Export user book data as csv (#1556)

Simple book data export

bookwyrm/templates/preferences/export.html

@@ -0,0 +1,22 @@
+{% extends 'preferences/layout.html' %}
+{% load i18n %}
+
+{% block title %}{% trans "CSV Export" %}{% endblock %}
+
+{% block header %}
+{% trans "CSV Export" %}
+{% endblock %}
+
+{% block panel %}
+<div class="block content">
+    <p class="notification">
+        {% trans "Your export will include all the books on your shelves, books you have reviewed, and books with reading activity." %}
+    </p>
+    <p>
+        <a href="{% url 'prefs-export-file' %}" class="button">
+            <span class="icon icon-download" aria-hidden="true"></span>
+            <span>Download file</span>
+        </a>
+    </p>
+</div>
+{% endblock %}

bookwyrm/templates/preferences/layout.html

@@ -24,6 +24,17 @@
                 <a href="{{ url }}"{% if url in request.path %} class="is-active" aria-selected="true"{% endif %}>{% trans "Delete Account" %}</a>
             </li>
         </ul>
+        <h2 class="menu-label">{% trans "Data" %}</h2>
+        <ul class="menu-list">
+            <li>
+                {% url 'import' as url %}
+                <a href="{{ url }}"{% if url in request.path %} class="is-active" aria-selected="true"{% endif %}>{% trans "Import" %}</a>
+            </li>
+            <li>
+                {% url 'prefs-export' as url %}
+                <a href="{{ url }}"{% if url in request.path %} class="is-active" aria-selected="true"{% endif %}>{% trans "CSV export" %}</a>
+            </li>
+        </ul>
         <h2 class="menu-label">{% trans "Relationships" %}</h2>
         <ul class="menu-list">
             <li>

bookwyrm/tests/views/test_export.py

@@ -0,0 +1,69 @@
+""" test for app action functionality """
+from unittest.mock import patch
+
+from django.http import StreamingHttpResponse
+from django.test import TestCase
+from django.test.client import RequestFactory
+
+from bookwyrm import models, views
+from bookwyrm.tests.validate_html import validate_html
+
+
+@patch("bookwyrm.activitystreams.add_book_statuses_task.delay")
+@patch("bookwyrm.activitystreams.add_status_task.delay")
+@patch("bookwyrm.activitystreams.populate_stream_task.delay")
+@patch("bookwyrm.suggested_users.rerank_suggestions_task.delay")
+@patch("bookwyrm.models.activitypub_mixin.broadcast_task.apply_async")
+class ExportViews(TestCase):
+    """viewing and creating statuses"""
+
+    def setUp(self):
+        """we need basic test data and mocks"""
+        self.factory = RequestFactory()
+        with patch("bookwyrm.suggested_users.rerank_suggestions_task.delay"), patch(
+            "bookwyrm.activitystreams.populate_stream_task.delay"
+        ):
+            self.local_user = models.User.objects.create_user(
+                "mouse@local.com",
+                "mouse@mouse.com",
+                "mouseword",
+                local=True,
+                localname="mouse",
+                remote_id="https://example.com/users/mouse",
+            )
+        self.work = models.Work.objects.create(title="Test Work")
+        self.book = models.Edition.objects.create(
+            title="Test Book",
+            remote_id="https://example.com/book/1",
+            parent_work=self.work,
+            isbn_13="9781234567890",
+            bnf_id="beep",
+        )
+
+    def tst_export_get(self, *_):
+        """request export"""
+        request = self.factory.get("")
+        request.user = self.local_user
+        result = views.Export.as_view()(request)
+        validate_html(result.render())
+
+    def test_export_file(self, *_):
+        """simple export"""
+        models.ShelfBook.objects.create(
+            shelf=self.local_user.shelf_set.first(),
+            user=self.local_user,
+            book=self.book,
+        )
+        request = self.factory.get("")
+        request.user = self.local_user
+        export = views.export_user_book_data(request)
+        self.assertIsInstance(export, StreamingHttpResponse)
+        self.assertEqual(export.status_code, 200)
+        result = list(export.streaming_content)
+        # pylint: disable=line-too-long
+        self.assertEqual(
+            result[0],
+            b"title,author_text,remote_id,openlibrary_key,inventaire_id,librarything_key,goodreads_key,bnf_id,viaf,wikidata,asin,isbn_10,isbn_13,oclc_number,rating,review_name,review_cw,review_content\r\n",
+        )
+        expected = f"Test Book,,{self.book.remote_id},,,,,beep,,,,123456789X,9781234567890,,,,,\r\n"
+        self.assertEqual(result[1].decode("utf-8"), expected)

