User import and export improvements

bookwyrm/importers/bookwyrm_import.py

@@ -22,6 +22,21 @@ def process_import(
         job = BookwyrmImportJob.objects.create(
             user=user, archive_file=archive_file, required=required
         )
+
+        return job
+
+    def create_retry_job(
+        self, user: User, original_job: BookwyrmImportJob
+    ) -> BookwyrmImportJob:
+        """retry items that didn't import"""
+
+        job = BookwyrmImportJob.objects.create(
+            user=user,
+            archive_file=original_job.archive_file,
+            required=original_job.required,
+            retry=True,
+        )
+
         return job
 
 

bookwyrm/migrations/0210_userrelationshipimport_and_more.py

@@ -0,0 +1,144 @@
+# Generated by Django 4.2.15 on 2024-09-03 11:22
+
+import bookwyrm.models.fields
+import django.contrib.postgres.fields
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ("bookwyrm", "0209_user_show_ratings"),
+    ]
+
+    operations = [
+        migrations.CreateModel(
+            name="UserRelationshipImport",
+            fields=[
+                (
+                    "childjob_ptr",
+                    models.OneToOneField(
+                        auto_created=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        parent_link=True,
+                        primary_key=True,
+                        serialize=False,
+                        to="bookwyrm.childjob",
+                    ),
+                ),
+                (
+                    "relationship",
+                    bookwyrm.models.fields.CharField(
+                        choices=[("follow", "Follow"), ("block", "Block")],
+                        max_length=10,
+                        null=True,
+                    ),
+                ),
+                (
+                    "remote_id",
+                    bookwyrm.models.fields.RemoteIdField(
+                        max_length=255,
+                        null=True,
+                        validators=[bookwyrm.models.fields.validate_remote_id],
+                    ),
+                ),
+                ("fail_reason", models.TextField(null=True)),
+            ],
+            options={
+                "abstract": False,
+            },
+            bases=("bookwyrm.childjob",),
+        ),
+        migrations.RemoveField(
+            model_name="bookwyrmexportjob",
+            name="json_completed",
+        ),
+        migrations.AddField(
+            model_name="bookwyrmimportjob",
+            name="retry",
+            field=models.BooleanField(default=False),
+        ),
+        migrations.AlterField(
+            model_name="bookwyrmimportjob",
+            name="required",
+            field=django.contrib.postgres.fields.ArrayField(
+                base_field=bookwyrm.models.fields.CharField(blank=True, max_length=50),
+                blank=True,
+                size=None,
+            ),
+        ),
+        migrations.CreateModel(
+            name="UserImportPost",
+            fields=[
+                (
+                    "childjob_ptr",
+                    models.OneToOneField(
+                        auto_created=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        parent_link=True,
+                        primary_key=True,
+                        serialize=False,
+                        to="bookwyrm.childjob",
+                    ),
+                ),
+                ("json", models.JSONField()),
+                (
+                    "status_type",
+                    bookwyrm.models.fields.CharField(
+                        choices=[
+                            ("comment", "Comment"),
+                            ("review", "Review"),
+                            ("quote", "Quotation"),
+                        ],
+                        default="comment",
+                        max_length=10,
+                        null=True,
+                    ),
+                ),
+                ("fail_reason", models.TextField(null=True)),
+                (
+                    "book",
+                    bookwyrm.models.fields.ForeignKey(
+                        on_delete=django.db.models.deletion.PROTECT,
+                        to="bookwyrm.edition",
+                    ),
+                ),
+            ],
+            options={
+                "abstract": False,
+            },
+            bases=("bookwyrm.childjob",),
+        ),
+        migrations.CreateModel(
+            name="UserImportBook",
+            fields=[
+                (
+                    "childjob_ptr",
+                    models.OneToOneField(
+                        auto_created=True,
+                        on_delete=django.db.models.deletion.CASCADE,
+                        parent_link=True,
+                        primary_key=True,
+                        serialize=False,
+                        to="bookwyrm.childjob",
+                    ),
+                ),
+                ("book_data", models.JSONField()),
+                ("fail_reason", models.TextField(null=True)),
+                (
+                    "book",
+                    models.ForeignKey(
+                        blank=True,
+                        null=True,
+                        on_delete=django.db.models.deletion.SET_NULL,
+                        to="bookwyrm.book",
+                    ),
+                ),
+            ],
+            options={
+                "abstract": False,
+            },
+            bases=("bookwyrm.childjob",),
+        ),
+    ]

bookwyrm/models/__init__.py

@@ -26,7 +26,12 @@
 from .group import Group, GroupMember, GroupMemberInvitation
 
 from .import_job import ImportJob, ImportItem
-from .bookwyrm_import_job import BookwyrmImportJob
+from .bookwyrm_import_job import (
+    BookwyrmImportJob,
+    UserImportBook,
+    UserImportPost,
+    import_book_task,
+)
 from .bookwyrm_export_job import BookwyrmExportJob
 
 from .move import MoveUser

bookwyrm/models/bookwyrm_export_job.py

@@ -6,7 +6,7 @@
 from boto3.session import Session as BotoSession
 from s3_tar import S3Tar
 
