Improve search relevance

[sercancicek]

Description

Problem and Solution

The assignSearchRelevance function goes through each result object and assigns a score based on certain criteria, using the following formula:

        let criteriaArr = {
            "name": 10,
            "tags": 5,
            "description": 3,
            "raw_code": 2,
            "columns": 1
        };
result.overallWeight += (count * criteriaArr[criteria]);

Although this method is effective, it introduces a problem. For example, when searching for the term dm_test in the data below, the entry dm_test_total ends up with a score of 19 (10 points from name + 9 points from description), while dm_test gets only 10 points.

name	description
dm_test	Test dm
dm_test_total	Sums dm_test, multiplies dm_test, calculates dm_test

To fix this, the pull request adds an additional step to the relevance calculation process. Now, a weight value is also computed and stored for the name field. Finally, the results are sorted first by overallNameWeight and then by overallWeight.

Checklist

• I have read the contributing guide and understand what's expected of me
• I have signed the CLA
• I have run this code in development and it appears to resolve the stated issue
• I have opened an issue to add/update docs, or docs changes are not required/relevant for this PR
• I have run changie new to create a changelog entry

[matthieucan]

Wow, this looks great! Couple questions:

• Could you open an upstream PR in https://github.com/dbt-labs/dbt-docs, and reference it in FORK.md?
• Did you explore the possibility of unit testing this feature?

[jochemvandooren]

This improves search so much 🙌

Just some small comments, but it works perfectly!

[jochemvandooren]

I think we are just missing a changelog entry! other than that it looks good 🤩

[matthieucan]

What was the latest result your exploration of unit-testing this feature?

[jochemvandooren]

🚀

[matthieucan]

🚀