use schema and predicate types in extractors

8 files changed, 109 insertions, 56 deletions

diff --git a/bsie/base/errors.py b/bsie/base/errors.py
index f86ffb2..eedce3b 100644
--- a/bsie/base/errors.py
+++ b/bsie/base/errors.py
@@ -8,15 +8,22 @@ Author: Matthias Baumgartner, 2022
 import typing
 
 # exports
-__all__: typing.Sequence[str] = []
+__all__: typing.Sequence[str] = (
+    'ExtractorError',
+    )
+
+
 
 
 ## code ##
 
-class _BSIE_Error(Exception):
+class _BSIEError(Exception):
     """Generic BSIE error."""
 
-class ReaderError(_BSIE_Error):
+class ExtractorError(_BSIEError):
+    """The Extractor failed to process the given content."""
+
+class ReaderError(_BSIEError):
     """The Reader failed to read the given file."""
 
 ## EOF ##
diff --git a/bsie/base/extractor.py b/bsie/base/extractor.py
index ea43925..a6a69c6 100644
--- a/bsie/base/extractor.py
+++ b/bsie/base/extractor.py
@@ -11,13 +11,38 @@ import typing
 # inner-module imports
 from . import reader
 from bsie.utils import node
-from bsie.utils.bsfs import URI, typename
+from bsie.utils.bsfs import schema as _schema, typename
 
 # exports
 __all__: typing.Sequence[str] = (
     'Extractor',
     )
 
+# constants
+
+# essential definitions typically used in extractor schemas.
+# NOTE: The definition here is only for convenience; Each Extractor must implement its use, if so desired.
+SCHEMA_PREAMBLE = '''
+    # common external prefixes
+    prefix owl: <http://www.w3.org/2002/07/owl#>
+    prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
+    prefix rdfs: <http://www.w3.org/2000/01/rdf-schema#>
+    prefix xsd: <http://www.w3.org/2001/XMLSchema#>
+    prefix schema: <http://schema.org/>
+
+    # common bsfs prefixes
+    prefix bsfs: <http://bsfs.ai/schema/>
+    prefix bse: <http://bsfs.ai/schema/Entity#>
+
+    # essential nodes
+    bsfs:Entity rdfs:subClassOf bsfs:Node .
+
+    # common definitions
+    xsd:string rdfs:subClassOf bsfs:Literal .
+    xsd:integer rdfs:subClassOf bsfs:Literal .
+
+    '''
+
 
 ## code ##
 
@@ -27,23 +52,37 @@ class Extractor(abc.ABC):
     # what type of content is expected (i.e. reader subclass).
     CONTENT_READER: typing.Optional[typing.Type[reader.Reader]] = None
 
+    # extractor schema.
+    schema: _schema.Schema
+
+    def __init__(self, schema: _schema.Schema):
+        self.schema = schema
+
     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."""
+
+    def predicates(self) -> typing.Iterator[_schema.Predicate]:
+        """Return the predicates that may be part of extracted triples."""
+        # NOTE: Some predicates in the schema might not occur in actual triples,
+        # but are defined due to predicate class hierarchy. E.g., bsfs:Predicate
+        # is part of every schema but should not be used in triples.
+        # Announcing all predicates might not be the most efficient way, however,
+        # it is the most safe one. Concrete extractors that produce additional
+        # predicates (e.g. auxiliary nodes with their own predicates) should
+        # overwrite this method to only include the principal predicates.
+        return self.schema.predicates()
 
     @abc.abstractmethod
     def extract(
             self,
             subject: node.Node,
             content: typing.Any,
-            predicates: typing.Iterable[URI],
-            ) -> typing.Iterator[typing.Tuple[node.Node, URI, typing.Any]]:
+            predicates: typing.Iterable[_schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[node.Node, _schema.Predicate, typing.Any]]:
         """Return (node, predicate, value) triples."""
 
 ## EOF ##
diff --git a/bsie/extractor/generic/constant.py b/bsie/extractor/generic/constant.py
index e243131..795bac6 100644
--- a/bsie/extractor/generic/constant.py
+++ b/bsie/extractor/generic/constant.py
@@ -7,9 +7,9 @@ Author: Matthias Baumgartner, 2022
 # imports
 import typing
 
-# inner-module imports
+# bsie imports
 from bsie.base import extractor
-from bsie.utils.bsfs import URI
+from bsie.utils.bsfs import URI, schema as _schema
 from bsie.utils.node import Node
 
 # exports
@@ -25,26 +25,26 @@ class Constant(extractor.Extractor):
 
     CONTENT_READER = None
 
+    # predicate/value pairs to be produced.
+    _tuples: typing.Tuple[typing.Tuple[_schema.Predicate, typing.Any], ...]
+
     def __init__(
             self,
             schema: str,
             tuples: typing.Iterable[typing.Tuple[URI, typing.Any]],
             ):
-        self._schema = schema
-        self._tuples = tuples
-        # FIXME: use schema instance for predicate checking
-        #self._tuples = [(pred, value) for pred, value in tuples if pred in schema]
+        super().__init__(_schema.Schema.from_string(extractor.SCHEMA_PREAMBLE + schema))
+        # NOTE: Raises a KeyError if the predicate is not part of the schema
+        self._tuples = tuple((self.schema.predicate(p_uri), value) for p_uri, value in tuples)
         # FIXME: use schema instance for value checking
 
-    def schema(self) -> str:
-        return self._schema
 
     def extract(
             self,
             subject: Node,
             content: None,
-            predicates: typing.Iterable[URI],
-            ) -> typing.Iterator[typing.Tuple[Node, URI, typing.Any]]:
+            predicates: typing.Iterable[_schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[Node, _schema.Predicate, typing.Any]]:
         for pred, value in self._tuples:
             if pred in predicates:
                 yield subject, pred, value
diff --git a/bsie/extractor/generic/path.py b/bsie/extractor/generic/path.py
index c39bbd2..f358a79 100644
--- a/bsie/extractor/generic/path.py
+++ b/bsie/extractor/generic/path.py
@@ -8,11 +8,10 @@ Author: Matthias Baumgartner, 2022
 import os
 import typing
 
-# inner-module imports
+# bsie imports
 from bsie.base import extractor
 from bsie.utils import node, ns
-from bsie.utils.bsfs import URI
-import bsie.reader.path
+from bsie.utils.bsfs import schema
 
 # exports
 __all__: typing.Sequence[str] = (
@@ -27,30 +26,31 @@ class Path(extractor.Extractor):
 
     CONTENT_READER = bsie.reader.path.Path
 
-    def __init__(self):
-        self.__callmap = {
-            ns.bse.filename:    self.__filename,
-            }
+    # mapping from predicate to handler function.
+    _callmap: typing.Dict[schema.Predicate, typing.Callable[[str], typing.Any]]
 
-    def schema(self) -> str:
-        return '''
-            bse:filename a bsfs:Predicate ;
+    def __init__(self):
+        super().__init__(schema.Schema.from_string(extractor.SCHEMA_PREAMBLE + '''
+            bse:filename rdfs:subClassOf bsfs:Predicate ;
                 rdfs:domain bsfs:Entity ;
                 rdfs:range xsd:string ;
-                rdf:label "File name"^^xsd:string ;
+                rdfs:label "File name"^^xsd:string ;
                 schema:description "Filename of entity in some filesystem."^^xsd:string ;
                 owl:maxCardinality "INF"^^xsd:number .
-            '''
+            '''))
+        self._callmap = {
+            self.schema.predicate(ns.bse.filename): self.__filename,
+            }
 
     def extract(
             self,
             subject: node.Node,
             content: CONTENT_READER.CONTENT_TYPE,
-            predicates: typing.Iterable[URI],
-            ) -> typing.Iterator[typing.Tuple[node.Node, URI, typing.Any]]:
+            predicates: typing.Iterable[schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[node.Node, schema.Predicate, typing.Any]]:
         for pred in predicates:
             # find callback
-            clbk = self.__callmap.get(pred)
+            clbk = self._callmap.get(pred)
             if clbk is None:
                 continue
             # get value
@@ -60,11 +60,15 @@ class Path(extractor.Extractor):
             # produce triple
             yield subject, pred, value
 
-    def __filename(self, path: str) -> str:
+    def __filename(self, path: str) -> typing.Optional[str]:
         try:
             return os.path.basename(path)
-        except Exception:
-            # FIXME: some kind of error reporting (e.g. logging)
+        except Exception: # some error, skip.
+            # FIXME: some kind of error reporting (e.g. logging)?
+            # Options: (a) Fail silently (current); (b) Skip and report to log;
+            # (c) Raise ExtractorError (aborts extraction); (d) separate content type
+            # checks from basename errors (report content type errors, skip basename
+            # errors)
             return None
 
 ## EOF ##
diff --git a/bsie/extractor/generic/stat.py b/bsie/extractor/generic/stat.py
index d74369c..e5387af 100644
--- a/bsie/extractor/generic/stat.py
+++ b/bsie/extractor/generic/stat.py
@@ -5,14 +5,13 @@ A copy of the license is provided with the project.
 Author: Matthias Baumgartner, 2022
 """
 # imports
+import os
 import typing
 
