[ENH] Min compaction size (chroma-core#2346)

## Description of changes

*Summarize the changes made by this PR.*
 - Improvements & Bug fixes
- I overhauled the property tests for the log service. There were bugs
and also it was underspecc'ed
- The main bug was the the len() being used was over collectionData not
of the collectionData[C] in question.
- I added invariants using rapids "" action, which is the way to add
invariants.
- The invariants check the log is as we expect and also the the
collections we expect get returned
	 - I actually use the enumerationoffset in the model now
	 - The model will _actually_ purge the log it has.
	 - The prop test will check the fields of its records for equality
 - New functionality
- This PR introduces the "Min compaction size" argument on
GetCollectionsToCompact, making it so that the compactors can skip logs
with only a handful of entries.
	 - This min compaction size is config for the compactor
	 - I updated the rust in memory log to respect the min size.
- The property test is extended to check that min_compaction_size works.

Other notes:
- The log service types all use int where they should be uint. I added a
cleanup task to go convert the incorrect types. The new type I introduce
is correct.

## Test plan
*How are these changes tested?*
- [x] Tests pass locally with `pytest` for python, `yarn test` for js,
`cargo test` for rust

## Documentation Changes
None

go/database/log/queries/queries.sql

@@ -15,6 +15,7 @@ with summary as (
  select r.collection_id, r.offset, r.timestamp, row_number() over(partition by r.collection_id order by r.offset) as rank
  from record_log r, collection c
  where r.collection_id = c.id
+ and (c.record_enumeration_offset_position - c.record_compaction_offset_position) >= sqlc.arg(min_compaction_size)
  and r.offset > c.record_compaction_offset_position
 )
 select * from summary

go/pkg/log/repository/log.go

@@ -3,10 +3,11 @@ package repository
 import (
  "context"
  "errors"
+ "time"
+
  log "github.com/chroma-core/chroma/go/database/log/db"
  "github.com/jackc/pgx/v5"
  "github.com/jackc/pgx/v5/pgxpool"
- "time"
 )
 
 type LogRepository struct {
@@ -76,8 +77,8 @@ func (r *LogRepository) PullRecords(ctx context.Context, collectionId string, of
  return
 }
 
-func (r *LogRepository) GetAllCollectionInfoToCompact(ctx context.Context) (collectionToCompact []log.GetAllCollectionsToCompactRow, err error) {
- collectionToCompact, err = r.queries.GetAllCollectionsToCompact(ctx)
+func (r *LogRepository) GetAllCollectionInfoToCompact(ctx context.Context, minCompactionSize uint64) (collectionToCompact []log.GetAllCollectionsToCompactRow, err error) {
+ collectionToCompact, err = r.queries.GetAllCollectionsToCompact(ctx, int64(minCompactionSize))
  if collectionToCompact == nil {
  collectionToCompact = []log.GetAllCollectionsToCompactRow{}
  }