Hatchway API Rewrite (#499)

Removes django-ninja and replaces it with a new API framework, "hatchway".

I plan to move hatchway into its own project very soon.

activities/models/post_interaction.py

@@ -188,7 +188,7 @@ class PostInteraction(StatorModel):
         return result
 
     @classmethod
-    def get_event_interactions(cls, events, identity):
+    def get_event_interactions(cls, events, identity) -> dict[str, set[str]]:
         """
         Returns a dict of {interaction_type: set(post_ids)} for all the posts
         within the events and the given identity, for use in templates.

api/decorators.py

@@ -16,4 +16,7 @@ def identity_required(function):
             return JsonResponse({"error": "identity_token_required"}, status=400)
         return function(request, *args, **kwargs)
 
+    # This is for the API only
+    inner.csrf_exempt = True
+
     return inner

api/pagination.py

@@ -1,22 +1,98 @@
 import dataclasses
 import urllib.parse
 from collections.abc import Callable
-from typing import Any
+from typing import Any, Generic, Protocol, TypeVar
 
 from django.db import models
 from django.http import HttpRequest
 
 from activities.models import PostInteraction
+from hatchway.http import ApiResponse
+
+T = TypeVar("T")
+
+
+class SchemaWithId(Protocol):
+    """
+    Little protocol type to represent schemas that have an ID attribute
+    """
+
+    id: str
+
+
+TI = TypeVar("TI", bound=SchemaWithId)
+TM = TypeVar("TM", bound=models.Model)
+
+
+class PaginatingApiResponse(ApiResponse[list[TI]]):
+    """
+    An ApiResponse subclass that also handles pagination link headers
+    """
+
+    def __init__(
+        self,
+        data: list[TI],
+        request: HttpRequest,
+        include_params: list[str],
+        **kwargs,
+    ):
+        # Call superclass
+        super().__init__(data, **kwargs)
+        # Figure out if we need link headers
+        self._request = request
+        self.extra_params = self.filter_params(self._request, include_params)
+        link_header = self.build_link_header()
+        if link_header:
+            self.headers["link"] = link_header
+
+    @staticmethod
+    def filter_params(request: HttpRequest, allowed_params: list[str]):
+        params = {}
+        for key in allowed_params:
+            value = request.GET.get(key, None)
+            if value:
+                params[key] = value
+        return params
+
+    def get_part(self, data_index: int, param_name: str, rel: str) -> str | None:
+        """
+        Used to get next/prev URLs
+        """
+        if not self.data:
+            return None
+        # Use the ID of the last object for the next page start
+        params = dict(self.extra_params)
+        params[param_name] = self.data[data_index].id
+        return (
+            "<"
+            + self._request.build_absolute_uri(self._request.path)
+            + "?"
+            + urllib.parse.urlencode(params)
+            + f'>; rel="{rel}"'
+        )
+
+    def build_link_header(self):
+        parts = [
+            entry
+            for entry in [
+                self.get_part(0, "min_id", "prev"),
+                self.get_part(-1, "max_id", "next"),
+            ]
+            if entry
+        ]
+        if not parts:
+            return None
+        return ", ".join(parts)
 
 
 @dataclasses.dataclass
-class PaginationResult:
+class PaginationResult(Generic[T]):
     """
     Represents a pagination result for Mastodon (it does Link header stuff)
     """
 
     #: A list of objects that matched the pagination query.
-    results: list[models.Model]
+    results: list[T]
 
     #: The actual applied limit, which may be different from what was requested.
     limit: int
@@ -130,12 +206,12 @@ class MastodonPaginator:
 
     def paginate(
         self,
-        queryset,
+        queryset: models.QuerySet[TM],
         min_id: str | None,
         max_id: str | None,
         since_id: str | None,
         limit: int | None,
-    ) -> PaginationResult:
+    ) -> PaginationResult[TM]:
         # These "does not start with interaction" checks can be removed after a
         # couple months, when clients have flushed them out.
         if max_id and not max_id.startswith("interaction"):

api/parser.py

@@ -1,22 +0,0 @@
-import json
-
-from ninja.parser import Parser
-
-
-class FormOrJsonParser(Parser):
-    """
-    If there's form data in a request, makes it into a JSON dict.
-    This is needed as the Mastodon API allows form data OR json body as input.
-    """
-
-    def parse_body(self, request):
-        # Did they submit JSON?
-        if request.content_type == "application/json" and request.body.strip():
-            return json.loads(request.body)
-        # Fall back to form data
-        value = {}
-        for key, item in request.POST.items():
-            value[key] = item
-        for key, item in request.GET.items():
-            value[key] = item
-        return value

api/schemas.py

@@ -1,6 +1,9 @@
 from typing import Literal, Optional, Union
 
-from ninja import Field, Schema
+from activities import models as activities_models
+from hatchway import Field, Schema
+from users import models as users_models
+from users.services import IdentityService
 
 
 class Application(Schema):
@@ -19,6 +22,10 @@ class CustomEmoji(Schema):
     visible_in_picker: bool
     category: str
 
+    @classmethod
+    def from_emoji(cls, emoji: activities_models.Emoji) -> "CustomEmoji":
+        return cls(**emoji.to_mastodon_json())
+
 
 class AccountField(Schema):
     name: str
@@ -53,6 +60,17 @@ class Account(Schema):
     following_count: int
     source: dict | None
 
+    @classmethod
+    def from_identity(
+        cls,
+        identity: users_models.Identity,
+        include_counts: bool = True,
+        source=False,
+    ) -> "Account":
+        return cls(
+            **identity.to_mastodon_json(include_counts=include_counts, source=source)
+        )
+
 
 class MediaAttachment(Schema):
     id: str
@@ -64,6 +82,12 @@ class MediaAttachment(Schema):
     description: str | None
     blurhash: str | None
 
+    @classmethod
+    def from_post_attachment(
+        cls, attachment: activities_models.PostAttachment
+    ) -> "MediaAttachment":
+        return cls(**attachment.to_mastodon_json())
+
 
 class StatusMention(Schema):
     id: str
@@ -108,6 +132,47 @@ class Status(Schema):
     bookmarked: bool | None
     pinned: bool | None
 
+    @classmethod
+    def from_post(
+        cls,
+        post: activities_models.Post,
+        interactions: dict[str, set[str]] | None = None,
+    ) -> "Status":
+        return cls(**post.to_mastodon_json(interactions=interactions))
+
+    @classmethod
+    def map_from_post(
+        cls,
+        posts: list[activities_models.Post],
+        identity: users_models.Identity,
+    ) -> list["Status"]:
+        interactions = activities_models.PostInteraction.get_post_interactions(
+            posts, identity
+        )
+        return [cls.from_post(post, interactions=interactions) for post in posts]
+
+    @classmethod
+    def from_timeline_event(
+        cls,
+        timeline_event: activities_models.TimelineEvent,
+        interactions: dict[str, set[str]] | None = None,
+    ) -> "Status":
+        return cls(**timeline_event.to_mastodon_status_json(interactions=interactions))
+
+    @classmethod
+    def map_from_timeline_event(
+        cls,
+        events: list[activities_models.TimelineEvent],
+        identity: users_models.Identity,
+    ) -> list["Status"]:
+        interactions = activities_models.PostInteraction.get_event_interactions(
+            events, identity
+        )
+        return [
+            cls.from_timeline_event(event, interactions=interactions)
+            for event in events
+        ]
+
 
 class Conversation(Schema):
     id: str
@@ -134,12 +199,26 @@ class Notification(Schema):
     account: Account
     status: Status | None
 
+    @classmethod
+    def from_timeline_event(
+        cls,
+        event: activities_models.TimelineEvent,
+    ) -> "Notification":
+        return cls(**event.to_mastodon_notification_json())
+
 
 class Tag(Schema):
     name: str
     url: str
     history: dict
 
+    @classmethod
+    def from_hashtag(
+        cls,
+        hashtag: activities_models.Hashtag,
+    ) -> "Tag":
+        return cls(**hashtag.to_mastodon_json())
+
 
 class Search(Schema):
     accounts: list[Account]
@@ -162,6 +241,16 @@ class Relationship(Schema):
     endorsed: bool
     note: str
 
+    @classmethod
+    def from_identity_pair(
+        cls,
+        identity: users_models.Identity,
+        from_identity: users_models.Identity,
+    ) -> "Relationship":
+        return cls(
+            **IdentityService(identity).mastodon_json_relationship(from_identity)
+        )
+
 
 class Context(Schema):
     ancestors: list[Status]
@@ -187,3 +276,11 @@ class Announcement(Schema):
     tags: list[Tag]
     emojis: list[CustomEmoji]
     reactions: list
+
+    @classmethod
+    def from_announcement(
+        cls,
+        announcement: users_models.Announcement,
+        user: users_models.User,
+    ) -> "Announcement":
+        return cls(**announcement.to_mastodon_json(user=user))

