Contextual Enrichment

[Ayoubimad]

Hi everyone,

I've encountered an issue related to contextual enrichment when ingesting documents using custom chunks.

This works (enrichment is triggered):
client.documents.create(raw_text=text_of_the_document)

This does NOT trigger enrichment:
client.documents.create(chunks=list_of_custom_chunks)

When raw_text is provided, the system handles chunking internally and contextual enrichment is correctly applied. However, when supplying a pre-chunked list, enrichment does not seem to be invoked.

Logs for comparison:

Using raw_text:
2025-05-09 15:15:36 - INFO - Initializing text splitter with method: ChunkingStrategy.RECURSIVE
2025-05-09 15:15:40 - INFO - Successful ingestion for document_id: 23b151b0-d1c1-55bd-b4be-b83e52b4017e, with vector count: 55
2025-05-09 15:15:40 - INFO - Enriching document with contextual chunks
2025-05-09 15:15:56 - INFO - Completed enrichment of 55 chunks for document 23b151b0-d1c1-55bd-b4be-b83e52b4017e

Using chunks:
2025-05-09 15:15:56 - INFO - 172.17.0.1:58048 - "POST /v3/documents HTTP/1.1" 202
2025-05-09 15:15:57 - INFO - Successful ingestion for document_id: 8881d82f-b835-55b3-94e3-24625d28945a, with vector count: 13
2025-05-09 15:15:57 - INFO - 172.17.0.1:58048 - "POST /v3/documents HTTP/1.1" 202

Is this the intended behavior?
If so, is there a recommended way to enable enrichment when using pre-chunked input?

Thanks!