Synthesis for channels, inputs/outputs, T-junctions #3
Conversation
(and they're harder to read anyways)
| import org.manifold.compiler.back.microfluidics.strategies.multiphase.TJunctionDeviceStrategy; | ||
| import org.manifold.compiler.middle.Schematic; | ||
|
|
||
| public class MultiPhaseStrategySet extends TranslationStrategy { |
There was a problem hiding this comment.
Is the name MultiPhaseStrategySet too general for this code? It seems to be restricted to a combination of DropletConstraintStrategy and TJunctionDeviceStrategy. I see a few possibilities here:
0. Everything is just fine.
- Rename the class to something more specific.
- Generalize the code using type parameters (but still a pair of strategies).
- Generalize the code using a collection to hold the strategies.
There was a problem hiding this comment.
The idea for this class is that as we create new strategies for different multi-phase devices, we add them here to isolate them from the single-phase devices.
|
I have a perhaps odd habit of declaring everything final, including method parameters and local variables. I think it's a good discipline to help focus on where mutation is happening. Of course, I think it would be better if this were the default, and variables that could be re-assigned needed to be declared as such. |
|
I didn't fully grasp how Divij's question about the area of nodes and channels is handled in this code. |
| }; | ||
|
|
||
| Schematic schematic = UtilSchematicConstruction | ||
| .instantiateSchematic("test"); |
There was a problem hiding this comment.
Maybe we should change the name to something more specific instead of "test". In this case it could be tjunction_test or something. In the singlephase-electrophoresis branch I had to make this change explicitly because once there are multiple tests, it helps for this to be specified.
There was a problem hiding this comment.
How is tjunction_test more informative or specific than testTJunctionSynthesis? I don't understand your comment about why it is necessary to specify this when there are multiple tests.
There was a problem hiding this comment.
The name of the function is fine, I am referring to the label used for instantiation in line 40. The label, which in this case is "test", used to instantiate the schematic also works as the prefix for the *.smt2 that is generated by the backend. When there are a number of tests being run, it creates multiple smt2 files, when it's useful to have labels that provide more information about the specific schematic the smt2 files describes.
There was a problem hiding this comment.
Ah, I see the issue -- I'll fix this next commit.
|
Opened #5 |
…ckend-microfluidics into tjunction-synthesis Conflicts: src/test/java/org/manifold/compiler/back/microfluidics/TestMicrofluidicsBackend.java
(also, set epsilon=0 for sharp-edged T-junctions)
and worst-case droplet stuff
|
Okay, now this is ready to review; I disabled the experimental stuff and kept the version of the code that has it all in another branch. (The stuff is disabled selectively, so we can boolean-flag it back on if we want to revisit.) Closes #13. |
|
Two more things I have to do, are not run the multi-phase stuff for non-multi-phase devices and constrain input and output pressures to be strictly positive. |
|
Latest commit closes #14. |
|
This was signed off, so merging. |
Synthesis for channels, inputs/outputs, T-junctions
No description provided.