default_transforms() Missing cosine_similarity Transform for Docs with Token Count 101–500

[rgrizzo-linksmt]

At:
[default.py#L157](

ragas/src/ragas/testset/transforms/default.py

Line 157 in 2bc29a2

ner_overlap_sim,

),

you forgot to add the cosine_sim_builder transform to the list of default transforms. While you instantiated the cosine_sim_builder object, you did not include it in Parallel(cosine_sim_builder, ner_overlap_sim), as you did in:

[default.py#L126](https://github.com/explodinggradients/ragas/blob/2bc29a2b8358ddb6b167fdf7ab0518ad9371463c/src/ragas/testset/transforms/default.py#L126C13-L126C59).

This omission might impact the number of relationships created in the knowledge graph.

Ragas version: ragas==0.2.13
Python version: Python 3.10.12

[Vidit-Ostwal]

Can you share the problem you are facing?

[rgrizzo-linksmt]

The default_ transforms function defined at src/ragas/testset/transforms/default.py has a problem with handling transforms for documents with 101-500 tokens.

The code divides the transforms configurations based on the document's token count. Several transforms are instantiated when the "101-500" token count bins the first quartile (Q1, among them the cosine_sim_builder. While cosine_sim_builder is correctly instantiated (line 139), it's then not included in the list of transforms that are actually returned (line 153).

It appears that cosine_sim_builder was likely unintentionally omitted from the returned transforms list. The intended behavior should probably mirror how ner_overlap_sim is handled (line 120), where cosine_sim_builder is instantiated and added to the returned list. The current code effectively instantiates cosine_sim_builder but then discards it. This omission might impact the number of relationships created in the knowledge graph.

Hoping this can help you understand the problem I'm facing!

[Vidit-Ostwal]

Accha got this,
In src/ragas/testset/transforms/default.py
Instead of line 157 to be ner_overlap_sim,
this should be Parallel(cosine_sim_builder, ner_overlap_sim).

This makes sense.

@rgrizzo-linksmt I think you should raise a PR for this.