content["val"] = location + content["val"].lstrip("..")
-            try:
-                # get_wrong_dependencies_versions raises exception if ultralytics is not installed at all  # noqa: E501 // docs
-                if format == "yolov8" and not get_wrong_dependencies_versions(
-                    dependencies_versions=[("ultralytics", "==", "8.0.196")]
-                ):
-                    content["train"] = "train/images"
-                    content["val"] = "valid/images"
+            has_test_split = self.splits.get("test", 0) > 0
+
+            # Modern ultralytics formats — relative paths
+            if format in ("yolov8", "yolov9"):
+                content["train"] = "train/images"
+                content["val"] = "valid/images"
+                if has_test_split:
                     content["test"] = "test/images"
-            except ModuleNotFoundError:
-                pass
+                elif "test" in content:
+                    del content["test"]
+                return content
+
+            # Legacy formats — absolute paths
+            if format in ("yolov5pytorch", "yolov7pytorch", "mt-yolov6"):
+                for key in ("train", "val", "test"):
+                    if key in content:
+                        if key == "test" and not has_test_split:
+                            del content["test"]
+                        else:
+                            content[key] = os.path.join(location, _strip_rel_prefix(content[key]))
+                return content
+
             return content
 
         if format in ["yolov5pytorch", "mt-yolov6", "yolov7pytorch", "yolov8", "yolov9"]:

tests/test_version.py

@@ -1,8 +1,10 @@
 import os
+import tempfile
 import unittest
 from unittest.mock import patch
 
