Merge branch 'mb/arch' into develop

1 files changed, 50 insertions, 0 deletions

diff --git a/bsie/base/extractor.py b/bsie/base/extractor.py
new file mode 100644
index 0000000..d5b0922
--- /dev/null
+++ b/bsie/base/extractor.py
@@ -0,0 +1,50 @@
+"""The Extractor classes transform content into triples.
+
+Part of the bsie module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+import abc
+import collections
+import typing
+
+# inner-module imports
+from . import reader
+from bsie.utils import node
+from bsie.utils.bsfs import URI, typename
+
+# exports
+__all__: typing.Sequence[str] = (
+    'Extractor',
+    )
+
+
+## code ##
+
+class Extractor(abc.ABC, collections.abc.Iterable, collections.abc.Callable):
+    """Produce (node, predicate, value)-triples from some content."""
+
+    # what type of content is expected (i.e. reader subclass).
+    CONTENT_READER: typing.Optional[typing.Type[reader.Reader]] = None
+
+    def __str__(self) -> str:
+        return typename(self)
+
+    def __repr__(self) -> str:
+        return f'{typename(self)}()'
+
+    @abc.abstractmethod
+    def schema(self) -> str:
+        """Return the schema (predicates and nodes) produced by this Extractor."""
+
+    @abc.abstractmethod
+    def extract(
+            self,
+            subject: node.Node,
+            content: typing.Any,
+            predicates: typing.Iterable[URI],
+            ) -> typing.Iterator[typing.Tuple[node.Node, URI, typing.Any]]:
+        """Return (node, predicate, value) triples."""
+
+## EOF ##