preview extractor

1 files changed, 99 insertions, 0 deletions

diff --git a/bsie/extractor/preview.py b/bsie/extractor/preview.py
new file mode 100644
index 0000000..1531d62
--- /dev/null
+++ b/bsie/extractor/preview.py
@@ -0,0 +1,99 @@
+"""
+
+Part of the bsie module.
+A copy of the license is provided with the project.
+Author: Matthias Baumgartner, 2022
+"""
+# imports
+import io
+import typing
+
+# external imports
+import PIL.Image
+
+# bsie imports
+from bsie.utils import bsfs, node, ns
+
+# inner-module imports
+from . import base
+
+# exports
+__all__: typing.Sequence[str] = (
+    'Preview',
+    )
+
+
+## code ##
+
+class Preview(base.Extractor):
+    """Extract previews."""
+
+    CONTENT_READER = 'bsie.reader.preview.Preview'
+
+    def __init__(self, max_sides: typing.Iterable[int]):
+        super().__init__(bsfs.schema.from_string(base.SCHEMA_PREAMBLE + '''
+
+            bsfs:Preview rdfs:subClassOf bsfs:Node .
+            bsfs:BinaryBlob rdfs:subClassOf bsfs:Literal .
+            bsfs:JPEG rdfs:subClassOf bsfs:BinaryBlob .
+
+            bse:preview rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsfs:File ;
+                rdfs:range bsfs:Preview ;
+                bsfs:unique "false"^^xsd:boolean .
+
+            bsp:width rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsfs:Preview ;
+                rdfs:range xsd:integer ;
+                bsfs:unique "true"^^xsd:boolean .
+
+            bsp:height rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsfs:Preview ;
+                rdfs:range xsd:integer ;
+                bsfs:unique "true"^^xsd:boolean .
+
+            bsp:asset rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsfs:Preview ;
+                rdfs:range bsfs:JPEG ;
+                bsfs:unique "true"^^xsd:boolean .
+
+            '''))
+        # initialize extra args
+        self.max_sides = set(max_sides)
+
+    def __eq__(self, other: typing.Any) -> bool:
+        return super().__eq__(other) \
+           and self.max_sides == other.max_sides
+
+    def __hash__(self) -> int:
+        return hash((super().__hash__(), tuple(sorted(self.max_sides))))
+
+    def extract(
+            self,
+            subject: node.Node,
+            content: typing.Callable[[int], PIL.Image.Image],
+            principals: typing.Iterable[bsfs.schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[node.Node, bsfs.schema.Predicate, typing.Any]]:
+        # check principals
+        if self.schema.predicate(ns.bse.preview) not in principals:
+            return
+
+        for max_side in self.max_sides:
+            # get the preview in the right resolution
+            img = content(max_side)
+            # convert the preview to jpeg
+            buffer = io.BytesIO()
+            img.save(buffer, format='jpeg')
+            # create a preview node
+            preview = node.Node(ns.bsfs.Preview,
+                ucid=bsfs.uuid.UCID.from_bytes(buffer.getvalue()),
+                size=max_side,
+                source=subject,
+                )
+            # yield triples
+            yield subject, self.schema.predicate(ns.bse.preview), preview
+            yield preview, self.schema.predicate(ns.bsp.width), img.width
+            yield preview, self.schema.predicate(ns.bsp.height), img.height
+            yield preview, self.schema.predicate(ns.bsp.asset), buffer.getvalue()
+
+## EOF ##