Add top_k_join parameter to JoinDocuments.run

[adri1wald]

Implements a solution to #2059

Proposed changes:

JoinDocuments.run takes an optional parameter top_k_join that overrides the instance top_k_join of JoinDocuments.

This allows top_k_join to be modified when runing a pipeline, e.g.

p = Pipeline()
p.add_node(component=retriever_1, name="R1", inputs=["Query"])
p.add_node(component=retriever_2, name="R2", inputs=["Query"])
p.add_node(component=JoinDocuments(join_mode="concatenate"), name="Join", inputs=["R1", "R2"])
results = p.run(query=query, params={ 'Join': { 'top_k_join': 1 } })
# len(results["documents"]) == 1

Status (please check what you already did):

• First draft (up for discussions & feedback)
• Final code
• Added tests
• Updated documentation

[ZanSara]

Hello @adri1wald, neat work! Thanks a lot for the test especially! I left only a comment with an idea to extend it a bit, but this is all looking ready to merge.

Do you think is there anything else to add to this PR? Otherwise once you modify the test and CI is green I'm going to merge 🙂

[adri1wald]

Hello @adri1wald, neat work! Thanks a lot for the test especially! I left only a comment with an idea to extend it a bit, but this is all looking ready to merge.

Do you think is there anything else to add to this PR? Otherwise once you modify the test and CI is green I'm going to merge 🙂

Hey @ZanSara, great idea to extend the test with two different top_k_join - have made the change!

I think it might be worth adding some info in the documentation? But not sure where the best place is to add it.

[ZanSara]

No problem for the documentation, the API docs should update automatically based on the docstring. In any case I will check once this PR is merged, and if necessary I can add a note myself later 👍