chore: dedupe project entries in oss-directory

[ccerv1]

What is it?

A few duplicates seem to have popped up in OSSD after recent bulk imports:

with cte as (
  select
    artifact_id,
    array_agg(distinct(project_name)) as projects,
    count(distinct project_name) as num_projects
  from `oso.artifacts_by_project_v1`
  where artifact_source = 'GITHUB'
  group by 1
  order by 3 desc
)
select * from cte where num_projects > 1

Here is an export from this query.

[ccerv1]

opensource-observer/oss-directory#595