api/urls.py

@@ -0,0 +1,72 @@
+from django.urls import path
+
+from api.views import (
+    accounts,
+    announcements,
+    apps,
+    emoji,
+    filters,
+    instance,
+    media,
+    notifications,
+    search,
+    statuses,
+    timelines,
+)
+from hatchway import methods
+
+urlpatterns = [
+    # Accounts
+    path("v1/accounts/verify_credentials", accounts.verify_credentials),
+    path("v1/accounts/update_credentials", accounts.update_credentials),
+    path("v1/accounts/relationships", accounts.account_relationships),
+    path("v1/accounts/familiar_followers", accounts.familiar_followers),
+    path("v1/accounts/search", accounts.accounts_search),
+    path("v1/accounts/lookup", accounts.lookup),
+    path("v1/accounts/<id>", accounts.account),
+    path("v1/accounts/<id>/statuses", accounts.account_statuses),
+    path("v1/accounts/<id>/follow", accounts.account_follow),
+    path("v1/accounts/<id>/unfollow", accounts.account_unfollow),
+    path("v1/accounts/<id>/block", accounts.account_block),
+    path("v1/accounts/<id>/unblock", accounts.account_unblock),
+    path("v1/accounts/<id>/mute", accounts.account_mute),
+    path("v1/accounts/<id>/unmute", accounts.account_unmute),
+    path("v1/accounts/<id>/following", accounts.account_following),
+    path("v1/accounts/<id>/followers", accounts.account_followers),
+    # Announcements
+    path("v1/announcements", announcements.announcement_list),
+    path("v1/announcements/<pk>/dismiss", announcements.announcement_dismiss),
+    # Apps
+    path("v1/apps", apps.add_app),
+    # Filters
+    path("v2/filters", filters.list_filters),
+    path("v1/filters", filters.list_filters),
+    # Instance
+    path("v1/instance", instance.instance_info_v1),
+    path("v2/instance", instance.instance_info_v2),
+    path("v1/timelines/home", timelines.home),
+    path("v1/timelines/public", timelines.public),
+    path("v1/timelines/tag/<hashtag>", timelines.hashtag),
+    path("v1/timelines/conversations", timelines.conversations),
+    path("v1/favourites", timelines.favourites),
+    path("v1/notifications", notifications.notifications),
+    path("v1/statuses", statuses.post_status),
+    path("v1/media", media.upload_media),
+    path("v2/media", media.upload_media),
+    path("v1/media/<id>", methods(get=media.get_media, put=media.update_media)),
+    path(
+        "v1/statuses/<id>",
+        methods(
+            get=statuses.status,
+            delete=statuses.delete_status,
+        ),
+    ),
+    path("v1/statuses/<id>/context", statuses.status_context),
+    path("v1/statuses/<id>/favourite", statuses.favourite_status),
+    path("v1/statuses/<id>/unfavourite", statuses.unfavourite_status),
+    path("v1/statuses/<id>/favourited_by", statuses.favourited_by),
+    path("v1/statuses/<id>/reblog", statuses.reblog_status),
+    path("v1/statuses/<id>/unreblog", statuses.unreblog_status),
+    path("v1/custom_emojis", emoji.emojis),
+    path("v2/search", search.search),
+]

api/views/__init__.py

@@ -1,12 +0,0 @@
-from .accounts import *  # noqa
-from .announcements import *  # noqa
-from .apps import *  # noqa
-from .emoji import *  # noqa
-from .filters import *  # noqa
-from .instance import *  # noqa
-from .media import *  # noqa
-from .notifications import *  # noqa
-from .oauth import *  # noqa
-from .search import *  # noqa
-from .statuses import *  # noqa
-from .timelines import *  # noqa

api/views/accounts.py

@@ -1,50 +1,49 @@
-from django.http import HttpRequest, HttpResponse, QueryDict
+from typing import Any
-from django.http.multipartparser import MultiPartParser
+
+from django.core.files import File
+from django.http import HttpRequest
 from django.shortcuts import get_object_or_404
-from ninja import Field, Schema
 
 from activities.models import Post
 from activities.services import SearchService
 from api import schemas
 from api.decorators import identity_required
-from api.pagination import MastodonPaginator
+from api.pagination import MastodonPaginator, PaginatingApiResponse, PaginationResult
-from api.views.base import api_router
 from core.models import Config
+from hatchway import ApiResponse, QueryOrBody, api_view
 from users.models import Identity
 from users.services import IdentityService
 from users.shortcuts import by_handle_or_404
 
 
-@api_router.get("/v1/accounts/verify_credentials", response=schemas.Account)
 @identity_required
-def verify_credentials(request):
+@api_view.get
-    return request.identity.to_mastodon_json(source=True)
+def verify_credentials(request) -> schemas.Account:
+    return schemas.Account.from_identity(request.identity, source=True)
 
 
-@api_router.patch("/v1/accounts/update_credentials", response=schemas.Account)
 @identity_required
+@api_view.patch
 def update_credentials(
     request,
-):
+    display_name: QueryOrBody[str | None] = None,
-    # Django won't load POST and FILES for patch methods, so we do it.
+    note: QueryOrBody[str | None] = None,
-    if request.content_type == "multipart/form-data":
+    discoverable: QueryOrBody[bool | None] = None,
-        POST, FILES = MultiPartParser(
+    source: QueryOrBody[dict[str, Any] | None] = None,
-            request.META, request, request.upload_handlers, request.encoding
+    fields_attributes: QueryOrBody[dict[str, dict[str, str]] | None] = None,
-        ).parse()
+    avatar: File | None = None,
-    elif request.content_type == "application/x-www-form-urlencoded":
+    header: File | None = None,
-        POST = QueryDict(request.body, encoding=request._encoding)
+) -> schemas.Account:
-        FILES = {}
-    else:
-        return HttpResponse(status=400)
     identity = request.identity
     service = IdentityService(identity)
-    if "display_name" in POST:
+    if display_name is not None:
-        identity.name = POST["display_name"]
+        identity.name = display_name
-    if "note" in POST:
+    if note is not None:
-        service.set_summary(POST["note"])
+        service.set_summary(note)
-    if "discoverable" in POST:
+    if discoverable is not None:
-        identity.discoverable = POST["discoverable"] == "checked"
+        identity.discoverable = discoverable
-    if "source[privacy]" in POST:
+    if source:
+        if "privacy" in source:
             privacy_map = {
                 "public": Post.Visibilities.public,
                 "unlisted": Post.Visibilities.unlisted,
@@ -54,78 +53,77 @@ def update_credentials(
             Config.set_identity(
                 identity,
                 "default_post_visibility",
-            privacy_map[POST["source[privacy]"]],
+                privacy_map[source["privacy"]],
             )
-    if "fields_attributes[0][name]" in POST:
+    if fields_attributes:
         identity.metadata = []
-        for i in range(4):
+        for attribute in fields_attributes.values():
-            name_name = f"fields_attributes[{i}][name]"
+            attr_name = attribute.get("name", None)
-            value_name = f"fields_attributes[{i}][value]"
+            attr_value = attribute.get("value", None)
-            if name_name and value_name in POST:
+            if attr_name:
                 # Empty value means delete this item
-                if not POST[value_name]:
+                if not attr_value:
                     break
-                identity.metadata.append(
+                identity.metadata.append({"name": attr_name, "value": attr_value})
-                    {"name": POST[name_name], "value": POST[value_name]}
+    if avatar:
-                )
+        service.set_icon(avatar)
-    if "avatar" in FILES:
+    if header:
-        service.set_icon(FILES["avatar"])
+        service.set_image(header)
-    if "header" in FILES:
-        service.set_image(FILES["header"])
     identity.save()
-    return identity.to_mastodon_json(source=True)
+    return schemas.Account.from_identity(identity, source=True)
 
 
-@api_router.get("/v1/accounts/relationships", response=list[schemas.Relationship])
 @identity_required
-def account_relationships(request):
+@api_view.get
-    ids = request.GET.getlist("id[]")
+def account_relationships(request, id: list[str] | None) -> list[schemas.Relationship]:
     result = []
-    for id in ids:
+    # ID is actually a list. Thanks Mastodon!
-        identity = get_object_or_404(Identity, pk=id)
+    ids = id or []
+    for actual_id in ids:
+        identity = get_object_or_404(Identity, pk=actual_id)
         result.append(
             IdentityService(identity).mastodon_json_relationship(request.identity)
         )
     return result
 
 
-@api_router.get(
-    "/v1/accounts/familiar_followers", response=list[schemas.FamiliarFollowers]
-)
 @identity_required
-def familiar_followers(request):
+@api_view.get
+def familiar_followers(
+    request, id: list[str] | None
+) -> list[schemas.FamiliarFollowers]:
     """
     Returns people you follow that also follow given account IDs
     """
-    ids = request.GET.getlist("id[]")
+    ids = id or []
     result = []
-    for id in ids:
+    for actual_id in ids:
-        target_identity = get_object_or_404(Identity, pk=id)
+        target_identity = get_object_or_404(Identity, pk=actual_id)
         result.append(
-            {
+            schemas.FamiliarFollowers(
-                "id": id,
+                id=actual_id,
-                "accounts": [
+                accounts=[
-                    identity.to_mastodon_json()
+                    schemas.Account.from_identity(identity)
                     for identity in Identity.objects.filter(
                         inbound_follows__source=request.identity,
                         outbound_follows__target=target_identity,
                     )[:20]
                 ],
-            }
+            )
         )
     return result
 
 
-@api_router.get("/v1/accounts/search", response=list[schemas.Account])
 @identity_required
-def search(
+@api_view.get
+def accounts_search(
     request,
     q: str,
-    fetch_identities: bool = Field(False, alias="resolve"),
+    resolve: bool = False,
     following: bool = False,
     limit: int = 20,
     offset: int = 0,
-):
+) -> list[schemas.Account]:
     """
     Handles searching for accounts by username or handle
     """
@@ -135,33 +133,33 @@ def search(
         return []
     searcher = SearchService(q, request.identity)
     search_result = searcher.search_identities_handle()
-    return [i.to_mastodon_json() for i in search_result]
+    return [schemas.Account.from_identity(i) for i in search_result]
 
 
-@api_router.get("/v1/accounts/lookup", response=schemas.Account)
+@api_view.get
-def lookup(request: HttpRequest, acct: str):
+def lookup(request: HttpRequest, acct: str) -> schemas.Account:
     """
     Quickly lookup a username to see if it is available, skipping WebFinger
     resolution.
     """
     identity = by_handle_or_404(request, handle=acct, local=False)
-    return identity.to_mastodon_json()
+    return schemas.Account.from_identity(identity)
 
 
-@api_router.get("/v1/accounts/{id}", response=schemas.Account)
+@api_view.get
 @identity_required
-def account(request, id: str):
+def account(request, id: str) -> schemas.Account:
     identity = get_object_or_404(
-        Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
+        Identity.objects.exclude(restriction=Identity.Restriction.blocked),
+        pk=id,
     )
-    return identity.to_mastodon_json()
+    return schemas.Account.from_identity(identity)
 
