New GAMS solver interface, writer, and solution loader. by AnhTran01 · Pull Request #3683 · Pyomo/pyomo

AnhTran01 · 2025-08-07T18:07:31Z

Fixes NA

Summary/Motivation:

This PR introduced a new GAMS solver interface that uses the LinearRepnVisitor to enable coefficient gathering when writing out the scalar GAMS model in gms format.

Changes proposed in this PR:

New GAMS solver interface in pyomo.contrib.solver.solvers
New GAMS writer in pyomo.repn.plugins
New GAMS solution loader in pyomo.contrib.solver.solvers

Legal Acknowledgement

By contributing to this software project, I have read the contribution guide and agree to the following terms and conditions for my contribution:

I agree my contributions are submitted under the BSD license.
I represent I am authorized to make the contributions and grant the license. If my employer has rights to intellectual property that includes these contributions, I represent that I have received permission to make contributions and grant the required license on behalf of that employer.

…ittest and coverage test

AnhTran01 · 2025-08-07T18:07:57Z

@boxblox
@abhosekar

mrmundt · 2025-08-07T18:11:30Z

@AnhTran01 - Please make sure to run black on your changes!

mrmundt

Immediate piece of feedback, recognizing that I have not thoroughly looked over anything - there are no tests anywhere, and there absolutely will need to be some.

Thanks for this, though, I am so excited!!

…riting process. Added WIP unittest for the new solver interface

…ion error

… and writing to gms.

pyomo/repn/plugins/gams_writer_v2.py

… constraint). Fixed bug of load_solution that contains constraint id in addition to variable id

…e_gdx_results

…arsing solver_options

jsiirola

(Some initial comments, mostly on configuration)

pyomo/contrib/solver/solvers/gams.py

jsiirola · 2025-08-15T21:42:22Z

pyomo/contrib/solver/solvers/gams.py

+        if config.logfile is not None:
+            config.logfile = os.path.abspath(config.logfile)
+
+        config.writer_config.put_results_format = 'gdx' if gdxcc_available else 'dat'


This should probably only overwrite the configuration if it is still None (that way users can set / override the interface)

…lable() and version()

…are no longer valid.

codecov · 2026-02-06T19:52:06Z

Codecov Report

❌ Patch coverage is 89.10162% with 74 lines in your changes missing coverage. Please review.
✅ Project coverage is 89.46%. Comparing base (f2d23a5) to head (e32b768).
⚠️ Report is 32 commits behind head on main.

Files with missing lines	Patch %	Lines
pyomo/contrib/solver/solvers/gams.py	89.29%	32 Missing ⚠️
pyomo/repn/plugins/gams_writer_v2.py	89.34%	31 Missing ⚠️
pyomo/contrib/solver/solvers/gms_sol_reader.py	87.35%	11 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #3683      +/-   ##
==========================================
- Coverage   89.48%   89.46%   -0.03%     
==========================================
  Files         904      907       +3     
  Lines      105385   106064     +679     
==========================================
+ Hits        94299    94885     +586     
- Misses      11086    11179      +93

Flag	Coverage Δ
builders	`28.96% <18.40%> (-0.08%)`	⬇️
default	`83.61% <89.10%> (?)`
expensive	`35.41% <18.40%> (?)`
linux	`86.71% <78.05%> (-2.50%)`	⬇️
linux_other	`86.71% <78.05%> (-0.06%)`	⬇️
oldsolvers	`29.62% <18.40%> (-0.08%)`	⬇️
osx	`82.88% <78.05%> (-0.04%)`	⬇️
win	`84.94% <77.31%> (-0.07%)`	⬇️
win_other	`84.94% <77.31%> (-0.07%)`	⬇️
xpress95	`?`

Flags with carried forward coverage won't be shown. Click here to find out more.

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

boxblox · 2026-02-06T20:11:33Z

seems like codecov is not ideal. most of the missed lines are when gdx results are being read in with our gamsapi. instead of with a dat file. what do you think? we'd need to add another test dependency to the pipeline.

mrmundt · 2026-02-09T23:15:33Z

seems like codecov is not ideal. most of the missed lines are when gdx results are being read in with our gamsapi. instead of with a dat file. what do you think? we'd need to add another test dependency to the pipeline.

Oop, sorry, I didn't see this. The coverage is actually fine. 89% is more than okay.

boxblox · 2026-02-10T13:31:41Z

seems like codecov is not ideal. most of the missed lines are when gdx results are being read in with our gamsapi. instead of with a dat file. what do you think? we'd need to add another test dependency to the pipeline.

