updated solution to use detect_dataset

Author: klakhov

datumaro/components/dataset.py

@@ -22,6 +22,7 @@
 from datumaro.components.errors import (
     CategoriesRedefinedError,
     ConflictingCategoriesError,
+    DatasetNotFoundError,
     MediaTypeError,
     MultipleFormatsMatchError,
     NoMatchingFormatsError,
@@ -1158,6 +1159,15 @@ def import_from(
 
         if not format:
             format = cls.detect(path, env=env)
+        else:
+            not_found_error_instance = DatasetNotFoundError(path)
+
+            def not_found_error(format_name, reason, human_message):
+                not_found_error_instance.reason = human_message
+
+            detected = env.detect_dataset(path, rejection_callback=not_found_error, depth=3, format=format)
+            if not detected:
+                raise not_found_error_instance
 
         # TODO: remove importers, put this logic into extractors
         if format in env.importers:
@@ -1227,7 +1237,14 @@ def import_from(
         return dataset
 
     @staticmethod
-    def detect(path: str, *, env: Optional[Environment] = None, depth: int = 2) -> str:
+    def detect(
+        path: str,
+        *,
+        env: Optional[Environment] = None,
+        depth: int = 2,
+        rejection_callback:Optional[Callable] = None,
+        format: Optional[str] = None,
+    ) -> str:
         """
         Attempts to detect dataset format of a given directory.
 
@@ -1247,7 +1264,7 @@ def detect(path: str, *, env: Optional[Environment] = None, depth: int = 2) -> s
         if depth < 0:
             raise ValueError("Depth cannot be less than zero")
 
-        matches = env.detect_dataset(path, depth=depth)
+        matches = env.detect_dataset(path, depth=depth, rejection_callback=rejection_callback, format=format)
         if not matches:
             raise NoMatchingFormatsError()
         elif 1 < len(matches):

datumaro/components/environment.py

@@ -252,15 +252,22 @@ def detect_dataset(
         path: str,
         depth: int = 1,
         rejection_callback: Optional[Callable[[str, RejectionReason, str], None]] = None,
+        format: Optional[str] = None
     ) -> List[str]:
         ignore_dirs = {"__MSOSX", "__MACOSX"}
         matched_formats = set()
+        detectors = []
+        if not format:
+            detectors = (
+                (format_name, importer.detect)
+                for format_name, importer in self.importers.items.items()
+            )
+        elif self.is_format_known(format):
+            detectors = [(format, self.importers.get(format).detect)]
+
         for _ in range(depth + 1):
             detected_formats = detect_dataset_format(
-                (
-                    (format_name, importer.detect)
-                    for format_name, importer in self.importers.items.items()
-                ),
+                detectors,
                 path,
                 rejection_callback=rejection_callback,
             )

datumaro/components/errors.py

@@ -273,14 +273,10 @@ def __str__(self):
 @define(auto_exc=False)
 class DatasetNotFoundError(DatasetImportError):
     path = field()
-    ext = field(default="")
-    filename = field(default="")
+    reason = field(default='')
 
     def __str__(self):
-        file_ext_info = (
-            f", file '{self.filename}{self.ext}' was not found" if self.ext or self.filename else ""
-        )
-        return f"Failed to find dataset at '{self.path}' {file_ext_info}"
+        return f"Failed to find dataset at '{self.path}' {self.reason}"
 
 
 @define(auto_exc=False)

datumaro/components/extractor.py

@@ -415,17 +415,6 @@ def get(self, id, subset=None):
 
 
 class Importer(CliPlugin):
-    def __init__(self):
-        self.__not_found_error_data = {"ext": "", "filename": ""}
-
-    @property
-    def _not_found_error_data(self):
-        return self.__not_found_error_data
-
-    @_not_found_error_data.setter
-    def _not_found_error_data_setter(self, val):
-        self.__not_found_error_data = val
-
     @classmethod
     def detect(
         cls,
@@ -450,11 +439,7 @@ def __call__(self, path, **extra_params):
 
         found_sources = self.find_sources_with_params(osp.normpath(path), **extra_params)
         if not found_sources:
-            raise DatasetNotFoundError(
-                path,
-                self._not_found_error_data.get("ext", ""),
-                self._not_found_error_data.get("filename", ""),
-            )
+            raise DatasetNotFoundError(path)
 
         sources = []
         for desc in found_sources:
@@ -520,9 +505,6 @@ def _find_sources_recursive(
                 if sources:
                     break
 
-        if not sources:
-            cls._not_found_error_data = {"ext": ext, "filename": filename}
-
         return sources