bookwyrm/urls.py

@@ -475,6 +475,12 @@ urlpatterns = [
         views.ChangePassword.as_view(),
         name="prefs-password",
     ),
+    re_path(r"^preferences/export/?$", views.Export.as_view(), name="prefs-export"),
+    re_path(
+        r"^preferences/export/file/?$",
+        views.export_user_book_data,
+        name="prefs-export-file",
+    ),
     re_path(r"^preferences/delete/?$", views.DeleteUser.as_view(), name="prefs-delete"),
     re_path(r"^preferences/block/?$", views.Block.as_view(), name="prefs-block"),
     re_path(r"^block/(?P<user_id>\d+)/?$", views.Block.as_view()),

bookwyrm/views/__init__.py

@@ -28,6 +28,7 @@ from .admin.user_admin import UserAdmin, UserAdminList
 # user preferences
 from .preferences.change_password import ChangePassword
 from .preferences.edit_user import EditUser
+from .preferences.export import Export, export_user_book_data
 from .preferences.delete_user import DeleteUser
 from .preferences.block import Block, unblock
 

bookwyrm/views/preferences/export.py

@@ -0,0 +1,97 @@
+""" Let users export their book data """
+import csv
+
+from django.contrib.auth.decorators import login_required
+from django.db.models import Q
+from django.http import StreamingHttpResponse
+from django.template.response import TemplateResponse
+from django.views import View
+from django.utils.decorators import method_decorator
+from django.views.decorators.http import require_GET
+
+from bookwyrm import models
+
+# pylint: disable=no-self-use
+@method_decorator(login_required, name="dispatch")
+class Export(View):
+    """Let users export data"""
+
+    def get(self, request):
+        """Request csv file"""
+        return TemplateResponse(request, "preferences/export.html")
+
+
+@login_required
+@require_GET
+def export_user_book_data(request):
+    """Streaming the csv file of a user's book data"""
+    data = (
+        models.Edition.viewer_aware_objects(request.user)
+        .filter(
+            Q(shelves__user=request.user)
+            | Q(readthrough__user=request.user)
+            | Q(review__user=request.user)
+            | Q(comment__user=request.user)
+            | Q(quotation__user=request.user)
+        )
+        .distinct()
+    )
+
+    generator = csv_row_generator(data, request.user)
+
+    pseudo_buffer = Echo()
+    writer = csv.writer(pseudo_buffer)
+    # for testing, if you want to see the results in the browser:
+    # from django.http import JsonResponse
+    # return JsonResponse(list(generator), safe=False)
+    return StreamingHttpResponse(
+        (writer.writerow(row) for row in generator),
+        content_type="text/csv",
+        headers={"Content-Disposition": 'attachment; filename="bookwyrm-export.csv"'},
+    )
+
+
+def csv_row_generator(books, user):
+    """generate a csv entry for the user's book"""
+    deduplication_fields = [
+        f.name
+        for f in models.Edition._meta.get_fields()  # pylint: disable=protected-access
+        if getattr(f, "deduplication_field", False)
+    ]
+    fields = (
+        ["title", "author_text"]
+        + deduplication_fields
+        + ["rating", "review_name", "review_cw", "review_content"]
+    )
+    yield fields
+    for book in books:
+        # I think this is more efficient than doing a subquery in the view? but idk
+        review_rating = (
+            models.Review.objects.filter(user=user, book=book, rating__isnull=False)
+            .order_by("-published_date")
+            .first()
+        )
+
+        book.rating = review_rating.rating if review_rating else None
+
+        review = (
+            models.Review.objects.filter(user=user, book=book, content__isnull=False)
+            .order_by("-published_date")
+            .first()
+        )
+        if review:
+            book.review_name = review.name
+            book.review_cw = review.content_warning
+            book.review_content = review.raw_content
+        yield [getattr(book, field, "") or "" for field in fields]
+
+
+class Echo:
+    """An object that implements just the write method of the file-like
+    interface. (https://docs.djangoproject.com/en/3.2/howto/outputting-csv/)
+    """
+
+    # pylint: disable=no-self-use
+    def write(self, value):
+        """Write the value by returning it, instead of storing in a buffer."""
+        return value