Extend Functionality to Use the Presolver Plugin and Add a Tutorial #1076
Conversation
…s in the knapsack model
…for fixed and aggregated variables
…, and example shiftbound.py
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## master #1076 +/- ##
==========================================
+ Coverage 54.13% 54.70% +0.56%
==========================================
Files 22 24 +2
Lines 5045 5327 +282
==========================================
+ Hits 2731 2914 +183
- Misses 2314 2413 +99 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
There was a problem hiding this comment.
Pull Request Overview
This PR extends PySCIPOpt to support custom presolver plugins by adding wrapper functions, tests, a working example, and documentation. The key enhancement enables users to implement custom presolvers in Python by subclassing the Presol class.
- Added
isActive()method for variables to check if they are active (not fixed or aggregated) - Added
aggregateVars()method to aggregate two variables with custom coefficients - Created a complete Shiftbound presolver example that replicates SCIP's boundshift.c logic
Reviewed Changes
Copilot reviewed 7 out of 7 changed files in this pull request and generated 3 comments.
Show a summary per file
| File | Description |
|---|---|
| src/pyscipopt/scip.pxi | Added isActive() and aggregateVars() wrapper methods for variables and models |
| src/pyscipopt/scip.pxd | Added C function declarations for SCIPvarIsActive and SCIPaggregateVars |
| tests/test_vars.py | Added basic test for isActive() method on newly created variables |
| tests/test_aggregate_vars.py | Added comprehensive tests for aggregateVars() functionality |
| examples/finished/shiftbound.py | Complete presolver example implementing boundshift logic |
| docs/tutorials/presolver.rst | Tutorial documentation explaining presolver plugin usage |
| CHANGELOG.md | Updated changelog with new features |
Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.
Co-authored-by: Copilot <[email protected]>
| if val > 0 and val >= scip.infinity(): | ||
| return scip.infinity() | ||
| if var.vtype() != "CONTINUOUS": | ||
| return scip.feasCeil(val) |
There was a problem hiding this comment.
I might be wrong, but shouldn't this be rounding to the nearest integer?
There was a problem hiding this comment.
Yes, you are right. It's actually only adjustedLB() and forgot to include adjustedUB(). I guess that is a good example for why we want to wrap functions instead of copying their functionality.
| for var in reversed(vars): | ||
| if var.vtype() == "BINARY": | ||
| continue |
There was a problem hiding this comment.
shouldn't this be an assert? I assume all binary variables are skipped from earlier.
| ) | ||
| # Aggregate old variable with new variable: | ||
| # x = y + lb (no flip), or |
There was a problem hiding this comment.
how about writing this as x - y = lb to be the same as what is passed to SCIPaggregateVars?
| assert not infeasible | ||
| assert aggregated | ||
| # model should stay consistent | ||
| model.optimize() |
There was a problem hiding this comment.
Could you please add assertion to check the bounds of the variables? also an assertion for the final status and the objective value of the model would be nice :)
|
|
||
| def aggregateVars(self, Variable varx, Variable vary, scalarx=1.0, scalary=-1.0, rhs=0.0): | ||
| """ | ||
| Aggregate varx and vary: calls SCIPaggregateVars and returns |
There was a problem hiding this comment.
Could you please add more info on what SCIPaggregateVars does? a formula would be very helpful here so that one wouldn't need to go the C-API documentation.
|
To also answer your questions:
I'm happy with the current test, but if you have the time yes, assertions in your presolver test on the aggregated variables would be nice.
I don't have a lot of experience with presolvers but I would say one should wrap the methods when a use-case/feature request pops up. If you see anything that could be useful to "preemptively" wrap I'd be happy to review it in future PRs :D |
|
I just realised that, while I was working on the documentation, the implementation of presol_boundshift.c had changed. Besides slight modifications to the sequence, it now no longer uses SCIPvarIsActive(). I'll check and align my documentation again but prioritise to make the presolver documentation/example self-contained/-explanatory. I'll also take a look again on the tests to capture your suggestions. Lastly, I think I will crawl through the other presolvers to see which SCIP functions are most frequently used and then wrap them (as part of future PRs). My goal is anyway to better grasp SCIP's source code, so it'll be a good exercise. Just let me know about any other feature that I could take a look at (ofc, I'll also take a look at the currently open issues). Thank you for your feedback :) |
3 participants
To clarify and document the usage of PySCIPOpt's presolver plugin, a working example and a tutorial were added. They required wrappers to be implemented, too. The wrapper functions are supplemented with tests. The example and tutorial replicate the Boundshift presolver.
However, some open issues remain: