Update APIs with new scroll DB #141
Conversation
| @@ -4,4 +4,5 @@ updates: | |||
| directory: "/" | |||
| schedule: | |||
| interval: weekly | |||
| open-pull-requests-limit: 5 | |||
| # Disable updates | |||
| open-pull-requests-limit: 0 | |||
There was a problem hiding this comment.
Suppose to disable dependabot (auto-updagrde-dependencies) for a while 😂.
Will enable again until upgrading rust-toolchain and fix issue #107.
silathdiir
dismissed stale reviews from colinlyguo and 0xmountaintop
via
d76ab85
Copy from scroll and update in commit c3fa8f0. |
Add |
src/db/queries/chunk_query.rs
Outdated
| "SELECT start_block_number FROM {} where index = $1", | ||
| table_name::CHUNK | ||
| "SELECT start_block_number FROM {} | ||
| where index = $1", |
There was a problem hiding this comment.
Sorry for not explaining this.
Just thought, could the start (and end) chunk (and block) be marked as deleted?
Or the start_chunk_index (and others) should be updated to batch table if the start chunk is marked as deleted (Seems no).
If so, seems I should not use start_chunk_index, end_chunk_index, start_block_number and end_block_number (in this backend).
Should I fetch start_chunk_index as (if this start chunk is marked as deleted):
select min(index) from chunk where batch_hash = XXX and deleted_at is null;
WDYT? Thanks.
There was a problem hiding this comment.
If limit deleted_at is NULL in this SQL, it may be NULL here 😂 (to use start_chunk_index to get start_block_number).
And break the process to fetch batch, Or I should use above SQL for start_chunk_index.
There was a problem hiding this comment.
After discussing with @HAOYUatHZ , in commit 3b99586, fix to use chunk hash to query start and end block number, and add deleted_at IS NULL condition.
Tested on local with db/tests/test.sql.
src/db/queries/chunk_query.rs
Outdated
| @@ -60,7 +64,8 @@ pub async fn get_end_block_number_by_index( | |||
| chunk_index: i64, | |||
| ) -> Result<Option<i64>> { | |||
| let stmt = format!( | |||
| "SELECT end_block_number FROM {} where index = $1", | |||
| "SELECT end_block_number FROM {} | |||
| WHERE index = $1", | |||
…mber`, and add `deleted_at IS NULL` condition.
|
@HAOYUatHZ deployed to staging? |
I have asked @Thegaram for helping deploy this (after deploying scroll DB). Thanks 🙏 |
Cannot deploy on staging, it's incompatible with pre-sepolia DB versions. |
|
So we will be testing this PR alongside scroll-tech/frontends#415 on the Sepolia testnet? |
It would be best to test it with a local testnet first, running this version: https://github.com/scroll-tech/testnet/pull/632 |
|
Where is this currently deployed to? |
Summary
Update APIS as:
Get /api/batch?index=BATCH_INDEX
Get /api/batches?page=PAGE&per_page=PER_PAGE
Get /api/batch_blocks?batch_index=BATCH_INDEX
Get /api/chunks?batch_index=BATCH_INDEX
Get /api/chunk_blocks?chunk_index=CHUNK_INDEX
Get /api/last_batch_indexes
Get /api/search?keyword=KEYWORD
Could reference detailed specific requests and reponses in docs/openapi.spec.
Test
Update db/tests/test.sql for fake data, and test on local (http://localhost:5001/).