Basic taxonomy management.

requirements.txt

 Django==1.11.2
 django-export-action==0.1.1
+django-mptt==0.8.7
 et-xmlfile==1.0.1
 jdcal==1.3
 olefile==0.44

website/specimens/admin.py

 from django.contrib import admin
 
 from django import forms
-from .models import Specimen, SpecimenLocation, Person, Fixation, Station, Expedition, SpecimenPicture
+
+from mptt.admin import DraggableMPTTAdmin
+
+from .models import Specimen, SpecimenLocation, Person, Fixation, Station, Expedition, SpecimenPicture, Taxon
 from .widgets import LatLongWidget
 
 
@@ -66,4 +69,8 @@ class ExpeditionAdmin(admin.ModelAdmin):
 class SpecimenPictureAdmin(admin.ModelAdmin):
     fields = ('specimen', 'image', 'high_interest')
 
+@admin.register(Taxon)
+class TaxonAdmin(DraggableMPTTAdmin):
+    pass
+
 admin.site.site_header = 'Astapor administration'
\ No newline at end of file

website/specimens/management/commands/csv_import.py → website/specimens/management/commands/import_specimens.py

@@ -11,7 +11,6 @@ from specimens.models import Person, SpecimenLocation, Specimen, Fixation, Exped
 MODELS_TO_TRUNCATE = [Station, Expedition, Fixation, Person, SpecimenLocation, Specimen]
 
 # TODO: document use of this script:
-# - export Google Sheet (specimens) as CSV (separator: comma)
 # - Column name is important, not column order
 # - Lat/lon use comma as a separator
 
@@ -35,7 +34,7 @@ def get_or_create_station_and_expedition(station_name, expedition_name):
 
 
 class Command(BaseCommand):
-    help = 'Initial data import to populate the tables'
+    help = 'Import specimens from a CSV file and attach them to the existing taxonomy.'
 
     def add_arguments(self, parser):
         parser.add_argument('csv_file')

website/specimens/management/commands/import_taxonomy.py

+import csv
+
+from django.core.management.base import BaseCommand
+
+from specimens.models import TaxonRank, Taxon, TaxonStatus, SPECIES_RANK_NAME, SUBGENUS_RANK_NAME
+
+MODELS_TO_TRUNCATE = [Taxon, TaxonRank, TaxonStatus]
+
+
+def create_initial_ranks():
+    TaxonRank.objects.bulk_create([
+        TaxonRank(name='Kingdom'),
+        TaxonRank(name='Phylum'),
+        TaxonRank(name='Class'),
+        TaxonRank(name='Order'),
+        TaxonRank(name='Family'),
+        TaxonRank(name='Genus'),
+        TaxonRank(name=SUBGENUS_RANK_NAME),
+        TaxonRank(name=SPECIES_RANK_NAME)
+    ])
+
+class Command(BaseCommand):
+    help = 'Import taxonomy from a CSV file.'
+
+    def add_arguments(self, parser):
+        parser.add_argument('csv_file')
+
+        parser.add_argument(
+            '--truncate',
+            action='store_true',
+            dest='truncate',
+            default=False,
+            help='Truncate all tables prior to import',
+        )
+
+    def handle(self, *args, **options):
+        self.stdout.write('Importing data from file...')
+
+        with open(options['csv_file']) as csv_file:
+            if options['truncate']:
+                for model in MODELS_TO_TRUNCATE:
+                    self.stdout.write('Truncate model {name} ...'.format(name=model.__name__), ending='')
+                    model.objects.all().delete()
+                    self.stdout.write(self.style.SUCCESS('OK'))
+
+            self.stdout.write('Creating initial ranks...')
+            create_initial_ranks()
+
+            for i, row in enumerate(csv.DictReader(csv_file, delimiter=',')):
+                self.stdout.write('Processing row #{i}...'.format(i=i), ending='')
+
+                species_status, _ = TaxonStatus.objects.get_or_create(name=row['Status'])
+
+                # Starting from the higher ranks
+                kingdom, _ = Taxon.objects.get_or_create(name=row['Kingdom'].strip(),
+                                                         rank=TaxonRank.objects.get(name="Kingdom"))
+                phylum, _ = Taxon.objects.get_or_create(name=row['Phylum'].strip(),
+                                                        rank=TaxonRank.objects.get(name="Phylum"),
+                                                        parent=kingdom)
+                class_taxon, _ = Taxon.objects.get_or_create(name=row['Class'].strip(),
+                                                          rank=TaxonRank.objects.get(name="Class"),
+                                                          parent=phylum)
+                order_taxon, _ = Taxon.objects.get_or_create(name=row['Order'].strip(),
+                                                             rank=TaxonRank.objects.get(name="Order"),
+                                                             parent=class_taxon)
+                family, _ = Taxon.objects.get_or_create(name=row['Family'].strip(),
+                                                        rank=TaxonRank.objects.get(name="Family"),
+                                                        parent=order_taxon)
+                genus, _ = Taxon.objects.get_or_create(name=row['Genus'].strip(),
+                                                        rank=TaxonRank.objects.get(name="Genus"),
+                                                        parent=family)
+                # Subgenus rank is optional
+                subgenus_source = row['Subgenus'].strip()
+                if subgenus_source:
+                    subgenus, _ = Taxon.objects.get_or_create(name=subgenus_source,
+                                                              rank=TaxonRank.objects.get(name=SUBGENUS_RANK_NAME),
+                                                              parent=genus)
+
+                species_parent = subgenus if subgenus_source else genus
+                species, _ = Taxon.objects.get_or_create(name=row['Species'].strip(),
+                                                         rank=TaxonRank.objects.get(name=SPECIES_RANK_NAME),
+                                                         parent=species_parent,
+                                                         status=species_status,
+                                                         aphia_id=row['Aphia_ID'].strip(),
+                                                         authority=row['Authority'].strip())
+
+
+                self.stdout.write(self.style.SUCCESS('OK'))
+

