Make reset_db great again

.gitignore

@@ -3,3 +3,4 @@
 */migrations/*_auto_*.py
 /.datastore/
 /.idea/
+/import.log

meanings/management/commands/import_plWordnet.py

 #! /usr/bin/python
 # -*- coding: utf-8 -*-
+import zipfile
 
 from django.core.management.base import BaseCommand
 
@@ -93,13 +94,14 @@ class Command(BaseCommand):
         import_plWordnet()
         
 def import_plWordnet():
-    xml_file = os.path.join(BASE_DIR, 'data', 'plwordnet', 'plwordnet_2_1.xml')
-    xml_path = os.path.join(os.path.dirname(os.path.abspath(__file__)), xml_file)
+    wordnet_dir = os.path.join(BASE_DIR, 'data', 'plwordnet')
+    zipped_xml_file = os.path.join(wordnet_dir, 'plwordnet_2_1.xml.zip')
 
     parser = make_parser()
     parser.setContentHandler(PlWNHandler())
     print("Parsing Wordnet...")
-    parser.parse(xml_path)
+    with zipfile.ZipFile(zipped_xml_file, 'r') as zip_file:
+        parser.parse(zip_file.open("data/plwordnet/plwordnet_2_1.xml"))
     print("...DONE")
 
     print()
@@ -140,12 +142,14 @@ def import_plWordnet():
         i += 1
         try:
             child = Synset.objects.get(id=child_id)
-            parents = [Synset.objects.get(id=parent_id) for parent_id in parent_ids]
-        except:
-            print('************', child_id, parent_ids)
+        except Synset.DoesNotExist:
+            print(f'************ Missing Synset {child_id}')
             continue
+        parents = list(Synset.objects.filter(id__in=parent_ids).only("id"))
+        missing_parent_ids = set(parent_ids) - {p.id for p in parents}
+        if missing_parent_ids:
+            print(f'************ Missing parent Synsets for {child_id}: {missing_parent_ids}')
         if i % 2000 == 0:
             print(i, child, parents)
         child.hypernyms.add(*parents)
-        child.save()
     print("...DONE")

reset_db.sh

 #!/bin/bash
 
-sudo su postgres -c "dropdb shellvalier"
-sudo su postgres -c "createdb shellvalier -E UTF8 -T template0 -l pl_PL.utf8"
+set -e
+
+python manage.py reset_db --noinput
 
 # reset the migrations since we create the DB from scratch
 find . -path "*/migrations/*.py" -not -name "__init__.py" -delete
@@ -10,7 +11,7 @@ find . -path "*/migrations/*.pyc"  -delete
 python manage.py makemigrations
 python manage.py migrate
 
-rm import.log
+rm import.log || true
 
 time python manage.py start_import
 time python manage.py import_expansions
@@ -21,7 +22,7 @@ head import.log
 wc import.log
 
 #python manage.py check_descriptions
-python manage.py generate_semantics_css
+#python manage.py generate_semantics_css
 
 # TODO dev only!!!
 python manage.py shell -c "from django.contrib.auth.models import User; User.objects.create_superuser('shell', '', 'valier')"

syntax/management/commands/import_tei.py

@@ -71,7 +71,6 @@ def import_constants():
     import_phrase_attributes()
     import_lemma_operators()
     import_modification_types()
-    pass
 
 def import_poses():
     poses = [u'unk', u'adj', u'noun', u'adv', u'verb']