 
-@api_router.get("/v1/accounts/{id}/statuses", response=list[schemas.Status])
+@api_view.get
 @identity_required
 def account_statuses(
     request: HttpRequest,
-    response: HttpResponse,
     id: str,
     exclude_reblogs: bool = False,
     exclude_replies: bool = False,
@@ -172,7 +170,7 @@ def account_statuses(
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Status]]:
     identity = get_object_or_404(
         Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
     )
@@ -191,27 +189,24 @@ def account_statuses(
         .order_by("-created")
     )
     if pinned:
-        return []
+        return ApiResponse([])
     if only_media:
         queryset = queryset.filter(attachments__pk__isnull=False)
     if tagged:
         queryset = queryset.tagged_with(tagged)
     # Get user posts with pagination
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[Post] = paginator.paginate(
         queryset,
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    # Convert those to the JSON form
+    return PaginatingApiResponse(
-    pager.jsonify_posts(identity=request.identity)
+        schemas.Status.map_from_post(pager.results, request.identity),
-    # Add a link header if we need to
+        request=request,
-    if pager.results:
+        include_params=[
-        response.headers["Link"] = pager.link_header(
-            request,
-            [
             "limit",
             "id",
             "exclude_reblogs",
@@ -221,147 +216,136 @@ def account_statuses(
             "tagged",
         ],
     )
-    return pager.json_results
 
 
-@api_router.post("/v1/accounts/{id}/follow", response=schemas.Relationship)
+@api_view.post
 @identity_required
-def account_follow(request, id: str, reblogs: bool = True):
+def account_follow(request, id: str, reblogs: bool = True) -> schemas.Relationship:
     identity = get_object_or_404(
         Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
     )
     service = IdentityService(identity)
     service.follow_from(request.identity, boosts=reblogs)
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-@api_router.post("/v1/accounts/{id}/unfollow", response=schemas.Relationship)
+@api_view.post
 @identity_required
-def account_unfollow(request, id: str):
+def account_unfollow(request, id: str) -> schemas.Relationship:
     identity = get_object_or_404(
         Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
     )
     service = IdentityService(identity)
     service.unfollow_from(request.identity)
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-@api_router.post("/v1/accounts/{id}/block", response=schemas.Relationship)
+@api_view.post
 @identity_required
-def account_block(request, id: str):
+def account_block(request, id: str) -> schemas.Relationship:
     identity = get_object_or_404(Identity, pk=id)
     service = IdentityService(identity)
     service.block_from(request.identity)
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-@api_router.post("/v1/accounts/{id}/unblock", response=schemas.Relationship)
+@api_view.post
 @identity_required
-def account_unblock(request, id: str):
+def account_unblock(request, id: str) -> schemas.Relationship:
     identity = get_object_or_404(Identity, pk=id)
     service = IdentityService(identity)
     service.unblock_from(request.identity)
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-class MuteDetailsSchema(Schema):
-    notifications: bool = True
-    duration: int = 0
-
-
-@api_router.post("/v1/accounts/{id}/mute", response=schemas.Relationship)
 @identity_required
-def account_mute(request, id: str, details: MuteDetailsSchema):
+@api_view.post
+def account_mute(
+    request,
+    id: str,
+    notifications: QueryOrBody[bool] = True,
+    duration: QueryOrBody[int] = 0,
+) -> schemas.Relationship:
     identity = get_object_or_404(Identity, pk=id)
     service = IdentityService(identity)
     service.mute_from(
         request.identity,
-        duration=details.duration,
+        duration=duration,
-        include_notifications=details.notifications,
+        include_notifications=notifications,
     )
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-@api_router.post("/v1/accounts/{id}/unmute", response=schemas.Relationship)
 @identity_required
-def account_unmute(request, id: str):
+@api_view.post
+def account_unmute(request, id: str) -> schemas.Relationship:
     identity = get_object_or_404(Identity, pk=id)
     service = IdentityService(identity)
     service.unmute_from(request.identity)
-    return service.mastodon_json_relationship(request.identity)
+    return schemas.Relationship.from_identity_pair(identity, request.identity)
 
 
-@api_router.get("/v1/accounts/{id}/following", response=list[schemas.Account])
+@api_view.get
 def account_following(
     request: HttpRequest,
-    response: HttpResponse,
     id: str,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 40,
-):
+) -> ApiResponse[list[schemas.Account]]:
     identity = get_object_or_404(
         Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
     )
 
     if not identity.config_identity.visible_follows and request.identity != identity:
-        return []
+        return ApiResponse([])
 
     service = IdentityService(identity)
 
     paginator = MastodonPaginator(max_limit=80)
-    pager = paginator.paginate(
+    pager: PaginationResult[Identity] = paginator.paginate(
         service.following(),
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    pager.jsonify_identities()
+    return PaginatingApiResponse(
-
+        [schemas.Account.from_identity(i) for i in pager.results],
-    if pager.results:
+        request=request,
-        response.headers["Link"] = pager.link_header(
+        include_params=["limit"],
-            request,
-            ["limit"],
     )
 
-    return pager.json_results
 
-
+@api_view.get
-@api_router.get("/v1/accounts/{id}/followers", response=list[schemas.Account])
 def account_followers(
     request: HttpRequest,
-    response: HttpResponse,
     id: str,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 40,
-):
+) -> ApiResponse[list[schemas.Account]]:
     identity = get_object_or_404(
         Identity.objects.exclude(restriction=Identity.Restriction.blocked), pk=id
     )
 
     if not identity.config_identity.visible_follows and request.identity != identity:
-        return []
+        return ApiResponse([])
 
     service = IdentityService(identity)
 
     paginator = MastodonPaginator(max_limit=80)
-    pager = paginator.paginate(
+    pager: PaginationResult[Identity] = paginator.paginate(
         service.followers(),
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    pager.jsonify_identities()
+    return PaginatingApiResponse(
-
+        [schemas.Account.from_identity(i) for i in pager.results],
-    if pager.results:
+        request=request,
-        response.headers["Link"] = pager.link_header(
+        include_params=["limit"],
-            request,
-            ["limit"],
     )
-
-    return pager.json_results

api/views/announcements.py

@@ -2,22 +2,22 @@ from django.shortcuts import get_object_or_404
 
 from api import schemas
 from api.decorators import identity_required
-from api.views.base import api_router
+from hatchway import api_view
 from users.models import Announcement
 from users.services import AnnouncementService
 
 
-@api_router.get("/v1/announcements", response=list[schemas.Announcement])
 @identity_required
-def announcement_list(request):
+@api_view.get
+def announcement_list(request) -> list[schemas.Announcement]:
     return [
-        a.to_mastodon_json(request.user)
+        schemas.Announcement.from_announcement(a, request.user)
         for a in AnnouncementService(request.user).visible()
     ]
 
 
-@api_router.post("/v1/announcements/{pk}/dismiss")
 @identity_required
+@api_view.post
 def announcement_dismiss(request, pk: str):
     announcement = get_object_or_404(Announcement, pk=pk)
     AnnouncementService(request.user).mark_seen(announcement)

api/views/apps.py

@@ -1,10 +1,9 @@
 import secrets
 
-from ninja import Schema
+from hatchway import Schema, api_view
 
 from .. import schemas
 from ..models import Application
-from .base import api_router
 
 
 class CreateApplicationSchema(Schema):
@@ -14,8 +13,8 @@ class CreateApplicationSchema(Schema):
     website: None | str = None
 
 
-@api_router.post("/v1/apps", response=schemas.Application)
+@api_view.post
-def add_app(request, details: CreateApplicationSchema):
+def add_app(request, details: CreateApplicationSchema) -> schemas.Application:
     client_id = "tk-" + secrets.token_urlsafe(16)
     client_secret = secrets.token_urlsafe(40)
     application = Application.objects.create(
@@ -26,4 +25,4 @@ def add_app(request, details: CreateApplicationSchema):
         redirect_uris=details.redirect_uris,
         scopes=details.scopes or "read",
     )
-    return application
+    return schemas.Application.from_orm(application)

api/views/base.py

@@ -1,5 +0,0 @@
-from ninja import NinjaAPI
-
-from api.parser import FormOrJsonParser
-
-api_router = NinjaAPI(parser=FormOrJsonParser())

api/views/emoji.py

@@ -1,8 +1,10 @@
 from activities.models import Emoji
 from api.schemas import CustomEmoji
-from api.views.base import api_router
+from hatchway import api_view
 
 
-@api_router.get("/v1/custom_emojis", response=list[CustomEmoji])
+@api_view.get
-def emojis(request):
+def emojis(request) -> list[CustomEmoji]:
-    return [e.to_mastodon_json() for e in Emoji.objects.usable().filter(local=True)]
+    return [
+        CustomEmoji.from_emoji(e) for e in Emoji.objects.usable().filter(local=True)
+    ]

api/views/filters.py

@@ -1,8 +1,8 @@
 from api.decorators import identity_required
-from api.views.base import api_router
+from hatchway import api_view
 
 
-@api_router.get("/v1/filters")
 @identity_required
-def status(request):
+@api_view.get
+def list_filters(request):
     return []

api/views/instance.py

@@ -1,15 +1,15 @@
 from django.conf import settings
 
 from activities.models import Post
+from api import schemas
 from core.models import Config
+from hatchway import api_view
 from takahe import __version__
 from users.models import Domain, Identity
 
-from .base import api_router
 
-
+@api_view.get
-@api_router.get("/v1/instance")
+def instance_info_v1(request):
-def instance_info(request):
     return {
         "uri": request.headers.get("host", settings.SETUP.MAIN_DOMAIN),
         "title": Config.system.site_name,
@@ -53,11 +53,17 @@ def instance_info(request):
     }
 
