Merge branch 'develop' into main

2 files changed, 110 insertions, 0 deletions

diff --git a/bsie/lib/__init__.py b/bsie/lib/__init__.py
new file mode 100644
index 0000000..578c2c4
--- /dev/null
+++ b/bsie/lib/__init__.py
@@ -0,0 +1,18 @@
+"""
+
+Part of the bsie module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+import typing
+
+# inner-module imports
+from .bsie import BSIE
+
+# exports
+__all__: typing.Sequence[str] = (
+    'BSIE',
+    )
+
+## EOF ##
diff --git a/bsie/lib/bsie.py b/bsie/lib/bsie.py
new file mode 100644
index 0000000..e087fa9
--- /dev/null
+++ b/bsie/lib/bsie.py
@@ -0,0 +1,92 @@
+"""
+
+Part of the bsie module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+import typing
+
+# bsie imports
+from bsie.tools import Pipeline
+from bsie.utils import bsfs, node, ns
+
+# exports
+__all__: typing.Sequence[str] = (
+    'BSIE',
+    )
+
+
+## code ##
+
+class BSIE():
+    """Extract triples from files.
+
+    Controls which predicates to extract (*collect*) and
+    which to not extract (*discard*). Note that this only affects
+    principal predicates not auxililary predicates like, e.g., tag labels.
+
+    """
+
+    # pipeline
+    _pipeline: Pipeline
+
+    # predicates to extract.
+    _principals: typing.Set[bsfs.URI]
+
+    # local schema.
+    _schema: bsfs.schema.Schema
+
+    def __init__(
+            self,
+            # pipeline builder.
+            pipeline: Pipeline,
+            # principals to extract at most. None implies all available w.r.t. extractors.
+            collect: typing.Optional[typing.Iterable[bsfs.URI]] = None,
+            # principals to discard.
+            discard: typing.Optional[typing.Iterable[bsfs.URI]] = None,
+            ):
+        # store pipeline
+        self._pipeline = pipeline
+        # start off with available principals
+        self._principals = {pred.uri for pred in self._pipeline.principals}
+        # limit principals to specified ones by argument.
+        if collect is not None:
+            collect = set(collect)
+            if len(collect) > 0:
+                self._principals &= collect
+        # discard principals.
+        if discard is not None:
+            self._principals -= set(discard)
+        # discard ns.bsfs.Predicate
+        self._principals.discard(ns.bsfs.Predicate)
+        # compile a schema that only contains the requested principals (and auxiliary predicates)
+        self._schema = self._pipeline.subschema(
+            self._pipeline.schema.predicate(pred) for pred in self._principals)
+
+    @property
+    def schema(self) -> bsfs.schema.Schema:
+        """Return the BSIE schema."""
+        return self._schema
+
+    @property
+    def principals(self) -> typing.Iterator[bsfs.URI]:
+        """Return an iterator to the principal predicates."""
+        return iter(self._principals)
+
+    def from_file(
+            self,
+            path: bsfs.URI,
+            principals: typing.Optional[typing.Iterable[bsfs.URI]] = None,
+            ) -> typing.Iterator[typing.Tuple[node.Node, bsfs.URI, typing.Any]]:
+        """Produce triples for a given *path*. Limit to *principals* if given."""
+        # get requested principals.
+        principals = set(principals) if principals is not None else self._principals
+        # filter through requested principals.
+        principals &= self._principals
+        # predicate lookup
+        principals = {self.schema.predicate(pred) for pred in principals}
+        # invoke pipeline
+        yield from self._pipeline(path, principals)
+
+## EOF ##