 import responses
+import yaml
 
 from roboflow.adapters import rfapi
 from roboflow.config import (
@@ -266,3 +268,199 @@ def test_detection_project_rejects_sem_model(self):
     def test_classification_project_rejects_detection(self):
         with self.assertRaises(ValueError):
             self._version(TYPE_CLASSICATION)._validate_against_project_type("yolov11")
+
+
+# ---------------------------------------------------------------------------
+# __reformat_yaml — fixing issue #240 (Incorrect Data Path in YOLOv8 Dataset)
+# ---------------------------------------------------------------------------
+
+_INITIAL_YAML = {
+    "train": "../train/images",
+    "val": "../valid/images",
+    "test": "../test/images",
+    "nc": 3,
+    "names": ["cat", "dog", "bird"],
+    "roboflow": {"license": "MIT", "project": "test-project", "version": 1},
+}
+
+
+def _write_data_yaml(directory: str, content: dict) -> str:
+    """Write a ``data.yaml`` file in *directory* and return its path."""
+    path = os.path.join(directory, "data.yaml")
+    with open(path, "w") as fh:
+        yaml.dump(content, fh)
+    return path
+
+
+def _read_data_yaml(directory: str) -> dict:
+    """Read the ``data.yaml`` file from *directory*."""
+    with open(os.path.join(directory, "data.yaml")) as fh:
+        return yaml.safe_load(fh)
+
+
+class TestReformatYaml(unittest.TestCase):
+    """Tests for ``Version.__reformat_yaml`` (issue #240)."""
+
+    # ------------------------------------------------------------------
+    # helpers
+    # ------------------------------------------------------------------
+
+    def _version_with_test(self):
+        return get_version(splits={"train": 210, "valid": 20, "test": 10})
+
+    def _version_without_test(self):
+        return get_version(splits={"train": 210, "valid": 20, "test": 0})
+
+    def _reformat(self, version, location, fmt):
+        """Invoke the name-mangled private method."""
+        version._Version__reformat_yaml(location, fmt)
+
+    # ------------------------------------------------------------------
+    # yolov8 — relative paths
+    # ------------------------------------------------------------------
+
+    def test_yolov8_paths_are_relative(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov8")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertEqual(content["test"], "test/images")
+
+    def test_yolov8_without_test_split_removes_test_key(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_without_test(), tmp, "yolov8")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertNotIn("test", content)
+
+    def test_yolov8_preserves_other_fields(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov8")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["nc"], 3)
+            self.assertEqual(content["names"], ["cat", "dog", "bird"])
+            self.assertEqual(content["roboflow"]["license"], "MIT")
+
+    # ------------------------------------------------------------------
+    # yolov9 — relative paths (was completely missing before fix)
+    # ------------------------------------------------------------------
+
+    def test_yolov9_paths_are_relative(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov9")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertEqual(content["test"], "test/images")
+
+    def test_yolov9_without_test_split_removes_test_key(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_without_test(), tmp, "yolov9")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertNotIn("test", content)
+
+    # ------------------------------------------------------------------
+    # Legacy formats — absolute paths
+    # ------------------------------------------------------------------
+
+    def test_yolov5pytorch_uses_absolute_paths(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov5pytorch")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], os.path.join(tmp, "train/images"))
+            self.assertEqual(content["val"], os.path.join(tmp, "valid/images"))
+            self.assertEqual(content["test"], os.path.join(tmp, "test/images"))
+
+    def test_yolov5pytorch_without_test_split_removes_test_key(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_without_test(), tmp, "yolov5pytorch")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], os.path.join(tmp, "train/images"))
+            self.assertEqual(content["val"], os.path.join(tmp, "valid/images"))
+            self.assertNotIn("test", content)
+
+    def test_yolov7pytorch_uses_absolute_paths(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov7pytorch")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], os.path.join(tmp, "train/images"))
+            self.assertEqual(content["val"], os.path.join(tmp, "valid/images"))
+            self.assertEqual(content["test"], os.path.join(tmp, "test/images"))
+
+    def test_mt_yolov6_uses_absolute_paths(self):
+        initial = dict(_INITIAL_YAML)
+        initial["train"] = "./train/images"
+        initial["val"] = "./valid/images"
+        initial["test"] = "./test/images"
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, initial)
+            self._reformat(self._version_with_test(), tmp, "mt-yolov6")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], os.path.join(tmp, "train/images"))
+            self.assertEqual(content["val"], os.path.join(tmp, "valid/images"))
+            self.assertEqual(content["test"], os.path.join(tmp, "test/images"))
+
+    # ------------------------------------------------------------------
+    # Edge cases
+    # ------------------------------------------------------------------
+
+    def test_strips_double_parent_prefix(self):
+        """``../../train/images`` should become ``<location>/train/images``."""
+        initial = dict(_INITIAL_YAML)
+        initial["train"] = "../../train/images"
+        initial["val"] = "../../valid/images"
+        initial["test"] = "../../test/images"
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, initial)
+            self._reformat(self._version_with_test(), tmp, "yolov5pytorch")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], os.path.join(tmp, "train/images"))
+            self.assertEqual(content["val"], os.path.join(tmp, "valid/images"))
+
+    def test_non_yolo_format_not_modified(self):
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "coco")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "../train/images")
+            self.assertEqual(content["val"], "../valid/images")
+
+    # ------------------------------------------------------------------
+    # Independence from ultralytics (the root cause of issue #240)
+    # ------------------------------------------------------------------
+
+    def test_yolov8_works_without_ultralytics(self):
+        """Paths must be fixed even when ultralytics is not installed."""
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            with patch(
+                "roboflow.util.versions.import_module",
+                side_effect=ModuleNotFoundError,
+            ):
+                self._reformat(self._version_with_test(), tmp, "yolov8")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertEqual(content["test"], "test/images")
+
+    def test_yolov8_works_with_any_ultralytics_version(self):
+        """Paths must be fixed regardless of the installed ultralytics version."""
+        with tempfile.TemporaryDirectory() as tmp:
+            _write_data_yaml(tmp, dict(_INITIAL_YAML))
+            self._reformat(self._version_with_test(), tmp, "yolov8")
+            content = _read_data_yaml(tmp)
+            self.assertEqual(content["train"], "train/images")
+            self.assertEqual(content["val"], "valid/images")
+            self.assertEqual(content["test"], "test/images")