aboutsummaryrefslogtreecommitdiffstats
path: root/rust
diff options
context:
space:
mode:
authorBryan Newbold <bnewbold@robocracy.org>2019-01-24 15:21:43 -0800
committerBryan Newbold <bnewbold@robocracy.org>2019-01-24 15:21:48 -0800
commitf955f66789b0078dcb973ce587d2d3b3184e73a7 (patch)
treee048502b461fdfa973d4697b3043f4a516694b2f /rust
parent206acf1b37a1a34d5338c744e17ef2035cd2db58 (diff)
downloadfatcat-f955f66789b0078dcb973ce587d2d3b3184e73a7.tar.gz
fatcat-f955f66789b0078dcb973ce587d2d3b3184e73a7.zip
allow importing contrib/refs lists
The motivation here isn't really to support these gigantic lists on principle, but to be able to ingest large corpuses without having to decide whether to filter out or crop such lists.
Diffstat (limited to 'rust')
-rw-r--r--rust/src/entity_crud.rs29
1 files changed, 25 insertions, 4 deletions
diff --git a/rust/src/entity_crud.rs b/rust/src/entity_crud.rs
index a92c45a6..73e7aa58 100644
--- a/rust/src/entity_crud.rs
+++ b/rust/src/entity_crud.rs
@@ -1964,7 +1964,26 @@ impl EntityCrud for ReleaseEntity {
if let Some(ref release_type) = entity.release_type {
check_release_type(release_type)?;
}
+ if let Some(ref abstracts) = entity.abstracts {
+ if abstracts.len() > 200 {
+ return Err(FatcatError::BadRequest(
+ "too many abstracts (sanity cap is 200)".to_string(),
+ ).into())
+ }
+ }
+ if let Some(ref refs) = entity.abstracts {
+ if refs.len() > 10000 {
+ return Err(FatcatError::BadRequest(
+ "too many refs (sanity cap is 10000)".to_string(),
+ ).into())
+ }
+ }
if let Some(ref contribs) = entity.contribs {
+ if contribs.len() > 10000 {
+ return Err(FatcatError::BadRequest(
+ "too many contributors (sanity cap is 10000)".to_string(),
+ ).into())
+ }
for contrib in contribs {
if let Some(ref role) = contrib.role {
check_contrib_role(role)?;
@@ -2160,18 +2179,20 @@ impl EntityCrud for ReleaseEntity {
}
}
- if !release_ref_rows.is_empty() {
+ // can't insert more than 65k rows at a time, so take chunks
+ for release_ref_batch in release_ref_rows.chunks(2000) {
insert_into(release_ref::table)
- .values(release_ref_rows)
+ .values(release_ref_batch)
.execute(conn)?;
}
- if !release_contrib_rows.is_empty() {
+ for release_contrib_batch in release_contrib_rows.chunks(2000) {
insert_into(release_contrib::table)
- .values(release_contrib_rows)
+ .values(release_contrib_batch)
.execute(conn)?;
}
+ // limit is much smaller for abstracts, so don't need to batch
if !abstract_rows.is_empty() {
// Sort of an "upsert"; only inserts new abstract rows if they don't already exist
insert_into(abstracts::table)