[F] Sdasm

metadata_mapper/mappers/flickr/sdasm_mapper.py

@@ -3,22 +3,6 @@
 
 
 class SdasmRecord(FlickrRecord):
-
-    IDENTIFIER_FIELDS = [
-        "PictionID",
-        "Catalog",
-        "Filename"
-    ]
-
-    DATE_FIELDS = [
-        "Date on Neg",
-        "Date"
-    ]
-
-    FIELD_BEFORE_LABEL = r"(^| - )"
-
-    FIELD_AFTER_LABEL = r":\s*"
-
     def UCLDC_map(self):
         return {
             "description": self.map_description,
@@ -29,48 +13,126 @@ def UCLDC_map(self):
     def source_description(self):
         return self.source_metadata.get("description", {}).get("_content")
 
-    def date_field_regex(self, field_name):
-        return self.field_wrap(field_name) + r"([\d/]+)"
-
-    def identifier_field_regex(self, field_name):
-        return self.field_wrap(field_name) + r"([-.\w]+)"
-
-    def field_wrap(self, field_name):
-        return self.FIELD_BEFORE_LABEL + re.escape(field_name) + \
-               self.FIELD_AFTER_LABEL
-
-    def map_description(self):
-        return self.source_description
-
-    def get_matches(self, regexes):
+    @staticmethod
+    def get_mapping_configuration():
+        """
+        Provides configuration for locating and extracting metadata values from
+        the description field using regex. The "key" field gives the regex a
+        name. The capture group in the regex is the value that we extract for
+        that field. The entire match (including the parts outside the capture
+        group), are replaced with an empty string in the description field.
+
+        There are two formats that SDASM uses. One separates values with " - ".
+        The other separates with linebreaks. The fields that use these two
+        formats don't appear to mix fields, which is why this works. It might be
+        possible to match both formats, but doesn't seem necessary at this time.
+        """
+        return [
+            {
+                "key": "piction_id",
+                "regex": r"(^| - )PictionID: ?(\S+)"
+            },
+            {
+                "key": "catalog",
+                "regex": r"(^| - )Catalog: ?(\S+)"
+            },
+            {
+                "key": "filename",
+                "regex": r"(^| - )Filename: ?(\S+)"
+            },
+            {
+                "key": "date",
+                "regex": r"(^| - )Date on Neg: ?(\S+)"
+            },
+            {
+                "key": "year",
+                "regex": r"(^| - )Year: ?([^\n]+)\n"
+            },
+            {
+                "key": "date_on_neg",
+                "regex": r"(^| - )Date: ?(\S+)",
+            },
+            {
+                "key": "sdasm_catalog",
+                "regex": r"^SDASM Catalog #: ?([^\n]+)\n"
+            },
+            {
+                "key": "corp_name",
+                "regex": r"^Corp. Name: ?([^\n]+)\n"
+            },
+            {
+                "key": "title",
+                "regex": r"^Title: ?([^\n]+)\n"
+            },
+            {
+                "key": "catalog_or_negative_number",
+                "regex": r"^Catalog or Negative #: ?([^\n]+)\n"
+            },
+            {
+                "key": "media_negative_size",
+                "regex": r"^Media +\(negative size\): ?([^\n]+)\n"
+            },
+            {
+               "key": "description",
+               "regex": r"Description: ?([^\n]*)\n"
+            },
+            {
+                "key": "repository",
+                "regex": r"Repository:(</b>)? ?([^\n]*)$"
+            }
+        ]
+
+    def split_description(self):
         description = self.source_description
+        description_parts = {}
 
-        matches = []
-        for regex in regexes:
-            matches = re.search(regex, description, re.IGNORECASE)
+        for field_configuration in self.get_mapping_configuration():
+            matches = re.search(field_configuration.get("regex"),
+                                self.source_description, re.MULTILINE)
             if not matches:
                 continue
 
-            matches.append(matches.group(2))
-        return matches
+            description = description.replace(matches.group(0), "")
 
-    def map_date(self):
-        if not self.source_description:
-            return
+            description_parts.update({field_configuration.get("key"):
+                                          matches.groups()[-1].strip()})
+
+        # Set the description if it wasn't provided as metadata in the
+        # description field
+        if "description" not in description_parts:
+            description_parts.update({"description": description})
 
-        date = self.get_matches([self.date_field_regex(field)
-                                 for field in self.DATE_FIELDS])
+        return description_parts
 
-        return date[-1] if date else None
+    def map_date(self):
+        return list(filter(None, [
+            self.split_description().get("date"),
+            self.split_description().get("date_on_neg"),
+            self.split_description().get("year")
+        ]))
 
     def map_identifier(self):
-        if not self.source_description:
-            return
+        return list(filter(None, [
+            self.split_description().get("piction_id"),
+            self.split_description().get("catalog"),
+            self.split_description().get("filename"),
+            self.split_description().get("sdasm_catalog"),
+            self.split_description().get("catalog_or_negative_number")
+        ]))
+
+    def map_description(self):
+        description = self.split_description().get("description")
+        description = description.replace("---Please Tag these images so that "
+                                          "the information can be permanently "
+                                          "stored with the digital file.---",
+                                          "")
+        description = re.sub(r"( +-){2,}", " -", description)
+        description = re.sub(r"\A -", "", description, re.MULTILINE)
 
-        return list(set([self.identifier_field_regex(field)
-                         for field in self.IDENTIFIER_FIELDS]))
+        # Extracting the title, from "Title:" to " - ", is possible, but would
+        # require some close analysis to review the results.
+        return description
 
 
 class SdasmVernacular(FlickrVernacular):
     record_cls = SdasmRecord
-