-# inner-module imports
+# bsie imports
 from bsie.base import extractor
 from bsie.utils import node, ns
-from bsie.utils.bsfs import URI
-import bsie.reader.stat
-
+from bsie.utils.bsfs import schema as _schema
 
 # exports
 __all__: typing.Sequence[str] = (
@@ -27,30 +26,31 @@ class Stat(extractor.Extractor):
 
     CONTENT_READER = bsie.reader.stat.Stat
 
-    def __init__(self):
-        self.__callmap = {
-            ns.bse.filesize:    self.__filesize,
-            }
+    # mapping from predicate to handler function.
+    _callmap: typing.Dict[_schema.Predicate, typing.Callable[[os.stat_result], typing.Any]]
 
-    def schema(self) -> str:
-        return '''
-            bse:filesize a bsfs:Predicate ;
+    def __init__(self):
+        super().__init__(_schema.Schema.from_string(extractor.SCHEMA_PREAMBLE + '''
+            bse:filesize rdfs:subClassOf bsfs:Predicate ;
                 rdfs:domain bsfs:Entity ;
                 rdfs:range xsd:integer ;
-                rdf:label "File size"^^xsd:string ;
+                rdfs:label "File size"^^xsd:string ;
                 schema:description "File size of entity in some filesystem."^^xsd:string ;
                 owl:maxCardinality "INF"^^xsd:number .
-            '''
+            '''))
+        self._callmap = {
+            self.schema.predicate(ns.bse.filesize): self.__filesize,
+            }
 
     def extract(
             self,
             subject: node.Node,
             content: CONTENT_READER.CONTENT_TYPE,
-            predicates: typing.Iterable[URI],
-            ) -> typing.Iterator[typing.Tuple[node.Node, URI, typing.Any]]:
+            predicates: typing.Iterable[_schema.Predicate],
+            ) -> typing.Iterator[typing.Tuple[node.Node, _schema.Predicate, typing.Any]]:
         for pred in predicates:
             # find callback
-            clbk = self.__callmap.get(pred)
+            clbk = self._callmap.get(pred)
             if clbk is None:
                 continue
             # get value
@@ -60,7 +60,7 @@ class Stat(extractor.Extractor):
             # produce triple
             yield subject, pred, value
 
-    def __filesize(self, content: CONTENT_READER.CONTENT_TYPE) -> int:
+    def __filesize(self, content: os.stat_result) -> typing.Optional[int]:
         """Return the file size."""
         try:
             return content.st_size
diff --git a/bsie/utils/bsfs.py b/bsie/utils/bsfs.py
index 1ae657c..01ec5d1 100644
--- a/bsie/utils/bsfs.py
+++ b/bsie/utils/bsfs.py
@@ -8,6 +8,7 @@ Author: Matthias Baumgartner, 2022
 import typing
 
 # bsfs imports
+from bsfs import schema
 from bsfs.namespace import Namespace
 from bsfs.utils import URI, typename
 
@@ -15,6 +16,7 @@ from bsfs.utils import URI, typename
 __all__: typing.Sequence[str] = (
     'Namespace',
     'URI',
+    'schema',
     'typename',
     )
 
diff --git a/bsie/utils/namespaces.py b/bsie/utils/namespaces.py
index 67ccc71..13be96b 100644
--- a/bsie/utils/namespaces.py
+++ b/bsie/utils/namespaces.py
@@ -7,13 +7,14 @@ Author: Matthias Baumgartner, 2022
 # imports
 import typing
 
-# bsie imports
+# inner-module imports
 from . import bsfs as _bsfs
 
 # constants
 bse = _bsfs.Namespace('http://bsfs.ai/schema/Entity#')
 bsfs = _bsfs.Namespace('http://bsfs.ai/schema/')
 bsm = _bsfs.Namespace('http://bsfs.ai/schema/meta#')
+xsd = _bsfs.Namespace('http://www.w3.org/2001/XMLSchema#')
 
 # export
 __all__: typing.Sequence[str] = (
diff --git a/bsie/utils/node.py b/bsie/utils/node.py
index 60863a4..3a0f06b 100644
--- a/bsie/utils/node.py
+++ b/bsie/utils/node.py
@@ -12,7 +12,7 @@ from bsie.utils.bsfs import URI
 
 # exports
 __all__: typing.Sequence[str] = (
-    'Node'
+    'Node',
     )