add recipes to source packaging tests to avoid undesired transmutation

[nuclearkatie]

I figured out why tests in #613 failed. After playing with it a bit, I kept getting results where the number of packaged resources was twice the number of transactions, which didn't make sense (every packaged resource should only be packaged in order to fulfill one transaction).

Then I realized that my tests didn't include an outrecipe, and the Mock Sink in the Mock Sim was for some reason requesting a different mock composition. So, immediately before trading, the Source transmuted the composition to the request composition, which creates a new line in the resource table, because the composition id (QualId I think) changed.

This is exactly the kind of extra and confusing resource database entries I mentioned above!

However, I still think it makes sense that resource should only be transmuted after they are packaged and ready to go, because we don't want to do transmuting until we're sure that the packaging will work. Which it normally will, but this keeps coming back to the challenge of requesting agents accepting partial-sized bids, which creates the chance of a failed trade by being an unpackagable amount.

In this case, I solved the issue by creating an outrecipe in my tests.

[github-actions]

Build Status Report - 8126c14 - 2024-07-18 13:00:44 -0500

Build FROM cyclus_20.04_apt/cyclus:latest

• Cycamore: Success ✅
• Cymetric: Success ✅

Build FROM cyclus_20.04_apt/cyclus:stable

• Cycamore: Failure ❌
• Cymetric: Skipped due to upstream failure ⚠️

Build FROM cyclus_20.04_conda/cyclus:latest

• Cycamore: Success ✅
• Cymetric: Success ✅

Build FROM cyclus_20.04_conda/cyclus:stable

• Cycamore: Failure ❌
• Cymetric: Skipped due to upstream failure ⚠️

Build FROM cyclus_22.04_apt/cyclus:latest

• Cycamore: Success ✅
• Cymetric: Success ✅

Build FROM cyclus_22.04_apt/cyclus:stable

• Cycamore: Failure ❌
• Cymetric: Skipped due to upstream failure ⚠️

Build FROM cyclus_22.04_conda/cyclus:latest

• Cycamore: Success ✅
• Cymetric: Success ✅

Build FROM cyclus_22.04_conda/cyclus:stable

• Cycamore: Failure ❌
• Cymetric: Skipped due to upstream failure ⚠️

[gonuke]

Thanks @nuclearkatie - this looks like it fixes the problem. Do you think it reduces the generality/breadth of the testing? should some version of this work without an outrecipe?

[nuclearkatie]

The only difference if I didn't add an outrecipe is that the transactions would stay the same but the number of packaged resource entries in the database would always be 2 * transactions. I didn't think that warranted an additional test.

Also in some of the demonstration input files I've made to run full, real simulations, the requested composition didn't necessarily change, which is why I added the bit about preventing transmutation if the requested composition is the same as the existing composition (I have to admit I don't really understand mock compositions here, and when they might be the same and when they might be different). So I'm not even sure this is would be an issue in full simulations.

Plus, it should show up on the resource table as the QualId having changed if transmutation occurs, which is a visible flag that something did actually change to warrant a new resource line