Make the indexing batch differently and more, show number indexed
This commit is contained in:
parent
365024abec
commit
ea6a6a1287
|
@ -28,17 +28,29 @@ def run(["index"]) do
|
|||
])
|
||||
)
|
||||
|
||||
Pleroma.Repo.chunk_stream(
|
||||
chunk_size = 100_000
|
||||
|
||||
Pleroma.Repo.transaction(
|
||||
fn ->
|
||||
Pleroma.Repo.stream(
|
||||
from(Pleroma.Object,
|
||||
# Only index public posts which are notes and have some text
|
||||
where:
|
||||
fragment("data->>'type' = 'Note'") and
|
||||
fragment("LENGTH(data->>'source') > 0") and
|
||||
fragment("data->'to' \\? ?", ^Pleroma.Constants.as_public())
|
||||
fragment("data->'to' \\? ?", ^Pleroma.Constants.as_public()),
|
||||
order_by: fragment("data->'published' DESC")
|
||||
),
|
||||
200,
|
||||
:batches
|
||||
timeout: :infinity
|
||||
)
|
||||
|> Stream.chunk_every(chunk_size)
|
||||
|> Stream.transform(0, fn objects, acc ->
|
||||
new_acc = acc + Enum.count(objects)
|
||||
|
||||
IO.puts("Indexed #{new_acc} entries")
|
||||
|
||||
{[objects], new_acc}
|
||||
end)
|
||||
|> Stream.map(fn objects ->
|
||||
Enum.map(objects, fn object ->
|
||||
data = object.data
|
||||
|
@ -51,10 +63,11 @@ def run(["index"]) do
|
|||
"#{endpoint}/indexes/objects/documents",
|
||||
Jason.encode!(objects)
|
||||
)
|
||||
|
||||
IO.puts("Indexed #{Enum.count(objects)} entries")
|
||||
end)
|
||||
|> Stream.run()
|
||||
end,
|
||||
timeout: :infinity
|
||||
)
|
||||
end
|
||||
|
||||
def run(["clear"]) do
|
||||
|
|
Loading…
Reference in New Issue