Oop, sorry, I didn't see this. The coverage is actually fine. 89% is more than okay.

Sounds good, thanks for the follow up. Is there anything I can do to help with this?

blnicho

@boxblox I'm going to publish my comments/questions in batches as I'm reviewing since we're trying to move quickly on this to get it into the release. Here is the first set.

blnicho · 2026-02-11T21:07:17Z

pyomo/contrib/solver/solvers/gms_sol_reader.py

+        self.primals: List[float] = []
+        self.duals: List[float] = []
+        self.var_suffixes: Dict[str, Dict[int, Any]] = {}
+        self.con_suffixes: Dict[str, Dict[Any]] = {}
+        self.obj_suffixes: Dict[str, Dict[int, Any]] = {}
+        self.problem_suffixes: Dict[str, List[Any]] = {}
+        self.other: List(str) = []


Do we still need to use typing.Dict and typing.List here or can we use the built-in Python dict and list?

if you want type hinting to catch that the dict structure should be a str key and a list value, then yes, you will need to keep these as is (same logic for the List) im fixing a typo on the List(str) should be List[str] (w/ brackets)

blnicho · 2026-02-11T21:46:45Z

pyomo/contrib/solver/solvers/gms_sol_reader.py

+        else:
+            for sym, obj in self._gms_info.var_symbol_map.bySymbol.items():
+                level = self._gdx_data[sym][0]
+                if obj.parent_component().ctype is Var:


Can var_symbol_map contain Pyomo component types other than Var?

I will need to let @AnhTran01 give more detail here, but with the simple models that we have in the test library only Var components appear in this dict.

@blnicho The var_symbol_map only contains Pyomo component type Var (from the internal data of _gms_info below).

In that case we can probably remove the ctype check on L61.

ok, I have removed this.

pyomo/contrib/solver/solvers/gms_sol_reader.py

Co-authored-by: Bethany Nicholson <blnicho@users.noreply.github.com>

pyomo/contrib/solver/solvers/gms_sol_reader.py

blnicho · 2026-02-12T05:28:46Z

pyomo/repn/plugins/gams_writer_v2.py

+                RangeSet,
+                Port,
+            },
+            targets={Suffix, SOSConstraint, Objective},


I don't see any code for handling Suffix or SOSConstraint components. I suspect this writer will silently ignore any of these components declared on the model which seems like a bug. We should check for these components and log a warning or throw an error if any are encountered.

@jsiirola would removing Suffix and SOSConstraint from the target argument be sufficient for getting the writer to treat these component types as "unknown" and throw an error?

throwing an appropriate error seems fine with me. Thank you for the help/guidance in crafting this properly.

pyomo/repn/plugins/gams_writer_v2.py

Co-authored-by: Bethany Nicholson <blnicho@users.noreply.github.com>

jsiirola · 2026-02-13T15:12:48Z

pyomo/contrib/solver/solvers/gams.py

+    def available(
+        self, config: Optional[GAMSConfig] = None, recheck: bool = False
+    ) -> Availability:


We have been re-working how we want to use / think about available(). For solvers that rely on an external executable, we are caching a dict that maps executable path to the version information. That allows the cache to be a class attribute (so we don't have to keep re-checking), and to be robust to users pointing the solver to a new executable. See ipopt.py for the current reference implementation

ok, I am taking a look now

yeah, i see now. good to not keep calling a subprocess if .solve() is in a loop. hopefully fe7b43f fixes this

fixing the errors now... stay tuned

apologies... my tests are are now passing locally

blnicho

Updating the copyright headers for the new files in this PR following #3846

pyomo/contrib/solver/solvers/gams.py

pyomo/contrib/solver/solvers/gms_sol_reader.py

pyomo/contrib/solver/tests/solvers/test_gams.py

pyomo/repn/plugins/gams_writer_v2.py

Added new GAMS solver interface, writer, and solution loader. WIP: un…

1177492

…ittest and coverage test

Norman_Tran and others added 2 commits August 7, 2025 14:14

Ran black -S -C on commited files

2385654

Typo!

b36086f

mrmundt requested changes Aug 7, 2025

View reviewed changes

Norman_Tran and others added 14 commits August 7, 2025 17:02

Added ignore checking for constant in the objective function during w…

e91e480

…riting process. Added WIP unittest for the new solver interface

Added support for add_options for multiple solver_options

b060096

typo

96ba646

ran black formatter

b0f16a3

added add_options in the solver_interface

dde656c

Fixed bug with unable to create tmpdir. Updated test_gams_v2.py

3476f72

Merge branch 'main' into GAMS_new_solver_interface

ddfd98c

Added handling single bounded variable case (None,#) or (#, None)

eb0a3db

Match solver interface and writer configs to prevent implicit definit…

c6d2d01

…ion error

Added new gams solver interface for test_all_solvers_list()

2a6dcb7

Fixed bug of handling valid ub when writing pyomo expression

b6ead27

Fixed bug when using ComponentMap to extract duals in gms_sol_reader.py

89f80e0

Corrected handling of solution_loader. Implemented bounded constraint…

b0c3450

… and writing to gms.

