devel can now output a decent csv with selectable fields

staffanm · Nov 22, 2018 · 9c93fed · 9c93fed
1 parent 101a9f4
commit 9c93fed
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 13 deletions.
diff --git a/ferenda/devel.py b/ferenda/devel.py
@@ -737,7 +737,7 @@ def dumpstore(self, format="turtle"):
 #        sub.debug('Sublog message at DEBUG level')
 
     @decorators.action
-    def csvinventory(self, alias):
+    def csvinventory(self, alias, predicates=None):
         """Create an inventory of documents, as a CSV file. 
 
         Only documents that have been parsed and yielded some minimum
@@ -747,14 +747,18 @@ def csvinventory(self, alias):
         :type  alias: str
 
         """
-        predicates = ['basefile',
-                      'subobjects',  # sections that have rdf:type
-                      'rdf:type',
-                      'dcterms:identifier',
-                      'dcterms:title',
-                      'dcterms:published',
-                      'prov:wasGeneratedBy',
-                      ]
+        if predicates is None:
+            predicates = ['basefile',
+                          'subobjects',  # sections that have rdf:type
+                          'rdf:type',
+                          'dcterms:identifier',
+                          'dcterms:title',
+                          'dcterms:published',
+                          'prov:wasGeneratedBy',
+            ]
+        else:
+            # predicates are given as a comma separated list, eg ./ferenda-build.py devel csvinventory kkv rpubl:malnummer,rpubl:avgorandedatum,rinfoex:instanstyp,rinfoex:domstol,rinfoex:upphandlande,rinfoex:leverantor,rinfoex:arendetyp,rinfoex:avgorande
+            predicates = predicates.split(",")
         import csv
         # if six.PY2:
         #     delimiter = b';'
@@ -771,7 +775,9 @@ def csvinventory(self, alias):
         for basefile in repo.store.list_basefiles_for("relate"):
             baseuri = URIRef(repo.canonical_uri(basefile))
             with repo.store.open_distilled(basefile) as fp:
-                row = {'basefile': basefile}
+                row = {}
+                if 'basefile' in predicates:
+                    row['basefile'] = basefile
                 g = Graph().parse(fp, format="xml")
                 for (p, o) in g.predicate_objects(baseuri):
                     qname = g.qname(p)
@@ -787,7 +793,8 @@ def csvinventory(self, alias):
                             # if six.PY2:
                             #     fld = fld.encode("latin-1", errors="replace")
                             row[qname] = fld
-                row['subobjects'] = len(list(g.subject_objects(RDF.type)))
+                if 'subobjects' in predicates:
+                    row['subobjects'] = len(list(g.subject_objects(RDF.type)))
                 writer.writerow(row)
 
     def _repo_from_alias(self, alias, datadir=None, repoconfig=None):

diff --git a/ferenda/sources/legal/se/kkv.py b/ferenda/sources/legal/se/kkv.py
@@ -120,7 +120,14 @@ def extract_head(self, fp, basefile):
     def infer_identifier(self, basefile):
         return self.identifiers[basefile]
 
-    lblmap = {"Domstol:": "dcterms:publisher", # really :creator (KKV is the :publisher) but swedishlegalsource.space.ttl isn't written like that...
+    lblmap = {"Domstol:": "rinfoex:domstol",  # this ad-hoc predicate
+                                              # keeps
+                                              # attributes_to_resource
+                                              # from converting the
+                                              # string into a URI,
+                                              # which we'd like to
+                                              # avoid for now
+              "Instans:": "rinfoex:instanstyp",
               "Målnummer:": "rpubl:malnummer",
               "Ärendemening:": "dcterms:title",
               "Beslutsdatum:": "rpubl:avgorandedatum",
@@ -141,7 +148,7 @@ def extract_metadata(self, rawhead, basefile):
                 assert lbl.endswith(":"), "invalid label %s" % lbl
                 d[self.lblmap[lbl]] = value
         d["dcterms:issued"] = d["rpubl:avgorandedatum"]
-        self.identifiers[basefile] = "%ss dom den %s i mål %s" % (d["dcterms:publisher"],
+        self.identifiers[basefile] = "%ss dom den %s i mål %s" % (d["rinfoex:domstol"],
                                                                   d["rpubl:avgorandedatum"],
                                                                   d["rpubl:malnummer"])
         return d