 
-@api_router.get("/v2/instance")
+@api_view.get
-def instance_info_v2(request):
+def instance_info_v2(request) -> dict:
     current_domain = Domain.get_domain(
         request.headers.get("host", settings.SETUP.MAIN_DOMAIN)
     )
+    if current_domain is None or not current_domain.local:
+        current_domain = Domain.get_domain(
+            request.headers.get(settings.SETUP.MAIN_DOMAIN)
+        )
+    if current_domain is None:
+        raise ValueError("No domain set up for MAIN_DOMAIN")
     admin_identity = (
         Identity.objects.filter(users__admin=True).order_by("created").first()
     )
@@ -116,7 +122,7 @@ def instance_info_v2(request):
         },
         "contact": {
             "email": "",
-            "account": admin_identity.to_mastodon_json(),
+            "account": schemas.Account.from_identity(admin_identity),
         },
         "rules": [],
     }

api/views/media.py

@@ -1,28 +1,22 @@
+from django.core.files import File
 from django.shortcuts import get_object_or_404
-from ninja import File, Schema
-from ninja.files import UploadedFile
 
 from activities.models import PostAttachment, PostAttachmentStates
 from api import schemas
-from api.views.base import api_router
 from core.files import blurhash_image, resize_image
+from hatchway import QueryOrBody, api_view
 
 from ..decorators import identity_required
 
 
-class UploadMediaSchema(Schema):
-    description: str = ""
-    focus: str = "0,0"
-
-
-@api_router.post("/v1/media", response=schemas.MediaAttachment)
-@api_router.post("/v2/media", response=schemas.MediaAttachment)
 @identity_required
+@api_view.post
 def upload_media(
     request,
-    file: UploadedFile = File(...),
+    file: File,
-    details: UploadMediaSchema | None = None,
+    description: QueryOrBody[str] = "",
-):
+    focus: QueryOrBody[str] = "0,0",
+) -> schemas.MediaAttachment:
     main_file = resize_image(
         file,
         size=(2000, 2000),
@@ -38,7 +32,7 @@ def upload_media(
         mimetype="image/webp",
         width=main_file.image.width,
         height=main_file.image.height,
-        name=details.description if details else None,
+        name=description or None,
         state=PostAttachmentStates.fetched,
     )
     attachment.file.save(
@@ -50,27 +44,28 @@ def upload_media(
         thumbnail_file,
     )
     attachment.save()
-    return attachment.to_mastodon_json()
+    return schemas.MediaAttachment.from_post_attachment(attachment)
 
 
-@api_router.get("/v1/media/{id}", response=schemas.MediaAttachment)
 @identity_required
+@api_view.get
 def get_media(
     request,
     id: str,
-):
+) -> schemas.MediaAttachment:
     attachment = get_object_or_404(PostAttachment, pk=id)
-    return attachment.to_mastodon_json()
+    return schemas.MediaAttachment.from_post_attachment(attachment)
 
 
-@api_router.put("/v1/media/{id}", response=schemas.MediaAttachment)
 @identity_required
+@api_view.put
 def update_media(
     request,
     id: str,
-    details: UploadMediaSchema | None = None,
+    description: QueryOrBody[str] = "",
-):
+    focus: QueryOrBody[str] = "0,0",
+) -> schemas.MediaAttachment:
     attachment = get_object_or_404(PostAttachment, pk=id)
-    attachment.name = details.description if details else None
+    attachment.name = description or None
     attachment.save()
-    return attachment.to_mastodon_json()
+    return schemas.MediaAttachment.from_post_attachment(attachment)

api/views/notifications.py

@@ -1,24 +1,23 @@
-from django.http import HttpRequest, HttpResponse
+from django.http import HttpRequest
 
 from activities.models import TimelineEvent
 from activities.services import TimelineService
 from api import schemas
 from api.decorators import identity_required
-from api.pagination import MastodonPaginator
+from api.pagination import MastodonPaginator, PaginatingApiResponse, PaginationResult
-from api.views.base import api_router
+from hatchway import ApiResponse, api_view
 
 
-@api_router.get("/v1/notifications", response=list[schemas.Notification])
 @identity_required
+@api_view.get
 def notifications(
     request: HttpRequest,
-    response: HttpResponse,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
     account_id: str | None = None,
-):
+) -> ApiResponse[list[schemas.Notification]]:
     # Types/exclude_types use weird syntax so we have to handle them manually
     base_types = {
         "favourite": TimelineEvent.Types.liked,
@@ -37,16 +36,15 @@ def notifications(
         [base_types[r] for r in requested_types if r in base_types]
     )
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[TimelineEvent] = paginator.paginate(
         queryset,
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    pager.jsonify_notification_events(identity=request.identity)
+    return PaginatingApiResponse(
-
+        [schemas.Notification.from_timeline_event(event) for event in pager.results],
-    if pager.results:
+        request=request,
-        response.headers["Link"] = pager.link_header(request, ["limit", "account_id"])
+        include_params=["limit", "account_id"],
-
+    )
-    return pager.json_results

api/views/oauth.py

@@ -1,4 +1,5 @@
 import base64
+import json
 import secrets
 from urllib.parse import urlparse, urlunparse
 
@@ -16,7 +17,6 @@ from django.views.decorators.csrf import csrf_exempt
 from django.views.generic import View
 
 from api.models import Application, Authorization, Token
-from api.parser import FormOrJsonParser
 
 
 class OauthRedirect(HttpResponseRedirect):
@@ -40,6 +40,19 @@ class OauthRedirect(HttpResponseRedirect):
         super().__init__(urlunparse(url_parts))
 
 
+def get_json_and_formdata(request):
+    # Did they submit JSON?
+    if request.content_type == "application/json" and request.body.strip():
+        return json.loads(request.body)
+    # Fall back to form data
+    value = {}
+    for key, item in request.POST.items():
+        value[key] = item
+    for key, item in request.GET.items():
+        value[key] = item
+    return value
+
+
 class AuthorizationView(LoginRequiredMixin, View):
     """
     Asks the user to authorize access.
@@ -87,7 +100,7 @@ class AuthorizationView(LoginRequiredMixin, View):
         return render(request, "api/oauth_authorize.html", context)
 
     def post(self, request):
-        post_data = FormOrJsonParser().parse_body(request)
+        post_data = get_json_and_formdata(request)
         # Grab the application and other details again
         redirect_uri = post_data["redirect_uri"]
         scope = post_data["scope"]
@@ -141,7 +154,7 @@ class TokenView(View):
         )
 
     def post(self, request):
-        post_data = FormOrJsonParser().parse_body(request)
+        post_data = get_json_and_formdata(request)
         auth_client_id, auth_client_secret = extract_client_info_from_basic_auth(
             request
         )
@@ -221,7 +234,7 @@ class TokenView(View):
 @method_decorator(csrf_exempt, name="dispatch")
 class RevokeTokenView(View):
     def post(self, request):
-        post_data = FormOrJsonParser().parse_body(request)
+        post_data = get_json_and_formdata(request)
         auth_client_id, auth_client_secret = extract_client_info_from_basic_auth(
             request
         )

api/views/search.py

@@ -1,16 +1,14 @@
 from typing import Literal
 
-from ninja import Field
-
 from activities.models import PostInteraction
 from activities.services.search import SearchService
 from api import schemas
 from api.decorators import identity_required
-from api.views.base import api_router
+from hatchway import Field, api_view
 
 
-@api_router.get("/v2/search", response=schemas.Search)
 @identity_required
+@api_view.get
 def search(
     request,
     q: str,
@@ -24,26 +22,31 @@ def search(
     min_id: str | None = None,
     limit: int = 20,
     offset: int = 0,
-):
+) -> schemas.Search:
     if limit > 40:
         limit = 40
     result: dict[str, list] = {"accounts": [], "statuses": [], "hashtags": []}
     # We don't support pagination for searches yet
     if max_id or since_id or min_id or offset:
-        return result
+        return schemas.Search(**result)
     # Run search
     searcher = SearchService(q, request.identity)
     search_result = searcher.search_all()
     if type is None or type == "accounts":
-        result["accounts"] = [i.to_mastodon_json() for i in search_result["identities"]]
+        result["accounts"] = [
+            schemas.Account.from_identity(i, include_counts=False)
+            for i in search_result["identities"]
+        ]
     if type is None or type == "hashtag":
-        result["hashtag"] = [h.to_mastodon_json() for h in search_result["hashtags"]]
+        result["hashtag"] = [
+            schemas.Tag.from_hashtag(h) for h in search_result["hashtags"]
+        ]
     if type is None or type == "statuses":
         interactions = PostInteraction.get_post_interactions(
             search_result["posts"], request.identity
         )
         result["statuses"] = [
-            p.to_mastodon_json(interactions=interactions)
+            schemas.Status.from_post(p, interactions=interactions)
             for p in search_result["posts"]
         ]
-    return result
+    return schemas.Search(**result)

api/views/statuses.py

@@ -1,9 +1,8 @@
 from typing import Literal
 
 from django.forms import ValidationError
-from django.http import HttpRequest, HttpResponse
+from django.http import HttpRequest
 from django.shortcuts import get_object_or_404
-from ninja import Schema
 
 from activities.models import (
     Post,
@@ -14,11 +13,10 @@ from activities.models import (
 )
 from activities.services import PostService
 from api import schemas
-from api.views.base import api_router
+from api.decorators import identity_required
+from api.pagination import MastodonPaginator, PaginationResult
 from core.models import Config
-
+from hatchway import ApiResponse, Schema, api_view
-from ..decorators import identity_required
-from ..pagination import MastodonPaginator
 
 
 class PostStatusSchema(Schema):
@@ -32,9 +30,9 @@ class PostStatusSchema(Schema):
     media_ids: list[str] = []
 