Added offdigit based on error log message suggestion

a9df1d7

boxblox reviewed Aug 14, 2025

View reviewed changes

pyomo/repn/plugins/gams_writer_v2.py Show resolved Hide resolved

mrmundt and others added 4 commits August 15, 2025 07:47

Merge branch 'main' into GAMS_new_solver_interface

e04c89f

Fixed bug of extracting constraint dual when alias exits (for bounded…

518758e

… constraint). Fixed bug of load_solution that contains constraint id in addition to variable id

Added aliases as part of parse_dat_results to be consistent with pars…

a2ef9cc

…e_gdx_results

Removed code duplication between the solver interface and writer by p…

c33095d

…arsing solver_options

jsiirola reviewed Aug 18, 2025

View reviewed changes

Norman_Tran and others added 4 commits August 19, 2025 14:31

Revised code based on John's review. WIP: Handling rehash within avai…

f609ca2

…lable() and version()

Re-added config into available and version, and deprecated test that …

ec5aa6a

…are no longer valid.

Merge branch 'main' into GAMS_new_solver_interface

e760a1c

Sanity check config.writer_config.put_results_format without setdefault

7bb817b

mrmundt added the AT: PRE-TEST INSPECTED label Feb 6, 2026

pyomo-autotest removed the AT: PRE-TEST INSPECTED label Feb 6, 2026

remove unnecessary code comments (ostream)

e32b768

mrmundt added the AT: PRE-TEST INSPECTED label Feb 6, 2026

pyomo-autotest removed the AT: PRE-TEST INSPECTED label Feb 6, 2026

blnicho reviewed Feb 11, 2026

View reviewed changes

boxblox and others added 8 commits February 11, 2026 20:22

fix typo in type hint syntax

7d6907c

Update pyomo/contrib/solver/solvers/gms_sol_reader.py

c78e5df

Co-authored-by: Bethany Nicholson <blnicho@users.noreply.github.com>

black formatting and remove unreachable code

71ac96b

black formatting again

d6a25b6

Update pyomo/contrib/solver/solvers/gms_sol_reader.py

8fe1401

Co-authored-by: Bethany Nicholson <blnicho@users.noreply.github.com>

simplifications and remove more unreachable code

714071a

black

e639869

Merge branch 'main' into GAMS_new_solver_interface

a6485a8

blnicho reviewed Feb 12, 2026

View reviewed changes

pyomo/contrib/solver/solvers/gms_sol_reader.py Outdated Show resolved Hide resolved

Fix typo in gms_sol_reader.py

f5002bd

blnicho reviewed Feb 12, 2026

View reviewed changes

boxblox and others added 2 commits February 12, 2026 10:15

Update pyomo/repn/plugins/gams_writer_v2.py

2ab3f6e

Co-authored-by: Bethany Nicholson <blnicho@users.noreply.github.com>

remove ctype check

2837299

jsiirola reviewed Feb 13, 2026

View reviewed changes

boxblox and others added 4 commits February 13, 2026 11:39

update version caching logic and clean up some unused imports

fe7b43f

.available() fixups

328b61a

adjust tests for new .available method

dca18e7

Merge branch 'main' into GAMS_new_solver_interface

840c5ae

blnicho reviewed Feb 14, 2026

View reviewed changes

Update copyright headers in new files

1d85594

Conversation

AnhTran01 commented Aug 7, 2025

Fixes NA

Summary/Motivation:

Changes proposed in this PR:

Legal Acknowledgement

Uh oh!

AnhTran01 commented Aug 7, 2025

Uh oh!

mrmundt commented Aug 7, 2025

Uh oh!

mrmundt left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

jsiirola left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codecov bot commented Feb 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Codecov Report

Uh oh!

boxblox commented Feb 6, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

mrmundt commented Feb 9, 2026

Uh oh!

boxblox commented Feb 10, 2026

Uh oh!

blnicho left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

boxblox Feb 12, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

codecov bot commented Feb 6, 2026 •

edited

Loading

boxblox commented Feb 6, 2026 •

edited

Loading

boxblox Feb 12, 2026 •

edited

Loading