I assume that you will need to write scripts or code using a driver to connect to your Cassandra cluster query data, reshape and bulk load to MongoDB databases and collections .
I noticed that Cassandra is supporting JSON format queries:
Another way might be to export data to json or csv files and load them to MongoDB , however, here reshaping might be a challenge.
However, to leverage MongoDB architecture I would suggest to verify that the data copied from Cassandra is wisely using the document model.
Verify you are not hitting any of the known antipatterns.