 
-@api_router.post("/v1/statuses", response=schemas.Status)
 @identity_required
-def post_status(request, details: PostStatusSchema):
+@api_view.post
+def post_status(request, details: PostStatusSchema) -> schemas.Status:
     # Check text length
     if len(details.status) > Config.system.post_length:
         raise ValidationError("Status is too long")
@@ -66,74 +64,74 @@ def post_status(request, details: PostStatusSchema):
     )
     # Add their own timeline event for immediate visibility
     TimelineEvent.add_post(request.identity, post)
-    return post.to_mastodon_json()
+    return schemas.Status.from_post(post)
 
 
-@api_router.get("/v1/statuses/{id}", response=schemas.Status)
 @identity_required
-def status(request, id: str):
+@api_view.get
+def status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     interactions = PostInteraction.get_post_interactions([post], request.identity)
-    return post.to_mastodon_json(interactions=interactions)
+    return schemas.Status.from_post(post, interactions=interactions)
 
 
-@api_router.delete("/v1/statuses/{id}", response=schemas.Status)
 @identity_required
-def delete_status(request, id: str):
+@api_view.delete
+def delete_status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     PostService(post).delete()
-    return post.to_mastodon_json()
+    return schemas.Status.from_post(post)
 
 
-@api_router.get("/v1/statuses/{id}/context", response=schemas.Context)
 @identity_required
-def status_context(request, id: str):
+@api_view.get
+def status_context(request, id: str) -> schemas.Context:
     post = get_object_or_404(Post, pk=id)
     service = PostService(post)
     ancestors, descendants = service.context(request.identity)
     interactions = PostInteraction.get_post_interactions(
         ancestors + descendants, request.identity
     )
-    return {
+    return schemas.Context(
-        "ancestors": [
+        ancestors=[
-            p.to_mastodon_json(interactions=interactions) for p in reversed(ancestors)
+            schemas.Status.from_post(p, interactions=interactions)
+            for p in reversed(ancestors)
         ],
-        "descendants": [
+        descendants=[
-            p.to_mastodon_json(interactions=interactions) for p in descendants
+            schemas.Status.from_post(p, interactions=interactions) for p in descendants
         ],
-    }
+    )
 
 
-@api_router.post("/v1/statuses/{id}/favourite", response=schemas.Status)
 @identity_required
-def favourite_status(request, id: str):
+@api_view.post
+def favourite_status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     service = PostService(post)
     service.like_as(request.identity)
     interactions = PostInteraction.get_post_interactions([post], request.identity)
-    return post.to_mastodon_json(interactions=interactions)
+    return schemas.Status.from_post(post, interactions=interactions)
 
 
-@api_router.post("/v1/statuses/{id}/unfavourite", response=schemas.Status)
 @identity_required
-def unfavourite_status(request, id: str):
+@api_view.post
+def unfavourite_status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     service = PostService(post)
     service.unlike_as(request.identity)
     interactions = PostInteraction.get_post_interactions([post], request.identity)
-    return post.to_mastodon_json(interactions=interactions)
+    return schemas.Status.from_post(post, interactions=interactions)
 
 
-@api_router.get("/v1/statuses/{id}/favourited_by", response=list[schemas.Account])
+@api_view.get
 def favourited_by(
     request: HttpRequest,
-    response: HttpResponse,
     id: str,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Account]]:
     """
     View who favourited a given status.
     """
@@ -142,7 +140,7 @@ def favourited_by(
     post = get_object_or_404(Post, pk=id)
 
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[PostInteraction] = paginator.paginate(
         post.interactions.filter(
             type=PostInteraction.Types.like,
             state__in=PostInteractionStates.group_active(),
@@ -152,32 +150,37 @@ def favourited_by(
         since_id=since_id,
         limit=limit,
     )
-    pager.jsonify_results(lambda r: r.identity.to_mastodon_json(include_counts=False))
 
+    headers = {}
     if pager.results:
-        response.headers["Link"] = pager.link_header(
+        headers = {"link": pager.link_header(request, ["limit"])}
-            request,
+    return ApiResponse(
-            ["limit"],
+        [
+            schemas.Account.from_identity(
+                interaction.identity,
+                include_counts=False,
+            )
+            for interaction in pager.results
+        ],
+        headers=headers,
     )
 
-    return pager.json_results
 
-
-@api_router.post("/v1/statuses/{id}/reblog", response=schemas.Status)
 @identity_required
-def reblog_status(request, id: str):
+@api_view.post
+def reblog_status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     service = PostService(post)
     service.boost_as(request.identity)
     interactions = PostInteraction.get_post_interactions([post], request.identity)
-    return post.to_mastodon_json(interactions=interactions)
+    return schemas.Status.from_post(post, interactions=interactions)
 
 
-@api_router.post("/v1/statuses/{id}/unreblog", response=schemas.Status)
 @identity_required
-def unreblog_status(request, id: str):
+@api_view.post
+def unreblog_status(request, id: str) -> schemas.Status:
     post = get_object_or_404(Post, pk=id)
     service = PostService(post)
     service.unboost_as(request.identity)
     interactions = PostInteraction.get_post_interactions([post], request.identity)
-    return post.to_mastodon_json(interactions=interactions)
+    return schemas.Status.from_post(post, interactions=interactions)

api/views/timelines.py

@@ -1,23 +1,23 @@
-from django.http import HttpRequest, HttpResponse, JsonResponse
+from django.http import HttpRequest
 
+from activities.models import Post
 from activities.services import TimelineService
 from api import schemas
 from api.decorators import identity_required
-from api.pagination import MastodonPaginator
+from api.pagination import MastodonPaginator, PaginatingApiResponse, PaginationResult
-from api.views.base import api_router
 from core.models import Config
+from hatchway import ApiError, ApiResponse, api_view
 
 
-@api_router.get("/v1/timelines/home", response=list[schemas.Status])
 @identity_required
+@api_view.get
 def home(
     request: HttpRequest,
-    response: HttpResponse,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Status]]:
     # Grab a paginated result set of instances
     paginator = MastodonPaginator()
     queryset = TimelineService(request.identity).home()
@@ -41,18 +41,16 @@ def home(
         since_id=since_id,
         limit=limit,
     )
-    # Convert those to the JSON form
+    return PaginatingApiResponse(
-    pager.jsonify_status_events(identity=request.identity)
+        schemas.Status.map_from_timeline_event(pager.results, request.identity),
-    # Add the link header if needed
+        request=request,
-    if pager.results:
+        include_params=["limit"],
-        response.headers["Link"] = pager.link_header(request, ["limit"])
+    )
-    return pager.json_results
 
 
-@api_router.get("/v1/timelines/public", response=list[schemas.Status])
+@api_view.get
 def public(
     request: HttpRequest,
-    response: HttpResponse,
     local: bool = False,
     remote: bool = False,
     only_media: bool = False,
@@ -60,9 +58,9 @@ def public(
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Status]]:
     if not request.identity and not Config.system.public_timeline:
-        return JsonResponse({"error": "public timeline is disabled"}, status=422)
+        raise ApiError(error="public timeline is disabled", status=422)
 
     if local:
         queryset = TimelineService(request.identity).local()
@@ -74,29 +72,24 @@ def public(
         queryset = queryset.filter(attachments__id__isnull=True)
     # Grab a paginated result set of instances
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[Post] = paginator.paginate(
         queryset,
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    # Convert those to the JSON form
+    return PaginatingApiResponse(
-    pager.jsonify_posts(identity=request.identity)
+        schemas.Status.map_from_post(pager.results, request.identity),
-    # Add the link header if needed
+        request=request,
-    if pager.results:
+        include_params=["limit", "local", "remote", "only_media"],
-        response.headers["Link"] = pager.link_header(
-            request,
-            ["limit", "local", "remote", "only_media"],
     )
-    return pager.json_results
 
 
-@api_router.get("/v1/timelines/tag/{hashtag}", response=list[schemas.Status])
 @identity_required
+@api_view.get
 def hashtag(
     request: HttpRequest,
-    response: HttpResponse,
     hashtag: str,
     local: bool = False,
     only_media: bool = False,
@@ -104,7 +97,7 @@ def hashtag(
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Status]]:
     if limit > 40:
         limit = 40
     queryset = TimelineService(request.identity).hashtag(hashtag)
@@ -114,63 +107,54 @@ def hashtag(
         queryset = queryset.filter(attachments__id__isnull=True)
     # Grab a paginated result set of instances
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[Post] = paginator.paginate(
         queryset,
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    # Convert those to the JSON form
+    return PaginatingApiResponse(
-    pager.jsonify_posts(identity=request.identity)
+        schemas.Status.map_from_post(pager.results, request.identity),
-    # Add a link header if we need to
+        request=request,
-    if pager.results:
+        include_params=["limit", "local", "remote", "only_media"],
-        response.headers["Link"] = pager.link_header(
-            request,
-            ["limit", "local", "remote", "only_media"],
     )
-    return pager.json_results
 
 
-@api_router.get("/v1/conversations", response=list[schemas.Status])
 @identity_required
+@api_view.get
 def conversations(
     request: HttpRequest,
-    response: HttpResponse,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> list[schemas.Status]:
     # We don't implement this yet
     return []
 
