pangwangshu commented on issue #40476: URL: https://github.com/apache/beam/issues/40476#issuecomment-6102452989
I've opened #40502 for this. It adds an optional `content_payload_key` to `QdrantWriteConfig`. When it is set, the converter stores `EmbeddableItem.content_string` under that key in the point payload next to the metadata. The default is `None`, which writes only the metadata as today, so existing pipelines are unaffected. A key that collides with a metadata key raises `ValueError` instead of overwriting it, and an empty key is rejected at construction time. @MichaelGruschke, since you wrote the Qdrant writer in #38142, could I ask you two things before this goes further? 1. Was leaving the content out of the payload deliberate (for example to keep payloads small)? The existing tests assert `payload == item.metadata`, so I want to be sure I'm not undoing a design decision. 2. Is an opt-in `content_payload_key` the right shape, or would you prefer a different name or default, such as always writing to a fixed `"content"` key? If you have time, a review of #40502 would be much appreciated. I'm happy to change the name or the default based on your answers. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
