Add logging to milisiearch index and make it use desc(id)

This commit is contained in:
Ekaterina Vaartis 2021-08-16 10:18:01 +03:00 committed by FloatingGhost
parent e961cf2689
commit 41db5c8653

View file

@ -3,8 +3,9 @@
# SPDX-License-Identifier: AGPL-3.0-only # SPDX-License-Identifier: AGPL-3.0-only
defmodule Mix.Tasks.Pleroma.Search.Meilisearch do defmodule Mix.Tasks.Pleroma.Search.Meilisearch do
import Mix.Pleroma require Logger
import Mix.Pleroma
import Ecto.Query import Ecto.Query
def run(["index"]) do def run(["index"]) do
@ -12,12 +13,25 @@ def run(["index"]) do
endpoint = Pleroma.Config.get([Pleroma.Search.Meilisearch, :url]) endpoint = Pleroma.Config.get([Pleroma.Search.Meilisearch, :url])
{:ok, _} =
Pleroma.HTTP.post(
"#{endpoint}/indexes/objects/settings/ranking-rules",
Jason.encode!([
"desc(id)",
"typo",
"words",
"proximity",
"attribute",
"wordsPosition",
"exactness"
])
)
Pleroma.Repo.chunk_stream( Pleroma.Repo.chunk_stream(
from(Pleroma.Object, from(Pleroma.Object,
limit: 200,
where: fragment("data->>'type' = 'Note'") and fragment("LENGTH(data->>'source') > 0") where: fragment("data->>'type' = 'Note'") and fragment("LENGTH(data->>'source') > 0")
), ),
100, 200,
:batches :batches
) )
|> Stream.map(fn objects -> |> Stream.map(fn objects ->
@ -26,12 +40,14 @@ def run(["index"]) do
%{id: object.id, source: data["source"], ap: data["id"]} %{id: object.id, source: data["source"], ap: data["id"]}
end) end)
end) end)
|> Stream.each(fn activities -> |> Stream.each(fn objects ->
{:ok, _} = {:ok, _} =
Pleroma.HTTP.post( Pleroma.HTTP.post(
"#{endpoint}/indexes/objects/documents", "#{endpoint}/indexes/objects/documents",
Jason.encode!(activities) Jason.encode!(objects)
) )
IO.puts("Indexed #{Enum.count(objects)} entries")
end) end)
|> Stream.run() |> Stream.run()
end end