 
-@api_router.get("/v1/favourites", response=list[schemas.Status])
 @identity_required
+@api_view.get
 def favourites(
     request: HttpRequest,
-    response: HttpResponse,
     max_id: str | None = None,
     since_id: str | None = None,
     min_id: str | None = None,
     limit: int = 20,
-):
+) -> ApiResponse[list[schemas.Status]]:
     queryset = TimelineService(request.identity).likes()
 
     paginator = MastodonPaginator()
-    pager = paginator.paginate(
+    pager: PaginationResult[Post] = paginator.paginate(
         queryset,
         min_id=min_id,
         max_id=max_id,
         since_id=since_id,
         limit=limit,
     )
-    # Convert those to the JSON form
+    return PaginatingApiResponse(
-    pager.jsonify_posts(identity=request.identity)
+        schemas.Status.map_from_post(pager.results, request.identity),
-
+        request=request,
-    # Add the link header if needed
+        include_params=["limit"],
-    if pager.results:
+    )
-        response.headers["Link"] = pager.link_header(request, ["limit"])
-
-    return pager.json_results

hatchway/__init__.py

@@ -0,0 +1,5 @@
+from .http import ApiError, ApiResponse  # noqa
+from .schema import Field, Schema  # noqa
+from .types import Body, BodyDirect, Path, Query, QueryOrBody  # noqa
+from .urls import methods  # noqa
+from .view import api_view  # noqa

hatchway/constants.py

@@ -0,0 +1,10 @@
+import enum
+
+
+class InputSource(str, enum.Enum):
+    path = "path"
+    query = "query"
+    body = "body"
+    body_direct = "body_direct"
+    query_and_body_direct = "query_and_body_direct"
+    file = "file"

hatchway/http.py

@@ -0,0 +1,47 @@
+import json
+from typing import Generic, TypeVar
+
+from django.core.serializers.json import DjangoJSONEncoder
+from django.http import HttpResponse
+
+T = TypeVar("T")
+
+
+class ApiResponse(Generic[T], HttpResponse):
+    """
+    A way to return extra information with a response if you want
+    headers, etc.
+    """
+
+    def __init__(
+        self,
+        data: T,
+        encoder=DjangoJSONEncoder,
+        json_dumps_params: dict[str, object] | None = None,
+        finalize: bool = False,
+        **kwargs
+    ):
+        self.data = data
+        self.encoder = encoder
+        self.json_dumps_params = json_dumps_params or {}
+        kwargs.setdefault("content_type", "application/json")
+        super().__init__(content=b"(unfinalised)", **kwargs)
+        if finalize:
+            self.finalize()
+
+    def finalize(self):
+        """
+        Converts whatever our current data is into HttpResponse content
+        """
+        # TODO: Automatically call this when we're asked to write output?
+        self.content = json.dumps(self.data, cls=self.encoder, **self.json_dumps_params)
+
+
+class ApiError(BaseException):
+    """
+    A handy way to raise an error with JSONable contents
+    """
+
+    def __init__(self, status: int, error: str):
+        self.status = status
+        self.error = error

hatchway/schema.py

@@ -0,0 +1,52 @@
+from typing import Any
+
+from django.db.models import Manager, QuerySet
+from django.db.models.fields.files import FieldFile
+from django.template import Variable, VariableDoesNotExist
+from pydantic.fields import Field  # noqa
+from pydantic.main import BaseModel
+from pydantic.utils import GetterDict
+
+
+class DjangoGetterDict(GetterDict):
+    def __init__(self, obj: Any):
+        self._obj = obj
+
+    def __getitem__(self, key: str) -> Any:
+        try:
+            item = getattr(self._obj, key)
+        except AttributeError:
+            try:
+                item = Variable(key).resolve(self._obj)
+            except VariableDoesNotExist as e:
+                raise KeyError(key) from e
+        return self._convert_result(item)
+
+    def get(self, key: Any, default: Any = None) -> Any:
+        try:
+            return self[key]
+        except KeyError:
+            return default
+
+    def _convert_result(self, result: Any) -> Any:
+        if isinstance(result, Manager):
+            return list(result.all())
+
+        elif isinstance(result, getattr(QuerySet, "__origin__", QuerySet)):
+            return list(result)
+
+        if callable(result):
+            return result()
+
+        elif isinstance(result, FieldFile):
+            if not result:
+                return None
+            return result.url
+
+        return result
+
+
+class Schema(BaseModel):
+    class Config:
+        orm_mode = True
+        getter_dict = DjangoGetterDict

hatchway/tests/test_types.py

@@ -0,0 +1,63 @@
+from typing import Literal, Optional, Union
+
+from django.core.files import File
+
+from hatchway.http import ApiResponse
+from hatchway.types import (
+    Query,
+    QueryType,
+    acceptable_input,
+    extract_output_type,
+    extract_signifier,
+    is_optional,
+)
+
+
+def test_is_optional():
+
+    assert is_optional(Optional[int]) == (True, int)
+    assert is_optional(Union[int, None]) == (True, int)
+    assert is_optional(Union[None, int]) == (True, int)
+    assert is_optional(int | None) == (True, int)
+    assert is_optional(None | int) == (True, int)
+    assert is_optional(int) == (False, int)
+    assert is_optional(Query[int]) == (False, Query[int])
+
+
+def test_extract_signifier():
+
+    assert extract_signifier(int) == (None, int)
+    assert extract_signifier(Query[int]) == (QueryType, int)
+    assert extract_signifier(Query[Optional[int]]) == (  # type:ignore
+        QueryType,
+        Optional[int],
+    )
+    assert extract_signifier(Query[int | None]) == (  # type:ignore
+        QueryType,
+        Optional[int],
+    )
+    assert extract_signifier(Optional[Query[int]]) == (QueryType, Optional[int])
+
+
+def test_extract_output_type():
+
+    assert extract_output_type(int) == int
+    assert extract_output_type(ApiResponse[int]) == int
+    assert extract_output_type(ApiResponse[int | str]) == int | str
+
+
+def test_acceptable_input():
+
+    assert acceptable_input(str) is True
+    assert acceptable_input(int) is True
+    assert acceptable_input(Query[int]) is True
+    assert acceptable_input(Optional[int]) is True
+    assert acceptable_input(int | None) is True
+    assert acceptable_input(int | str | None) is True
+    assert acceptable_input(Query[int | None]) is True  # type: ignore
+    assert acceptable_input(File) is True
+    assert acceptable_input(list[str]) is True
+    assert acceptable_input(dict[str, int]) is True
+    assert acceptable_input(Literal["a", "b"]) is True
+    assert acceptable_input(frozenset) is False
+    assert acceptable_input(dict[str, frozenset]) is False

hatchway/tests/test_view.py

@@ -0,0 +1,244 @@
+import json
+
+import pytest
+from django.core import files
+from django.core.files.uploadedfile import SimpleUploadedFile
+from django.http import QueryDict
+from django.test import RequestFactory
+from django.test.client import MULTIPART_CONTENT
+from pydantic import BaseModel
+
+from hatchway import ApiError, Body, QueryOrBody, api_view
+from hatchway.view import ApiView
+
+
+def test_basic_view():
+    """
+    Tests that a view with simple types works correctly
+    """
+
+    @api_view
+    def test_view(
+        request,
+        a: int,
+        b: QueryOrBody[int | None] = None,
+        c: str = "x",
+    ) -> str:
+        if b is None:
+            return c * a
+        else:
+            return c * (a - b)
+
+    # Call it with a few different patterns to verify it's type coercing right
+    factory = RequestFactory()
+
+    # Implicit query param
+    response = test_view(factory.get("/test/?a=4"))
+    assert json.loads(response.content) == "xxxx"
+
+    # QueryOrBody pulling from query
+    response = test_view(factory.get("/test/?a=4&b=2"))
+    assert json.loads(response.content) == "xx"
+
+    # QueryOrBody pulling from formdata body
+    response = test_view(factory.post("/test/?a=4", {"b": "3"}))
+    assert json.loads(response.content) == "x"
+
+    # QueryOrBody pulling from JSON body
+    response = test_view(
+        factory.post(
+            "/test/?a=4", json.dumps({"b": 3}), content_type="application/json"
+        )
+    )
+    assert json.loads(response.content) == "x"
+
+    # Implicit Query not pulling from body
+    with pytest.raises(TypeError):
+        test_view(factory.post("/test/", {"a": 4, "b": 3}))
+
+
+def test_body_direct():
+    """
+    Tests that a Pydantic model with BodyDirect gets its fields from the top level
+    """
+
+    class TestModel(BaseModel):
+        number: int
+        name: str
+
+    @api_view
+    def test_view(request, data: TestModel) -> int:
+        return data.number
+
+    factory = RequestFactory()
+
+    # formdata version
+    response = test_view(factory.post("/test/", {"number": "123", "name": "Andrew"}))
+    assert json.loads(response.content) == 123
+
+    # JSON body version
+    response = test_view(
+        factory.post(
+            "/test/",
+            json.dumps({"number": "123", "name": "Andrew"}),
+            content_type="application/json",
+        )
+    )
+    assert json.loads(response.content) == 123
+
+
+def test_list_response():
+    """
+    Tests that a view with a list response type works correctly with both
+    dicts and pydantic model instances.
+    """
+
+    class TestModel(BaseModel):
+        number: int
+        name: str
+
+    @api_view
+    def test_view_dict(request) -> list[TestModel]:
+        return [
+            {"name": "Andrew", "number": 1},  # type:ignore
+            {"name": "Alice", "number": 0},  # type:ignore
+        ]
+
+    @api_view
+    def test_view_model(request) -> list[TestModel]:
+        return [TestModel(name="Andrew", number=1), TestModel(name="Alice", number=0)]
+
+    response = test_view_dict(RequestFactory().get("/test/"))
+    assert json.loads(response.content) == [
+        {"name": "Andrew", "number": 1},
+        {"name": "Alice", "number": 0},
+    ]
+
+    response = test_view_model(RequestFactory().get("/test/"))
+    assert json.loads(response.content) == [
+        {"name": "Andrew", "number": 1},
+        {"name": "Alice", "number": 0},
+    ]
+
+
+def test_patch_body():
+    """
+    Tests that PATCH also gets its body parsed
+    """
+
+    @api_view.patch
+    def test_view(request, a: Body[int]):
+        return a
+
+    factory = RequestFactory()
+    response = test_view(
+        factory.patch(
+            "/test/",
+            content_type=MULTIPART_CONTENT,
+            data=factory._encode_data({"a": "42"}, MULTIPART_CONTENT),
+        )
+    )
+    assert json.loads(response.content) == 42
+
+
+def test_file_body():
+    """
+    Tests that file uploads work right
+    """
+
+    @api_view.post
+    def test_view(request, a: Body[int], b: files.File) -> str:
+        return str(a) + b.read().decode("ascii")
+
+    factory = RequestFactory()
+    uploaded_file = SimpleUploadedFile(
+        "file.txt",
+        b"MY FILE IS AMAZING",
+        content_type="text/plain",
+    )
+    response = test_view(
+        factory.post(
+            "/test/",
+            data={"a": 42, "b": uploaded_file},
+        )
+    )
+    assert json.loads(response.content) == "42MY FILE IS AMAZING"
+
+
+def test_no_response():
+    """
+    Tests that a view with no response type returns the contents verbatim
+    """
+
+    @api_view
+    def test_view(request):
+        return [1, "woooooo"]
+
+    response = test_view(RequestFactory().get("/test/"))
+    assert json.loads(response.content) == [1, "woooooo"]
+
+
+def test_wrong_method():
+    """
+    Tests that a view with a method limiter works
+    """
+
+    @api_view.get
+    def test_view(request):
+        return "yay"
+
+    response = test_view(RequestFactory().get("/test/"))
+    assert json.loads(response.content) == "yay"
+
+    response = test_view(RequestFactory().post("/test/"))
+    assert response.status_code == 405
+
+
+def test_api_error():
+    """
+    Tests that ApiError propagates right
+    """
+
+    @api_view.get
+    def test_view(request):
+        raise ApiError(401, "you did a bad thing")
+
+    response = test_view(RequestFactory().get("/test/"))
+    assert json.loads(response.content) == {"error": "you did a bad thing"}
+    assert response.status_code == 401
+
+
+def test_unusable_type():
+    """
+    Tests that you get a nice error when you use a type on an input that
+    Pydantic doesn't understand.
+    """
+
+    with pytest.raises(ValueError):
+
+        @api_view.get
+        def test_view(request, a: RequestFactory):
+            pass
+
+
+def test_get_values():
+    """
+    Tests that ApiView.get_values correctly handles lists
+    """
+
+    assert ApiView.get_values({"a": 2, "b": [3, 4]}) == {"a": 2, "b": [3, 4]}
+    assert ApiView.get_values({"a": 2, "b[]": [3, 4]}) == {"a": 2, "b": [3, 4]}
+    assert ApiView.get_values(QueryDict("a=2&b=3&b=4")) == {"a": "2", "b": ["3", "4"]}
+    assert ApiView.get_values(QueryDict("a=2&b[]=3&b[]=4")) == {
+        "a": "2",
+        "b": ["3", "4"],
+    }
+    assert ApiView.get_values(QueryDict("a=2&b=3")) == {"a": "2", "b": "3"}
+    assert ApiView.get_values(QueryDict("a=2&b[]=3")) == {"a": "2", "b": ["3"]}
+    assert ApiView.get_values(QueryDict("a[b]=1")) == {"a": {"b": "1"}}
+    assert ApiView.get_values(QueryDict("a[b]=1&a[c]=2")) == {"a": {"b": "1", "c": "2"}}
+    assert ApiView.get_values(QueryDict("a[b][c]=1")) == {"a": {"b": {"c": "1"}}}
+    assert ApiView.get_values(QueryDict("a[b][c][]=1")) == {"a": {"b": {"c": ["1"]}}}
+    assert ApiView.get_values(QueryDict("a[b][]=1&a[b][]=2")) == {
+        "a": {"b": ["1", "2"]}
+    }

