Merge pull request #1602 from open-apparel-registry/feature/jcw/paren…

…t-company

CHANGELOG.md

@@ -25,6 +25,7 @@ and this project adheres to [Semantic Versioning](http://semver.org/spec/v2.0.0.
 
 - Add extended field handling to facility details [#1593](https://github.com/open-apparel-registry/open-apparel-registry/pull/1593)
 - Add searchability to embed config [#1598](https://github.com/open-apparel-registry/open-apparel-registry/pull/1598)
+- Create parent company extended fields [#1602](https://github.com/open-apparel-registry/open-apparel-registry/pull/1602)
 
 ### Changed
 

src/app/src/util/constants.js

@@ -720,7 +720,7 @@ export const EXTENDED_FIELD_TYPES = [
     {
         label: 'Parent Company',
         fieldName: 'parent_company',
-        formatValue: v => v,
+        formatValue: v => v.contributor_name || v.name || v.raw_value,
     },
     {
         label: 'Facility Type',

src/django/api/extended_fields.py

@@ -1,5 +1,5 @@
 import re
-from api.models import ExtendedField
+from api.models import Contributor, ExtendedField
 
 
 def extract_range_value(value):
@@ -12,6 +12,19 @@ def create_extendedfield(field, field_value, item, contributor):
     if field_value is not None and field_value != "":
         if field == ExtendedField.NUMBER_OF_WORKERS:
             field_value = extract_range_value(field_value)
+        elif field == ExtendedField.PARENT_COMPANY:
+            matches = Contributor.objects.filter_by_name(field_value)
+            if matches.exists():
+                field_value = {
+                    'raw_value': field_value,
+                    'contributor_name': matches[0].name,
+                    'contributor_id': matches[0].id
+                }
+            else:
+                field_value = {
+                    'raw_value': field_value,
+                    'name': field_value
+                }
         ExtendedField.objects.create(
             contributor=contributor,
             facility_list_item=item,
@@ -21,7 +34,8 @@ def create_extendedfield(field, field_value, item, contributor):
 
 
 RAW_DATA_FIELDS = (ExtendedField.NUMBER_OF_WORKERS,
-                   ExtendedField.NATIVE_LANGUAGE_NAME)
+                   ExtendedField.NATIVE_LANGUAGE_NAME,
+                   ExtendedField.PARENT_COMPANY)
 
 
 def create_extendedfields_for_single_item(item, raw_data):

src/django/api/migrations/0076_fill_clean_name_and_address.py

@@ -3,8 +3,6 @@
 
 
 def populate_cleaned_fields(apps, schema_editor):
-    print('')
-    print('Started filling clean name and address')
     count = 0
     FacilityListItem = apps.get_model('api', 'FacilityListItem')
     for list_item in FacilityListItem.objects.exclude(name='', address='').iterator():
@@ -14,7 +12,6 @@ def populate_cleaned_fields(apps, schema_editor):
         count += 1
         if count % 1000 == 0:
             print('Filled ' + str(count))
-    print('Finished filling clean name and address')
 
 
 def do_nothing_on_reverse(apps, schema_editor):

src/django/api/models.py

@@ -8,6 +8,7 @@
 from django.contrib.postgres import fields as postgres
 from django.contrib.postgres.indexes import GinIndex
 from django.contrib.postgres.aggregates.general import ArrayAgg
+from django.contrib.postgres.search import TrigramSimilarity
 from django.db import models, transaction
 from django.db.models import (F, Q, ExpressionWrapper)
 from django.db.models.signals import post_save
@@ -85,6 +86,33 @@ def create_superuser(self, email, password, **extra_fields):
         return self._create_user(email, password, **extra_fields)
 
 
+class ContributorManager(models.Manager):
+    TRIGRAM_SIMILARITY_THRESHOLD = 0.5
+
+    def filter_by_name(self, name):
+        """
+        Perform a fuzzy match on contributor name where the match exceeds a
+        confidence trheshold. The results are ordered by similarity, then
+        whether the Contirbutor is verified, then by whether the contributor
+        has active sources.
+
+        False is less than True so we order_by boolean fields in descending
+        order
+        """
+        threshold = ContributorManager.TRIGRAM_SIMILARITY_THRESHOLD
+        matches = self \
+            .annotate(active_source_count=models.Count(
+                Q(source__is_active=True))) \
+            .annotate(
+                has_active_sources=ExpressionWrapper(
+                    Q(active_source_count__gt=0),
+                    models.BooleanField())) \
+            .annotate(similarity=TrigramSimilarity('name', name)) \
+            .filter(similarity__gte=threshold) \
+            .order_by('-similarity', '-is_verified', '-has_active_sources')
+        return matches
+
+
 class Contributor(models.Model):
     """
     A participant in or observer of the supply chain that will
@@ -188,6 +216,7 @@ class Contributor(models.Model):
     created_at = models.DateTimeField(auto_now_add=True)
     updated_at = models.DateTimeField(auto_now=True)
 
+    objects = ContributorManager()
     history = HistoricalRecords()
 