website/specimens/migrations/0001_initial.py

 # -*- coding: utf-8 -*-
-# Generated by Django 1.11.2 on 2017-06-21 07:43
+# Generated by Django 1.11.2 on 2017-06-22 09:08
 from __future__ import unicode_literals
 
 import django.contrib.gis.db.models.fields
 import django.contrib.postgres.fields.ranges
 from django.db import migrations, models
 import django.db.models.deletion
+import mptt.fields
 
 
 class Migration(migrations.Migration):
@@ -78,6 +79,47 @@ class Migration(migrations.Migration):
                 ('expedition', models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, to='specimens.Expedition')),
             ],
         ),
+        migrations.CreateModel(
+            name='Taxon',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('name', models.CharField(max_length=100)),
+                ('aphia_id', models.IntegerField(blank=True, null=True)),
+                ('authority', models.CharField(blank=True, max_length=100, null=True)),
+                ('lft', models.PositiveIntegerField(db_index=True, editable=False)),
+                ('rght', models.PositiveIntegerField(db_index=True, editable=False)),
+                ('tree_id', models.PositiveIntegerField(db_index=True, editable=False)),
+                ('level', models.PositiveIntegerField(db_index=True, editable=False)),
+                ('parent', mptt.fields.TreeForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, related_name='children', to='specimens.Taxon')),
+            ],
+            options={
+                'abstract': False,
+            },
+        ),
+        migrations.CreateModel(
+            name='TaxonRank',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('name', models.CharField(max_length=100)),
+            ],
+        ),
+        migrations.CreateModel(
+            name='TaxonStatus',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('name', models.CharField(max_length=100)),
+            ],
+        ),
+        migrations.AddField(
+            model_name='taxon',
+            name='rank',
+            field=models.ForeignKey(on_delete=django.db.models.deletion.CASCADE, to='specimens.TaxonRank'),
+        ),
+        migrations.AddField(
+            model_name='taxon',
+            name='status',
+            field=models.ForeignKey(blank=True, null=True, on_delete=django.db.models.deletion.CASCADE, to='specimens.TaxonStatus'),
+        ),
         migrations.AddField(
             model_name='specimen',
             name='specimen_location',

website/specimens/models.py

 from django.contrib.gis.db import models
 from django.contrib.postgres.fields import FloatRangeField
 
+from mptt.models import MPTTModel, TreeForeignKey
+
 UNKNOWN_STATION_NAME = '<Unknown>'  # Sometimes we need a "fake" station to link Specimen to Expedition
 
+# Ranks: Sometimes names are used as an identifier...
+SPECIES_RANK_NAME = "Species"
+SUBGENUS_RANK_NAME = "Subgenus"
+
+class TaxonRank(models.Model):
+    name = models.CharField(max_length=100)
+
+    def __str__(self):
+        return self.name
+
+class TaxonStatus(models.Model):
+    name = models.CharField(max_length=100)
+
+class Taxon(MPTTModel):
+    name = models.CharField(max_length=100)
+    rank = models.ForeignKey(TaxonRank)
+    status = models.ForeignKey(TaxonStatus, null=True, blank=True)
+    aphia_id = models.IntegerField(null=True, blank=True)
+    authority = models.CharField(max_length=100, null=True, blank=True)
+
+    parent = TreeForeignKey('self', null=True, blank=True, related_name='children', db_index=True)
+
+    def is_species(self):
+        return self.rank.name == SPECIES_RANK_NAME
+
+    def is_subgenus(self):
+        return self.rank.name == SUBGENUS_RANK_NAME
+
+    def __str__(self):
+        # Specific representation for Species
+        if self.is_species():
+            if self.parent.is_subgenus():
+                name = "{genus_name} ({subgenus_name}) {species_name}".format(species_name=self.name,
+                                                                             subgenus_name=self.parent.name,
+                                                                             genus_name=self.parent.parent.name)
+            else:
+                name = "{genus_name} {species_name} ".format(species_name=self.name, genus_name=self.parent.name)
+        else:
+            name = self.name
+
+        return "{name} [{rank}]".format(name=name, rank=self.rank)
+
+    class MPTTMeta:
+        order_insertion_by = ['name']
+
+
 class Person(models.Model):
     first_name = models.CharField(max_length=100)
     last_name = models.CharField(max_length=100)

website/website/settings.py

@@ -37,6 +37,7 @@ INSTALLED_APPS = [
     'django.contrib.gis',
     'export_action',
     'specimens',
+    'mptt',
 ]
 
 MIDDLEWARE = [