hatchway/types.py

@@ -0,0 +1,145 @@
+from types import NoneType, UnionType
+from typing import (  # type: ignore[attr-defined]
+    Annotated,
+    Any,
+    Literal,
+    Optional,
+    TypeVar,
+    Union,
+    _AnnotatedAlias,
+    _GenericAlias,
+    get_args,
+    get_origin,
+)
+
+from django.core import files
+from pydantic import BaseModel
+
+from .http import ApiResponse
+
+T = TypeVar("T")
+
+
+class PathType:
+    """
+    An input pulled from the path (url resolver kwargs)
+    """
+
+
+class QueryType:
+    """
+    An input pulled from the query parameters (request.GET)
+    """
+
+
+class BodyType:
+    """
+    An input pulled from the POST body (request.POST or a JSON body)
+    """
+
+
+class FileType:
+    """
+    An input pulled from the POST body (request.POST or a JSON body)
+    """
+
+
+class BodyDirectType:
+    """
+    A Pydantic model whose keys are all looked for in the top-level
+    POST data, rather than in a dict under a key named after the input.
+    """
+
+
+class QueryOrBodyType:
+    """
+    An input pulled from either query parameters or post data.
+    """
+
+
+Path = Annotated[T, PathType]
+Query = Annotated[T, QueryType]
+Body = Annotated[T, BodyType]
+File = Annotated[T, FileType]
+BodyDirect = Annotated[T, BodyDirectType]
+QueryOrBody = Annotated[T, QueryOrBodyType]
+
+
+def is_optional(annotation) -> tuple[bool, Any]:
+    """
+    If an annotation is Optional or | None, returns (True, internal type).
+    Returns (False, annotation) otherwise.
+    """
+    if (isinstance(annotation, _GenericAlias) and annotation.__origin__ is Union) or (
+        isinstance(annotation, UnionType)
+    ):
+        args = get_args(annotation)
+        if len(args) > 2:
+            return False, annotation
+        if args[0] is NoneType:
+            return True, args[1]
+        if args[1] is NoneType:
+            return True, args[0]
+        return False, annotation
+    return False, annotation
+
+
+def extract_signifier(annotation) -> tuple[Any, Any]:
+    """
+    Given a type annotation, looks to see if it can find a input source
+    signifier (Path, Query, etc.)
+
+    If it can, returns (signifier, annotation_without_signifier)
+    If not, returns (None, annotation)
+    """
+    our_generics = {
+        PathType,
+        QueryType,
+        BodyType,
+        FileType,
+        BodyDirectType,
+        QueryOrBodyType,
+    }
+    # Remove any optional-style wrapper
+    optional, internal_annotation = is_optional(annotation)
+    # Is it an annotation?
+    if isinstance(internal_annotation, _AnnotatedAlias):
+        args = get_args(internal_annotation)
+        for arg in args[1:]:
+            if arg in our_generics:
+                if optional:
+                    return (arg, Optional[args[0]])
+                else:
+                    return (arg, args[0])
+    return None, annotation
+
+
+def extract_output_type(annotation):
+    """
+    Returns the right response type for a function
+    """
+    # If the type is ApiResponse, we want to pull out its inside
+    if isinstance(annotation, _GenericAlias):
+        if get_origin(annotation) == ApiResponse:
+            return get_args(annotation)[0]
+    return annotation
+
+
+def acceptable_input(annotation) -> bool:
+    """
+    Returns if this annotation is something we think we can accept as input
+    """
+    _, inner_type = extract_signifier(annotation)
+    try:
+        if issubclass(inner_type, BaseModel):
+            return True
+    except TypeError:
+        pass
+    if inner_type in [str, int, list, tuple, bool, Any, files.File, type(None)]:
+        return True
+    origin = get_origin(inner_type)
+    if origin == Literal:
+        return True
+    if origin in [Union, UnionType, dict, list, tuple]:
+        return all(acceptable_input(a) for a in get_args(inner_type))
+    return False

hatchway/urls.py

@@ -0,0 +1,32 @@
+from collections.abc import Callable
+from typing import Any
+
+from django.http import HttpResponseNotAllowed
+
+
+class Methods:
+    """
+    Allows easy multi-method dispatch to different functions
+    """
+
+    csrf_exempt = True
+
+    def __init__(self, **callables: Callable):
+        self.callables = {
+            method.lower(): callable for method, callable in callables.items()
+        }
+        unknown_methods = set(self.callables.keys()).difference(
+            {"get", "post", "patch", "put", "delete"}
+        )
+        if unknown_methods:
+            raise ValueError(f"Cannot route methods: {unknown_methods}")
+
+    def __call__(self, request, *args, **kwargs) -> Any:
+        method = request.method.lower()
+        if method in self.callables:
+            return self.callables[method](request, *args, **kwargs)
+        else:
+            return HttpResponseNotAllowed(self.callables.keys())
+
+
+methods = Methods

