Merge pull request #1113 from bookwyrm-social/book-format-choices

Adds book format field with choices

bookwyrm/activitypub/book.py

@@ -54,6 +54,7 @@ class Edition(Book):
     asin: str = ""
     pages: int = None
     physicalFormat: str = ""
+    physicalFormatDetail: str = ""
     publishers: List[str] = field(default_factory=lambda: [])
     editionRank: int = 0
 

bookwyrm/connectors/abstract_connector.py

@@ -9,6 +9,7 @@ from requests.exceptions import RequestException
 
 from bookwyrm import activitypub, models, settings
 from .connector_manager import load_more_data, ConnectorException
+from .format_mappings import format_mappings
 
 
 logger = logging.getLogger(__name__)
@@ -312,3 +313,25 @@ class Mapping:
             return self.formatter(value)
         except:  # pylint: disable=bare-except
             return None
+
+
+def infer_physical_format(format_text):
+    """try to figure out what the standardized format is from the free value"""
+    format_text = format_text.lower()
+    if format_text in format_mappings:
+        # try a direct match
+        return format_mappings[format_text]
+    # failing that, try substring
+    matches = [v for k, v in format_mappings.items() if k in format_text]
+    if not matches:
+        return None
+    return matches[0]
+
+
+def unique_physical_format(format_text):
+    """only store the format if it isn't diretly in the format mappings"""
+    format_text = format_text.lower()
+    if format_text in format_mappings:
+        # try a direct match, so saving this would be redundant
+        return None
+    return format_text

bookwyrm/connectors/format_mappings.py

@@ -0,0 +1,43 @@
+""" comparing a free text format to the standardized one """
+format_mappings = {
+    "paperback": "Paperback",
+    "soft": "Paperback",
+    "pamphlet": "Paperback",
+    "peperback": "Paperback",
+    "tapa blanda": "Paperback",
+    "turtleback": "Paperback",
+    "pocket": "Paperback",
+    "spiral": "Paperback",
+    "ring": "Paperback",
+    "平装": "Paperback",
+    "简装": "Paperback",
+    "hardcover": "Hardcover",
+    "hardcocer": "Hardcover",
+    "hardover": "Hardcover",
+    "hardback": "Hardcover",
+    "library": "Hardcover",
+    "tapa dura": "Hardcover",
+    "leather": "Hardcover",
+    "clothbound": "Hardcover",
+    "精装": "Hardcover",
+    "ebook": "EBook",
+    "e-book": "EBook",
+    "digital": "EBook",
+    "computer file": "EBook",
+    "epub": "EBook",
+    "online": "EBook",
+    "pdf": "EBook",
+    "elektronische": "EBook",
+    "electronic": "EBook",
+    "audiobook": "AudiobookFormat",
+    "audio": "AudiobookFormat",
+    "cd": "AudiobookFormat",
+    "dvd": "AudiobookFormat",
+    "mp3": "AudiobookFormat",
+    "cassette": "AudiobookFormat",
+    "kindle": "AudiobookFormat",
+    "talking": "AudiobookFormat",
+    "sound": "AudiobookFormat",
+    "comic": "GraphicNovel",
+    "graphic": "GraphicNovel",
+}

bookwyrm/connectors/inventaire.py

@@ -8,7 +8,7 @@ from .connector_manager import ConnectorException
 
 
 class Connector(AbstractConnector):
-    """instantiate a connector for OL"""
+    """instantiate a connector for inventaire"""
 
     def __init__(self, identifier):
         super().__init__(identifier)

bookwyrm/connectors/openlibrary.py

@@ -3,7 +3,7 @@ import re
 
 from bookwyrm import models
 from .abstract_connector import AbstractConnector, SearchResult, Mapping
-from .abstract_connector import get_data
+from .abstract_connector import get_data, infer_physical_format, unique_physical_format
 from .connector_manager import ConnectorException
 from .openlibrary_languages import languages
 
@@ -43,7 +43,16 @@ class Connector(AbstractConnector):
             ),
             Mapping("publishedDate", remote_field="publish_date"),
             Mapping("pages", remote_field="number_of_pages"),
-            Mapping("physicalFormat", remote_field="physical_format"),
+            Mapping(
+                "physicalFormat",
+                remote_field="physical_format",
+                formatter=infer_physical_format,
+            ),
+            Mapping(
+                "physicalFormatDetail",
+                remote_field="physical_format",
+                formatter=unique_physical_format,
+            ),
             Mapping("publishers"),
         ]
 

