ProgressiveProofer refactor 1/N: ThreatMetrix

[matthinz]

🛠 Summary of changes

Ahead of integrating a new vendor into the ProgressiveProofer, I'm going to do a series of PRs that extract code for individual vendors out into separate files/classes. I'm calling these "plugins" because I think that will ultimately describe the kind of architecture we're going for here, but for now these are basically just the old proof_with_X_if_needed methods extracted out into a new class with a call method.

Right now, each plugin's call method lays out exactly what its dependencies are. Each plugin also gets its own spec file, mainly focused on these primary scenarios:

• Remote unsupervised proofing
• In-person proofing (residential address == id address)
• In-person proofing (residential address != id address)

In some cases I am pulling spec code out of the existing ProgressiveProofer spec, and in some cases I am modifying it slightly to ensure it addresses the above scenarios.

This PR implements the first (and simplest) "plugin", which does the ThreatMetrix Session Query API call. There are not significant differences in implementation for the Remote and IPP flows here, and there's not really any "new" code being introduced--this is mostly moving existing code around.

See also

• The "How Login.gov does identity resolution and validation" section of the Identity Verification Business Logic doc, which is essentially an breakdown explanation of what exactly the ProgressiveProofer is doing.
• This RFC about pulling the ProgressiveProofer apart. This work corresponds roughly to the "Phase One" identified in the RFC.

📜 Testing Plan

Provide a checklist of steps to confirm the changes.

• Intend to commit fraud (signaling this using the Mock device profiling behavior box on the SSN screen), run through IdV, and be successfully caught
• Be pure of heart (also signaling this using the Mock device profiling behavior box on the SSN screen) and complete IdV successfully.

[matthinz]

This diff is a little messy, but my goal is to, at the end of this refactor, have two top level contexts here: remote unsupervised proofing and in-person proofing.

[mdiarra3]

Just a heads up. Will be updating the config to add a "policy" attribute since we have decided for Authentication we will have a different policy. But otherwise. Taking a look this shouldn't affect our implementation much. may be a merge conflict depending on whats merged first.