Max Isom
Max Isom
https://github.com/chroma-core/chroma/pull/4264
Still happening. :( https://github.com/chroma-core/chroma/actions/runs/15546384321/job/43768574573?pr=4794 https://github.com/chroma-core/chroma/actions/runs/15571877236/job/43849306265?pr=4807
This seems to happen even with parallelization disabled on windows: https://github.com/chroma-core/chroma/actions/runs/17304176643/job/49122770358
Have you run `yarn` to install dependencies?
I planned to many years ago. Sounds good to me.
#### :package: :robot: A new release has been made for this pull request. To play around with this PR, pull `ghcr.io/museofficial/muse:pr-1245`. Images are available for x86_64 and ARM64. > Latest...
* **#3440** 👈 (View in Graphite) * **#3432** : 1 other dependent PR ([#3457](https://github.com/chroma-core/chroma/pull/3457) ) * `main` This stack of pull requests is managed by Graphite. Learn more about stacking.
I think the main problem with auto-batching during ingest is that since Chroma doesn't support transactions, there needs to be a good way to report partial failures. How does this...
> could we call `upsert` instead so that the client doesn't need to know about any of its db internals? Maybe we want to change `upsert` to return some indication...
oh, so something like ```python records: (Iterator[ID], Iterator[Document], Iterator[Embedding]) # .batch_upsert() requires IDs to be specified for ingested_ids in collection.batch_upsert(records): user_db.documents.update(ingested_ids, ingested=True) # optional, can use to avoid repeating work...