Create SEER operators for finding SEER cancer variables under one cancer

[jenniferduryea]

Originally in MarksJigsawIssues.xlsx

When first primary cancer is not index, pulling all stage or other SEER variables gives the results for all prior cancers in addition to the one of interest.  Can use co-reported, but would be nice if we can make algorithms generic instead of cancer specific.

Seems like the ultimate idea is when selecting a certain histology, confirmation, etc, we want to default to use clinical_codes that are co-reported, or have the same context_id to ensure we are getting SEER variables for the one cancer.

Related to #240

Tagging everyone for discussion for actual implementation.