Still getting chunks from the same source document which haven't been merged.

[helenCNode]

e.g., user input: "Summarise the key methods from the Collective Intelligence Design Playbook"

The webpage: "Five examples of how collective intelligence can help us address climate change" appears twice in the retrieved sources.

It's possible that there are subtle differences in the url in the chunk metadata.