hatchway/view.py

@@ -0,0 +1,297 @@
+import json
+from collections.abc import Callable
+from typing import Any, Optional, get_type_hints
+
+from django.core import files
+from django.http import HttpRequest, HttpResponseNotAllowed, QueryDict
+from django.http.multipartparser import MultiPartParser
+from pydantic import BaseModel, create_model
+
+from .constants import InputSource
+from .http import ApiError, ApiResponse
+from .types import (
+    BodyDirectType,
+    BodyType,
+    FileType,
+    PathType,
+    QueryOrBodyType,
+    QueryType,
+    acceptable_input,
+    extract_output_type,
+    extract_signifier,
+    is_optional,
+)
+
+
+class ApiView:
+    """
+    A view 'wrapper' object that replaces the API view for anything further
+    up the stack.
+
+    Unlike Django's class-based views, we don't need an as_view pattern
+    as we are careful never to write anything per-request to self.
+    """
+
+    csrf_exempt = True
+
+    def __init__(
+        self,
+        view: Callable,
+        input_types: dict[str, Any] | None = None,
+        output_type: Any = None,
+        implicit_lists: bool = True,
+        method: str | None = None,
+    ):
+        self.view = view
+        self.implicit_lists = implicit_lists
+        self.view_name = getattr(view, "__name__", "unknown_view")
+        self.method = method
+        # Extract input/output types from view annotations if we need to
+        self.input_types = input_types
+        if self.input_types is None:
+            self.input_types = get_type_hints(view, include_extras=True)
+            if "return" in self.input_types:
+                del self.input_types["return"]
+        self.output_type = output_type
+        if self.output_type is None:
+            try:
+                self.output_type = extract_output_type(
+                    get_type_hints(view, include_extras=True)["return"]
+                )
+            except KeyError:
+                self.output_type = None
+        self.compile()
+
+    @classmethod
+    def get(cls, view: Callable):
+        return cls(view=view, method="get")
+
+    @classmethod
+    def post(cls, view: Callable):
+        return cls(view=view, method="post")
+
+    @classmethod
+    def put(cls, view: Callable):
+        return cls(view=view, method="put")
+
+    @classmethod
+    def patch(cls, view: Callable):
+        return cls(view=view, method="patch")
+
+    @classmethod
+    def delete(cls, view: Callable):
+        return cls(view=view, method="delete")
+
+    @classmethod
+    def sources_for_input(cls, input_type) -> tuple[list[InputSource], Any]:
+        """
+        Given a type that can appear as a request parameter type, returns
+        what sources it can come from, and what its type is as understood
+        by Pydantic.
+        """
+        signifier, input_type = extract_signifier(input_type)
+        if signifier is QueryType:
+            return ([InputSource.query], input_type)
+        elif signifier is BodyType:
+            return ([InputSource.body], input_type)
+        elif signifier is BodyDirectType:
+            if not issubclass(input_type, BaseModel):
+                raise ValueError(
+                    "You cannot use BodyDirect on something that is not a Pydantic model"
+                )
+            return ([InputSource.body_direct], input_type)
+        elif signifier is PathType:
+            return ([InputSource.path], input_type)
+        elif (
+            signifier is FileType
+            or input_type is files.File
+            or is_optional(input_type)[1] is files.File
+        ):
+            return ([InputSource.file], input_type)
+        elif signifier is QueryOrBodyType:
+            return ([InputSource.query, InputSource.body], input_type)
+        # Is it a Pydantic model, which means it's implicitly body?
+        elif isinstance(input_type, type) and issubclass(input_type, BaseModel):
+            return ([InputSource.body], input_type)
+        # Otherwise, we look in the path first and then the query
+        else:
+            return ([InputSource.path, InputSource.query], input_type)
+
+    @classmethod
+    def get_values(cls, data, use_square_brackets=True) -> dict[str, Any]:
+        """
+        Given a QueryDict or normal dict, returns data taking into account
+        lists made by repeated values or by suffixing names with [].
+        """
+        result: dict[str, Any] = {}
+        for key, value in data.items():
+            # If it's a query dict with multiple values, make it a list
+            if isinstance(data, QueryDict):
+                values = data.getlist(key)
+                if len(values) > 1:
+                    value = values
+            # If it is in dict-ish/list-ish syntax, adhere to that
+            # TODO: Make this better handle badly formed keys
+            if "[" in key and use_square_brackets:
+                parts = key.split("[")
+                target = result
+                last_key = parts[0]
+                for part in parts[1:]:
+                    part = part.rstrip("]")
+                    if not part:
+                        target = target.setdefault(last_key, [])
+                    else:
+                        target = target.setdefault(last_key, {})
+                        last_key = part
+                if isinstance(target, list):
+                    if isinstance(value, list):
+                        target.extend(value)
+                    else:
+                        target.append(value)
+                else:
+                    target[last_key] = value
+            else:
+                result[key] = value
+        return result
+
+    def compile(self):
+        self.sources: dict[str, list[InputSource]] = {}
+        amount_from_body = 0
+        pydantic_model_dict = {}
+        self.input_files = set()
+        last_body_type = None
+        # For each input item, work out where to pull it from
+        for name, input_type in self.input_types.items():
+            # Do some basic typechecking to stop things that aren't allowed
+            if isinstance(input_type, type) and issubclass(input_type, HttpRequest):
+                continue
+            if not acceptable_input(input_type):
+                # Strip away any singifiers for the error
+                _, inner_type = extract_signifier(input_type)
+                raise ValueError(
+                    f"Input argument {name} has an unsupported type {inner_type}"
+                )
+            sources, pydantic_type = self.sources_for_input(input_type)
+            self.sources[name] = sources
+            # Keep count of how many are pulling from the body
+            if InputSource.body in sources:
+                amount_from_body += 1
+                last_body_type = pydantic_type
+            if InputSource.file in sources:
+                self.input_files.add(name)
+            else:
+                pydantic_model_dict[name] = (Optional[pydantic_type], ...)
+        # If there is just one thing pulling from the body and it's a BaseModel,
+        # signify that it's actually pulling from the body keys directly and
+        # not a sub-dict
+        if amount_from_body == 1:
+            for name, sources in self.sources.items():
+                if (
+                    InputSource.body in sources
+                    and isinstance(last_body_type, type)
+                    and issubclass(last_body_type, BaseModel)
+                ):
+                    self.sources[name] = [
+                        x for x in sources if x != InputSource.body
+                    ] + [InputSource.body_direct]
+        # Turn all the main arguments into Pydantic parsing models
+        try:
+            self.input_model = create_model(
+                f"{self.view_name}_input", **pydantic_model_dict
+            )
+        except RuntimeError:
+            raise ValueError(
+                f"One or more inputs on view {self.view_name} have a bad configuration"
+            )
+        if self.output_type is not None:
+            self.output_model = create_model(
+                f"{self.view_name}_output", value=(self.output_type, ...)
+            )
+
+    def __call__(self, request: HttpRequest, *args, **kwargs):
+        """
+        Entrypoint when this is called as a view.
+        """
+        # Do a method check if we have one set
+        if self.method and self.method.upper() != request.method:
+            return HttpResponseNotAllowed([self.method])
+        # For each item we can source, go find it if we can
+        query_values = self.get_values(request.GET)
+        body_values = self.get_values(request.POST)
+        files_values = self.get_values(request.FILES)
+        # If it's a PUT or PATCH method, work around Django not handling FILES
+        # or POST on those requests
+        if request.method in ["PATCH", "PUT"]:
+            if request.content_type == "multipart/form-data":
+                POST, FILES = MultiPartParser(
+                    request.META, request, request.upload_handlers, request.encoding
+                ).parse()
+                body_values = self.get_values(POST)
+                files_values = self.get_values(FILES)
+            elif request.content_type == "application/x-www-form-urlencoded":
+                POST = QueryDict(request.body, encoding=request._encoding)
+                body_values = self.get_values(POST)
+        # If there was a JSON body, go load that
+        if request.content_type == "application/json" and request.body.strip():
+            body_values.update(self.get_values(json.loads(request.body)))
+        values = {}
+        for name, sources in self.sources.items():
+            for source in sources:
+                if source == InputSource.path:
+                    if name in kwargs:
+                        values[name] = kwargs[name]
+                        break
+                elif source == InputSource.query:
+                    if name in query_values:
+                        values[name] = query_values[name]
+                        break
+                elif source == InputSource.body:
+                    if name in body_values:
+                        values[name] = body_values[name]
+                        break
+                elif source == InputSource.file:
+                    if name in files_values:
+                        values[name] = files_values[name]
+                        break
+                elif source == InputSource.body_direct:
+                    values[name] = body_values
+                    break
+                elif source == InputSource.query_and_body_direct:
+                    values[name] = dict(query_values)
+                    values[name].update(body_values)
+                    break
+                else:
+                    raise ValueError(f"Unknown source {source}")
+            else:
+                values[name] = None
+        # Give that to the Pydantic model to make it handle stuff
+        model_instance = self.input_model(**values)
+        kwargs = {
+            name: getattr(model_instance, name)
+            for name in model_instance.__fields__
+            if values[name] is not None  # Trim out missing fields
+        }
+        # Add in any files
+        # TODO: HTTP error if file is not optional
+        for name in self.input_files:
+            kwargs[name] = files_values.get(name, None)
+        # Call the view with those as kwargs
+        try:
+            response = self.view(request, **kwargs)
+        except ApiError as error:
+            return ApiResponse(
+                {"error": error.error}, status=error.status, finalize=True
+            )
+        # If it's not an ApiResponse, make it one
+        if not isinstance(response, ApiResponse):
+            response = ApiResponse(response)
+        # Get pydantic to coerce the output response
+        if self.output_type is not None:
+            response.data = self.output_model(value=response.data).dict()["value"]
+        elif isinstance(response.data, BaseModel):
+            response.data = response.data.dict()
+        response.finalize()
+        return response
+
+
+api_view = ApiView

requirements.txt

@@ -6,7 +6,6 @@ django-cache-url~=3.4.2
 django-cors-headers~=3.13.0
 django-debug-toolbar~=3.8.1
 django-htmx~=1.13.0
-django-ninja~=0.19.1
 django-oauth-toolkit~=2.2.0
 django-storages[google,boto3]~=1.13.1
 django~=4.1

takahe/urls.py

@@ -3,7 +3,7 @@ from django.contrib import admin as djadmin
 from django.urls import include, path, re_path
 
 from activities.views import compose, debug, explore, follows, posts, search, timelines
-from api.views import api_router, oauth
+from api.views import oauth
 from core import views as core
 from mediaproxy import views as mediaproxy
 from stator import views as stator
@@ -296,7 +296,7 @@ urlpatterns = [
     path("actor/outbox/", activitypub.EmptyOutbox.as_view()),
     path("inbox/", activitypub.Inbox.as_view(), name="shared_inbox"),
     # API/Oauth
-    path("api/", api_router.urls),
+    path("api/", include("api.urls")),
     path("oauth/authorize", oauth.AuthorizationView.as_view()),
     path("oauth/token", oauth.TokenView.as_view()),
     path("oauth/revoke", oauth.RevokeTokenView.as_view()),

tests/activities/models/test_post.py

@@ -281,7 +281,9 @@ def test_inbound_posts(
     InboxMessage.objects.create(message=message)
 
     # Run stator and ensure that made the post
+    print("prestat")
     stator.run_single_cycle_sync()
+    print("poststat")
     post = Post.objects.get(object_uri="https://remote.test/test-post")
     assert post.content == "post version one"
     assert post.published.day == 13