-from django.db.models import BooleanField, FileField, JSONField
+from django.db.models import FileField, JSONField
 from django.core.serializers.json import DjangoJSONEncoder
 from django.core.files.base import ContentFile
 from django.core.files.storage import storages
@@ -17,7 +17,7 @@
 from bookwyrm.models import Review, Comment, Quotation
 from bookwyrm.models import Edition
 from bookwyrm.models import UserFollows, User, UserBlocks
-from bookwyrm.models.job import ParentJob
+from bookwyrm.models.job import ParentJob, ParentTask
 from bookwyrm.tasks import app, IMPORTS
 from bookwyrm.utils.tar import BookwyrmTarFile
 
@@ -42,38 +42,41 @@ class BookwyrmExportJob(ParentJob):
 
     export_data = FileField(null=True, storage=select_exports_storage)
     export_json = JSONField(null=True, encoder=DjangoJSONEncoder)
-    json_completed = BooleanField(default=False)
 
     def start_job(self):
         """schedule the first task"""
 
-        task = create_export_json_task.delay(job_id=self.id)
-        self.task_id = task.id
-        self.save(update_fields=["task_id"])
+        self.set_status("active")
+        create_export_json_task.delay(job_id=self.id)
 
 
-@app.task(queue=IMPORTS)
-def create_export_json_task(job_id):
+@app.task(queue=IMPORTS, base=ParentTask)
+def create_export_json_task(**kwargs):
     """create the JSON data for the export"""
 
-    job = BookwyrmExportJob.objects.get(id=job_id)
-
+    job = BookwyrmExportJob.objects.get(id=kwargs["job_id"])
     # don't start the job if it was stopped from the UI
-    if job.complete:
+    if job.status == "stopped":
         return
 
     try:
-        job.set_status("active")
-
-        # generate JSON structure
-        job.export_json = export_json(job.user)
+        # generate JSON
+        data = export_user(job.user)
+        data["settings"] = export_settings(job.user)
+        data["goals"] = export_goals(job.user)
+        data["books"] = export_books(job.user)
+        data["saved_lists"] = export_saved_lists(job.user)
+        data["follows"] = export_follows(job.user)
+        data["blocks"] = export_blocks(job.user)
+        job.export_json = data
         job.save(update_fields=["export_json"])
 
-        # create archive in separate task
+        # trigger task to create tar file
         create_archive_task.delay(job_id=job.id)
+
     except Exception as err:  # pylint: disable=broad-except
         logger.exception(
-            "create_export_json_task for %s failed with error: %s", job, err
+            "create_export_json_task for job %s failed with error: %s", job.id, err
         )
         job.set_status("failed")
 
@@ -94,21 +97,20 @@ def add_file_to_s3_tar(s3_tar: S3Tar, storage, file, directory=""):
     )
 
 
-@app.task(queue=IMPORTS)
-def create_archive_task(job_id):
+@app.task(queue=IMPORTS, base=ParentTask)
+def create_archive_task(**kwargs):
     """create the archive containing the JSON file and additional files"""
 
-    job = BookwyrmExportJob.objects.get(id=job_id)
+    job = BookwyrmExportJob.objects.get(id=kwargs["job_id"])
 
     # don't start the job if it was stopped from the UI
-    if job.complete:
+    if job.status == "stopped":
         return
 
     try:
         export_task_id = str(job.task_id)
         archive_filename = f"{export_task_id}.tar.gz"
         export_json_bytes = DjangoJSONEncoder().encode(job.export_json).encode("utf-8")
-
         user = job.user
         editions = get_books_for_user(user)
 
@@ -169,25 +171,15 @@ def create_archive_task(job_id):
                             tar.add_image(edition.cover, directory="images")
             job.save(update_fields=["export_data"])
 
-        job.set_status("completed")
+        job.complete_job()
 
     except Exception as err:  # pylint: disable=broad-except
-        logger.exception("create_archive_task for %s failed with error: %s", job, err)
+        logger.exception(
+            "create_archive_task for job %s failed with error: %s", job.id, err
+        )
         job.set_status("failed")
 
 
-def export_json(user: User):
-    """create export JSON"""
-    data = export_user(user)  # in the root of the JSON structure
-    data["settings"] = export_settings(user)
-    data["goals"] = export_goals(user)
-    data["books"] = export_books(user)
-    data["saved_lists"] = export_saved_lists(user)
-    data["follows"] = export_follows(user)
-    data["blocks"] = export_blocks(user)
-    return data
-
-
 def export_user(user: User):
     """export user data"""
     data = user.to_activity()
@@ -316,11 +308,9 @@ def export_book(user: User, edition: Edition):
 def get_books_for_user(user):
     """
     Get all the books and editions related to a user.
-
     We use union() instead of Q objects because it creates
-    multiple simple queries in stead of a much more complex DB query
+    multiple simple queries instead of a complex DB query
     that can time out.
-
     """
 
     shelf_eds = Edition.objects.select_related("parent_work").filter(shelves__user=user)