Test cases to see if chunking API exports is working properly

[dcarsey]

MORE IMPORTANT TEST CASES

• At least >2 proposals to test if pagination works

• Special characters in PI names

• Special characters in Proposal name / description about project

• Proposals that are HEAL - non-HEAL

• If some fields are intentionally left null

  Spaces OR blank?

• Look at actual REDCap data to see how it comes back

• Sometimes it’s an empty string

• Need a more targeted approach

• Look at source code and see which fields it uses that does not check for null and write cases for those fields / run test case and show there’s a bug and fix it

OR

• more time required - systematically take out each field then generate a bunch of proposals w/ each proposal w/ 1 field taken out and then run it

• Take one that’s complete and write a json script to remove a different field each time

• Shruti can try it and use Hao as a resource

• Checking logic to make sure the most recent data is captured / in the right state

• Anything that’s already in one of the github issues we should do

LESS IMPORTANT TEST CASES

• Proposal ID to be continuous

• Proposal ID to handle large gaps

• Proposal IDs that are a max integer

• Spacing in PI names

• Very large proposal ID (larger than maximum integer)

• Non-proposal proposals

• • Demos
• • Pilots

• Funding column - can it handle numbers AND text ?

• Funding institute >1 field

• Removed services - check that it captures this

• Proposal status - esp. Withdrawn by PI

• Dawn added after meeting: check resubmission

FIELDS THAT USE LOGIC

grant resubmission that was funded 2nd time thru:

• funding = 2 or 3
• AND new funding source = 4 (not funded)
• AND they do plan to resubmit
• look at updated_funding_status
• AND value in updated funding status = 1 OR new funding source =1

there are 6 fields to look at for funding source (and there may be >1 funding source for a proposal)

• look at sources_1
• DON'T look at funding source
• Look at institute center
• Look at Institute center 2
• Look at institute center 3

For OTHER funding look at

• other funding 2
• other funding 3
• other funding 4
• other funding 5
• • (4 and 5 are descriptive / text fields: not standardized)

Sometimes they update their funding source so also look at

• ins_center
• ins center 2
• ins center 3

Then for other look at

• OFS 1 (Other Funding Source)
• ofs 2
• ofs 3
• ofs 4 and ofs 5 and ofs 6 are all the 2nd spot for other potential funding sources

We have 2 main sources:

• proposal intake form
• network support status additional data elements (info from PL)

AND OTHER LOGIC SITUATIONS THAT TRIP US UP FROM TIME TO TIME

• updated funding status=1
• new funding source=1
• Using protocol status removes the need to use logic for determining the status of a proposal.
• actualGrantAwardDate maps to release_of_funds_2

[rajshruti18]

Created a synthetic dataset that covers following use cases:

• At least >2 proposals to test if pagination works
• Special characters in PI names
• Special characters in Proposal name / description about project
• Proposals that are HEAL and non-HEAL
• Proposal ID to be continuous
• Proposal ID to handle large gaps

Dataset: https://github.com/RENCI/tic-map-pipeline-script/blob/master/syntheticDataset.json

[krobasky]

I'm not sure if the synthetic data is useful here @rajshruti18 but it would be good to include all the described use cases in the synthetic dataset, as well as code to actually test functionality with that data.

[krobasky]

related to #433

[dcarsey]

@dcarsey add google doc here