bookwyrm/migrations/0101_auto_20210929_1847.py

@@ -0,0 +1,56 @@
+# Generated by Django 3.2 on 2021-05-21 00:17
+
+from django.db import migrations
+import bookwyrm
+from bookwyrm.connectors.abstract_connector import infer_physical_format
+
+
+def infer_format(app_registry, schema_editor):
+    """set the new phsyical format field based on existing format data"""
+    db_alias = schema_editor.connection.alias
+
+    editions = (
+        app_registry.get_model("bookwyrm", "Edition")
+        .objects.using(db_alias)
+        .filter(physical_format_detail__isnull=False)
+    )
+    for edition in editions:
+        free_format = edition.physical_format_detail.lower()
+        edition.physical_format = infer_physical_format(free_format)
+        edition.save()
+
+
+def reverse(app_registry, schema_editor):
+    """doesn't need to do anything"""
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("bookwyrm", "0100_shelf_description"),
+    ]
+
+    operations = [
+        migrations.RenameField(
+            model_name="edition",
+            old_name="physical_format",
+            new_name="physical_format_detail",
+        ),
+        migrations.AddField(
+            model_name="edition",
+            name="physical_format",
+            field=bookwyrm.models.fields.CharField(
+                blank=True,
+                choices=[
+                    ("AudiobookFormat", "Audiobook"),
+                    ("EBook", "eBook"),
+                    ("GraphicNovel", "Graphic novel"),
+                    ("Hardcover", "Hardcover"),
+                    ("Paperback", "Paperback"),
+                ],
+                max_length=255,
+                null=True,
+            ),
+        ),
+        migrations.RunPython(infer_format, reverse),
+    ]

bookwyrm/models/book.py

@@ -6,6 +6,7 @@ from django.contrib.postgres.indexes import GinIndex
 from django.db import models, transaction
 from django.db.models import Prefetch
 from django.dispatch import receiver
+from django.utils.translation import gettext_lazy as _
 from model_utils import FieldTracker
 from model_utils.managers import InheritanceManager
 from imagekit.models import ImageSpecField
@@ -226,6 +227,16 @@ class Work(OrderedCollectionPageMixin, Book):
     deserialize_reverse_fields = [("editions", "editions")]
 
 
+# https://schema.org/BookFormatType
+FormatChoices = [
+    ("AudiobookFormat", _("Audiobook")),
+    ("EBook", _("eBook")),
+    ("GraphicNovel", _("Graphic novel")),
+    ("Hardcover", _("Hardcover")),
+    ("Paperback", _("Paperback")),
+]
+
+
 class Edition(Book):
     """an edition of a book"""
 
@@ -243,7 +254,10 @@ class Edition(Book):
         max_length=255, blank=True, null=True, deduplication_field=True
     )
     pages = fields.IntegerField(blank=True, null=True)
-    physical_format = fields.CharField(max_length=255, blank=True, null=True)
+    physical_format = fields.CharField(
+        max_length=255, choices=FormatChoices, null=True, blank=True
+    )
+    physical_format_detail = fields.CharField(max_length=255, blank=True, null=True)
     publishers = fields.ArrayField(
         models.CharField(max_length=255), blank=True, default=list
     )

bookwyrm/templates/book/edit_book.html

@@ -251,13 +251,28 @@
 
             <div class="block">
                 <h2 class="title is-4">{% trans "Physical Properties" %}</h2>
+                <div class="columns">
+                    <div class="column is-one-third">
                         <div class="field">
                             <label class="label" for="id_physical_format">{% trans "Format:" %}</label>
+                            <div class="select">
                                 {{ form.physical_format }}
+                            </div>
                             {% for error in form.physical_format.errors %}
                             <p class="help is-danger">{{ error | escape }}</p>
                             {% endfor %}
                         </div>
+                    </div>
+                    <div class="column">
+                        <div class="field">
+                            <label class="label" for="id_physical_format_detail">{% trans "Format details:" %}</label>
+                            {{ form.physical_format_detail }}
+                            {% for error in form.physical_format_detail.errors %}
+                            <p class="help is-danger">{{ error | escape }}</p>
+                            {% endfor %}
+                        </div>
+                    </div>
+                </div>
 
                 <div class="field">
                     <label class="label" for="id_pages">{% trans "Pages:" %}</label>

bookwyrm/templates/book/publisher_info.html

@@ -4,13 +4,15 @@
 {% load humanize %}
 
 <p>
-    {% with format=book.physical_format pages=book.pages %}
+    {% firstof book.physical_format_detail book.physical_format as format %}
+    {% firstof book.physical_format book.physical_format_detail as format_property %}
+    {% with pages=book.pages %}
         {% if format %}
             {% comment %}
                 @todo The bookFormat property is limited to a list of values whereas the book edition is free text.
                 @see https://schema.org/bookFormat
             {% endcomment %}
-            <meta itemprop="bookFormat" content="{{ format }}">
+            <meta itemprop="bookFormat" content="{{ format_property }}">
         {% endif %}
 
         {% if pages %}