Fix ParadeDB key field handling

Use uuid ids in the ParadeDB documents table to satisfy BM25 key_field requirements and adjust indexing/search to dump FlakeIds consistently. Also fix the ParadeDB backend unit test to avoid invalid pinning on struct fields.
This commit is contained in:
Lain Soykaf 2026-03-04 13:04:41 +04:00
commit dee7ac0415
No known key found for this signature in database
2 changed files with 58 additions and 30 deletions

View file

@ -67,37 +67,43 @@ defmodule Pleroma.Search.ParadeDB do
maybe_search_data = object_to_search_data(activity.object) maybe_search_data = object_to_search_data(activity.object)
if activity.data["type"] == "Create" and maybe_search_data do if activity.data["type"] == "Create" and maybe_search_data do
dumped_activity_id = dump_activity_id(activity.id)
actor_ap_id = activity.data["actor"] actor_ap_id = activity.data["actor"]
published_at = DateTime.from_unix!(maybe_search_data.published) published_at = DateTime.from_unix!(maybe_search_data.published)
sql = if is_nil(dumped_activity_id) do
""" Logger.error("ParadeDB add_to_index failed: invalid activity id #{inspect(activity.id)}")
INSERT INTO #{table()} (id, object_id, object_ap_id, actor_ap_id, content, published_at) :ok
VALUES ($1, $2, $3, $4, $5, $6) else
ON CONFLICT (id) DO UPDATE SET sql =
object_id = EXCLUDED.object_id, """
object_ap_id = EXCLUDED.object_ap_id, INSERT INTO #{table()} (id, object_id, object_ap_id, actor_ap_id, content, published_at)
actor_ap_id = EXCLUDED.actor_ap_id, VALUES ($1, $2, $3, $4, $5, $6)
content = EXCLUDED.content, ON CONFLICT (id) DO UPDATE SET
published_at = EXCLUDED.published_at object_id = EXCLUDED.object_id,
""" object_ap_id = EXCLUDED.object_ap_id,
actor_ap_id = EXCLUDED.actor_ap_id,
content = EXCLUDED.content,
published_at = EXCLUDED.published_at
"""
params = [ params = [
activity.id, dumped_activity_id,
maybe_search_data.id, maybe_search_data.id,
maybe_search_data.ap, maybe_search_data.ap,
actor_ap_id, actor_ap_id,
maybe_search_data.content, maybe_search_data.content,
published_at published_at
] ]
case client_impl().query(sql, params) do case client_impl().query(sql, params) do
{:ok, _} -> {:ok, _} ->
:ok :ok
{:error, error} -> {:error, error} ->
Logger.error("ParadeDB add_to_index failed: #{inspect(error)}") Logger.error("ParadeDB add_to_index failed: #{inspect(error)}")
{:error, error} {:error, error}
end
end end
else else
:ok :ok
@ -139,7 +145,10 @@ defmodule Pleroma.Search.ParadeDB do
end end
defp fetch_activities(ids, user, query, author) do defp fetch_activities(ids, user, query, author) do
dumped_ids = Enum.map(ids, &FlakeId.from_string/1) dumped_ids =
ids
|> Enum.map(&dump_activity_id/1)
|> Enum.reject(&is_nil/1)
from(a in Activity, where: a.id in ^dumped_ids) from(a in Activity, where: a.id in ^dumped_ids)
|> Activity.with_preloaded_object() |> Activity.with_preloaded_object()
@ -189,7 +198,7 @@ defmodule Pleroma.Search.ParadeDB do
defp create_table_sql(table) do defp create_table_sql(table) do
""" """
CREATE TABLE IF NOT EXISTS #{table} ( CREATE TABLE IF NOT EXISTS #{table} (
id text PRIMARY KEY, id uuid PRIMARY KEY,
object_id bigint NOT NULL, object_id bigint NOT NULL,
object_ap_id text NOT NULL, object_ap_id text NOT NULL,
actor_ap_id text, actor_ap_id text,
@ -210,4 +219,15 @@ defmodule Pleroma.Search.ParadeDB do
WITH (key_field = 'id') WITH (key_field = 'id')
""" """
end end
defp dump_activity_id(<<_::binary-size(16)>> = uuid), do: uuid
defp dump_activity_id(id) when is_binary(id) do
case FlakeId.Ecto.CompatType.dump(id) do
{:ok, <<_::binary-size(16)>> = uuid} -> uuid
_ -> nil
end
end
defp dump_activity_id(_), do: nil
end end

View file

@ -35,7 +35,9 @@ defmodule Pleroma.Search.ParadeDBTest do
[activity_id, object_id, object_ap_id, actor_ap_id, content, published_at] = params [activity_id, object_id, object_ap_id, actor_ap_id, content, published_at] = params
assert activity_id == activity.id {:ok, dumped_activity_id} = FlakeId.Ecto.CompatType.dump(activity.id)
assert activity_id == dumped_activity_id
assert is_integer(object_id) assert is_integer(object_id)
assert is_binary(object_ap_id) assert is_binary(object_ap_id)
assert actor_ap_id == activity.data["actor"] assert actor_ap_id == activity.data["actor"]
@ -144,12 +146,18 @@ defmodule Pleroma.Search.ParadeDBTest do
{:ok, activity1} = CommonAPI.post(user, %{status: "first swamp", visibility: "public"}) {:ok, activity1} = CommonAPI.post(user, %{status: "first swamp", visibility: "public"})
{:ok, activity2} = CommonAPI.post(user, %{status: "second swamp", visibility: "public"}) {:ok, activity2} = CommonAPI.post(user, %{status: "second swamp", visibility: "public"})
activity1_id = activity1.id
activity2_id = activity2.id
{:ok, dumped_activity1_id} = FlakeId.Ecto.CompatType.dump(activity1_id)
{:ok, dumped_activity2_id} = FlakeId.Ecto.CompatType.dump(activity2_id)
ClientMock ClientMock
|> expect(:query, fn sql, params -> |> expect(:query, fn sql, params ->
assert sql =~ "SELECT id FROM pleroma_search_documents" assert sql =~ "SELECT id FROM pleroma_search_documents"
assert ["swamp", _limit, _offset] = params assert ["swamp", _limit, _offset] = params
{:ok, %{rows: [[activity2.id], [activity1.id]]}} {:ok, %{rows: [[dumped_activity2_id], [dumped_activity1_id]]}}
end) end)
Config Config
@ -161,7 +169,7 @@ defmodule Pleroma.Search.ParadeDBTest do
"pleroma_search_documents" "pleroma_search_documents"
end) end)
assert [%{id: ^activity2.id}, %{id: ^activity1.id}] = assert [%{id: ^activity2_id}, %{id: ^activity1_id}] =
ParadeDB.search(nil, "swamp", limit: 40, offset: 0) ParadeDB.search(nil, "swamp", limit: 40, offset: 0)
end end
end end