     @staticmethod

src/django/api/tests.py

@@ -33,7 +33,8 @@
                         RequestLog, DownloadLog, FacilityLocation, Source,
                         ApiLimit, ApiBlock, ContributorNotifications,
                         EmbedConfig, EmbedField, NonstandardField,
-                        FacilityActivityReport, ExtendedField)
+                        FacilityActivityReport, ExtendedField,
+                        ContributorManager)
 from api.oar_id import make_oar_id, validate_oar_id
 from api.matching import match_facility_list_items, GazetteerCache
 from api.processing import (parse_facility_list_item,
@@ -7273,3 +7274,155 @@ def test_inactive_match(self):
         self.assertEquals(None, field['value'])
         self.assertEquals('ExtraTwo', field_two['label'])
         self.assertEquals(None, field_two['value'])
+
+
+class ContributorManagerTest(TestCase):
+    fixtures = ['users', 'contributors']
+
+    def test_filter_by_name(self):
+        matches = Contributor.objects.filter_by_name('factory a')
+        self.assertGreater(matches.count(), 0)
+        self.assertEquals(1.0, matches[0].similarity)
+
+        matches = Contributor.objects.filter_by_name('factory')
+        self.assertGreater(matches.count(), 0)
+        self.assertLess(matches[0].similarity, 1.0)
+        self.assertGreater(matches[0].similarity,
+                           ContributorManager.TRIGRAM_SIMILARITY_THRESHOLD)
+
+    def test_filter_by_name_verified(self):
+        user1 = User.objects.create(email='test1@test.com')
+        user2 = User.objects.create(email='test2@test.com')
+        c1 = Contributor \
+            .objects \
+            .create(admin=user1,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+        c2 = Contributor \
+            .objects \
+            .create(admin=user2,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+
+        matches = Contributor.objects.filter_by_name('TESTING')
+        self.assertEqual(2, matches.count())
+        # When the names are the same and neither is verified than the second
+        # contributor happens to sort first
+        self.assertEqual(c2, matches[0])
+
+        c1.is_verified = True
+        c1.save()
+        matches = Contributor.objects.filter_by_name('TESTING')
+        self.assertEqual(2, matches.count())
+        # Marking c1 as verified forces it to sort first
+        self.assertEqual(c1, matches[0])
+
+    def test_filter_by_name_source(self):
+        user1 = User.objects.create(email='test1@test.com')
+        user2 = User.objects.create(email='test2@test.com')
+        c1 = Contributor \
+            .objects \
+            .create(admin=user1,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+        c2 = Contributor \
+            .objects \
+            .create(admin=user2,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+
+        matches = Contributor.objects.filter_by_name('TESTING')
+        self.assertEqual(2, matches.count())
+        # When the names are the same and neither is verified than the second
+        # contributor happens to sort first
+        self.assertEqual(c2, matches[0])
+
+        Source \
+            .objects \
+            .create(source_type=Source.SINGLE,
+                    is_active=True,
+                    is_public=True,
+                    contributor=c1)
+
+        matches = Contributor.objects.filter_by_name('TESTING')
+        self.assertEqual(2, matches.count())
+        # An active source forces it to sort first
+        self.assertEqual(c1, matches[0])
+
+    def test_filter_by_name_verified_and_source(self):
+        user1 = User.objects.create(email='test1@test.com')
+        user2 = User.objects.create(email='test2@test.com')
+        c1 = Contributor \
+            .objects \
+            .create(admin=user1,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+        c2 = Contributor \
+            .objects \
+            .create(admin=user2,
+                    name='TESTING',
+                    contrib_type=Contributor.OTHER_CONTRIB_TYPE)
+
+        Source \
+            .objects \
+            .create(source_type=Source.SINGLE,
+                    is_active=True,
+                    is_public=True,
+                    contributor=c1)
+
+        c2.is_verified = True
+        c2.save()
+
+        matches = Contributor.objects.filter_by_name('TESTING')
+        self.assertEqual(2, matches.count())
+        # A verified contributor sorts before one with a source
+        self.assertEqual(c2, matches[0])
+
+
+class ParentCompanyTestCase(FacilityAPITestCaseBase):
+    def setUp(self):
+        super(ParentCompanyTestCase, self).setUp()
+        self.url = reverse('facility-list')
+
+    def join_group_and_login(self):
+        self.client.logout()
+        group = auth.models.Group.objects.get(
+            name=FeatureGroups.CAN_SUBMIT_FACILITY,
+        )
+        self.user.groups.set([group.id])
+        self.user.save()
+        self.client.login(email=self.user_email,
+                          password=self.user_password)
+
+    def test_submit_parent_company_no_match(self):
+        self.join_group_and_login()
+        self.client.post(self.url, {
+            'country': "US",
+            'name': "Azavea",
+            'address': "990 Spring Garden St., Philadelphia PA 19123",
+            'parent_company': 'A random value'
+        })
+        self.assertEqual(1, ExtendedField.objects.all().count())
+        ef = ExtendedField.objects.first()
+        self.assertEqual(ExtendedField.PARENT_COMPANY, ef.field_name)
+        self.assertEqual({
+            'raw_value': 'A random value',
+            'name': 'A random value'
+        }, ef.value)
+
+    def test_submit_parent_company_fuzzy_match(self):
+        self.join_group_and_login()
+        self.client.post(self.url, {
+            'country': "US",
+            'name': "Azavea",
+            'address': "990 Spring Garden St., Philadelphia PA 19123",
+            'parent_company': 'TEST CNTRIBUTOR'
+        })
+        self.assertEqual(1, ExtendedField.objects.all().count())
+        ef = ExtendedField.objects.first()
+        self.assertEqual(ExtendedField.PARENT_COMPANY, ef.field_name)
+        self.assertEqual({
+            'raw_value': 'TEST CNTRIBUTOR',
+            'contributor_name': self.contributor.name,
+            'contributor_id': self.contributor.id
+        }, ef.value)