IPTC tag extraction

4 files changed, 108 insertions, 0 deletions

diff --git a/bsie/extractor/image/iptc.py b/bsie/extractor/image/iptc.py
new file mode 100644
index 0000000..195eff7
--- /dev/null
+++ b/bsie/extractor/image/iptc.py
@@ -0,0 +1,70 @@
+
+# standard imports
+import typing
+
+# bsie imports
+from bsie.utils import bsfs, node, ns
+
+# inner-module imports
+from .. import base
+
+# exports
+__all__: typing.Sequence[str] = (
+    'Iptc',
+    )
+
+
+## code ##
+
+class Iptc(base.Extractor):
+    """Turn IPTC keywords into tags."""
+
+    CONTENT_READER = 'bsie.reader.exif.Iptc'
+
+    def __init__(self):
+        super().__init__(bsfs.schema.from_string(base.SCHEMA_PREAMBLE + '''
+            bsn:Tag rdfs:subClassOf bsfs:Node .
+
+            bse:tag rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsn:Entity ;
+                rdfs:range bsn:Tag .
+
+            <https://schema.bsfs.io/ie/Node/Tag#label> rdfs:subClassOf bsfs:Predicate ;
+                rdfs:domain bsn:Tag ;
+                rdfs:range xsd:string ;
+                bsfs:unique "true"^^xsd:boolean .
+
+            '''))
+        self._callmap = {
+            self.schema.predicate(ns.bse.tag): self._keywords,
+            }
+
+    def extract(
+            self,
+            subject: node.Node,
+            content: dict,
+            principals: typing.Iterable[bsfs.schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[node.Node, bsfs.schema.Predicate, typing.Any]]:
+        for pred in principals:
+            # find callback
+            clbk = self._callmap.get(pred)
+            if clbk is None:
+                continue
+            # produce triples
+            yield from clbk(subject, content)
+
+    def _keywords(
+            self,
+            subject: node.Node,
+            content: dict,
+            ) -> typing.Iterator[typing.Tuple[node.Node, bsfs.schema.Predicate, typing.Any]]:
+        if 'Iptc.Application2.Keywords' not in content:
+            return
+        for keyword in content['Iptc.Application2.Keywords']:
+            tag = node.Node(ns.bsn.Tag, label=keyword)
+            yield subject, self.schema.predicate(ns.bse.tag), tag
+            yield tag, self.schema.predicate(ns.bst.label), keyword
+
+
+
+## EOF ##
diff --git a/bsie/lib/naming_policy.py b/bsie/lib/naming_policy.py
index 9b9a45d..3e7c940 100644
--- a/bsie/lib/naming_policy.py
+++ b/bsie/lib/naming_policy.py
@@ -4,6 +4,9 @@ import abc
 import os
 import typing
 
+# external imports
+import urllib.parse
+
 # bsie imports
 from bsie.utils import bsfs, errors, ns
 from bsie.utils.node import Node
@@ -84,6 +87,8 @@ class DefaultNamingPolicy(NamingPolicy):
             return self.name_file(node)
         if node.node_type == ns.bsn.Preview:
             return self.name_preview(node)
+        if node.node_type == ns.bsn.Tag:
+            return self.name_tag(node)
         raise errors.ProgrammingError('no naming policy available for {node.node_type}')
 
     def name_file(self, node: Node) -> Node:
@@ -112,4 +117,14 @@ class DefaultNamingPolicy(NamingPolicy):
         node.uri = getattr(self._prefix.preview(), fragment)
         return node
 
+    def name_tag(self, node: Node) -> Node:
+        # NOTE: Must ensure to produce the same name for that tags with the same label.
+        if 'label' in node.hints: # tag label
+            fragment = urllib.parse.quote(node.hints['label'])
+        else: # random name
+            fragment = self._uuid()
+        # FIXME: match to existing tags in bsfs storage!
+        node.uri = getattr(self._prefix.tag(), fragment)
+        return node
+
 ## EOF ##
diff --git a/bsie/reader/exif.py b/bsie/reader/exif.py
index 2d0428b..7ec7574 100644
--- a/bsie/reader/exif.py
+++ b/bsie/reader/exif.py
@@ -17,6 +17,7 @@ MATCH_RULE = 'mime=image/jpeg'
 # exports
 __all__: typing.Sequence[str] = (
     'Exif',
+    'Iptc',
     )
 
 
@@ -41,4 +42,24 @@ class Exif(base.Reader):
         except (TypeError, OSError, RuntimeError) as err:
             raise errors.ReaderError(path) from err
 
+
+class Iptc(base.Reader):
+    """Use pyexiv2 to read iptc metadata from image files."""
+
+    def __init__(self):
+        self._match = filematcher.parse(MATCH_RULE)
+
+    def __call__(self, path: str) -> dict:
+        # perform quick checks first
+        if not self._match(path):
+            raise errors.UnsupportedFileFormatError(path)
+
+        try:
+            # open the file
+            img = pyexiv2.Image(path)
+            # read metadata
+            return img.read_iptc()
+        except (TypeError, OSError, RuntimeError) as err:
+            raise errors.ReaderError(path) from err
+
 ## EOF ##
diff --git a/bsie/utils/namespaces.py b/bsie/utils/namespaces.py
index 4a66048..9357253 100644
--- a/bsie/utils/namespaces.py
+++ b/bsie/utils/namespaces.py
@@ -20,6 +20,7 @@ bsf = bsie.Literal.Array.Feature
 bsl = bsfs.Literal
 bsn = bsie.Node
 bsp = bsie.Node.Preview()
+bst = bsie.Node.Tag()
 
 # export
 __all__: typing.Sequence[str] = (
@@ -32,6 +33,7 @@ __all__: typing.Sequence[str] = (
     'bsl',
     'bsn',
     'bsp',
+    'bst',
     'xsd',
     )