Issue 5356 modelica measures squashed #5557

kbenne · 2025-12-16T03:03:15Z

Summary

Introduces a full Modelica workflow stack—ModelicaFile ownership/AST management, generated ANTLR parser artifacts, SWIG bindings, and OSWorkflow plumbing—so ModelicaMeasure authors can
parse, edit, and run Modelica directly within OpenStudio.
Adds ModelicaMeasure and ModelicaParameters, extends OSRunner + WorkflowJSON (seed file metadata, snapshotting, last result path), and captures Modelica runs in run/stdout-modelica
alongside the .mat result file.
Ships an end-to-end example under resources/Examples/compact_osw/, thoroughly documented in MODELICA_README.md—start there for an overview of the workflow, the SetModelicaZones measure,
and the sample Modelica libraries you can run via the CLI.

Key Areas to Review

resources/Examples/compact_osw/
- MODELICA_README.md: walkthrough for running the Modelica OSW, inspecting results, and troubleshooting.
- compact_modelica.osw, modellica/OpenStudioExample/*, and measures/SetModelicaZones/ (measure code, README, tests) show how the workflow is wired together.
src/modelica/
- ModelicaFile., ModelicaFileImpl., the generated ANTLR sources, Modelica.i, ModelicaAPI.hpp, and src/modelica/CMakeLists.txt implement parsing, ownership, and bindings.
src/measure/
- ModelicaMeasure., ModelicaParameters., and the updated OSRunner.* surface Modelica parameters and last-result-file info to measures.
src/utilities/filetypes/WorkflowJSON* & src/workflow/*
- Workflow JSON now carries modelicaSeedFile; OSWorkflow tracks snapshots and adds runModelicaMeasures/runModelica stages.
- RunModelicaMeasures.cpp applies Modelica measures, and RunModelica.cpp builds the .mos, launches OpenModelica, parses tagged output, and records the result file.
Supporting changes across CMakeLists.txt, Ruby bindings (Measure.i, InitRubyBindings.cpp, ruby/module/CMakeLists.txt), etc., to integrate the new module.

Notes / Follow-ups

Only OpenModelica is supported at the moment; Dymola hooks can follow once we can validate them.
There is still no automated CLI regression for the Modelica OSW; the environment setup is involved.

- add a full Modelica parsing stack (ModelicaFile ownership/AST management, ANTLR grammar artifacts, SWIG interface, and CLI driver) so measures can inspect and edit Modelica source safely - expose ModelicaMeasure/ModelicaParameters, extend OSRunner + WorkflowJSON to pass parameters and surface the last simulation result, and wire everything through OSWorkflow with runModelica/runModelicaMeasures steps - ship a compact Modelica example workflow, measure, Modelica libraries, and README to demonstrate running the CLI end-to-end, while tightening logging, bindings, and CMake integration to make the feature production-ready

jmarrec · 2026-01-06T10:25:25Z

ruby/engine/CMakeLists.txt

+  # KSB: This is an abomination. the basic idea is that we need a full path to the crypto library
+  # so that we can use -force_load
+  # It would really be nice to have a simpler solution, but this appears to work


jmarrec · 2026-01-06T10:26:43Z

src/modelica/antlr.zip

@kbenne did you mean to commit this?

jmarrec · 2026-01-06T10:28:07Z

src/modelica/main.cpp

+  auto classDefinitions = moFile.getClassDefinitions();
+
+  for (const auto& classDef : classDefinitions) {
+    std::cout << "Found class name: " << classDef.longClassSpecifier() << std::endl;  // NOLINT


You're already importing fmt, so why not use it instead of iostream?

jmarrec · 2026-01-06T10:29:39Z

src/modelica/ModelicaFileImpl.hpp

+  ModelicaSyntax(std::shared_ptr<ModelicaFileImpl> modelicaFile, T* ctx)  // NOLINT
+    : m_modelicaFile(std::move(modelicaFile)), m_ctx(ctx), m_id(Derived::idFromCTX(ctx)) {
+    if (auto file = m_modelicaFile.lock()) {
+      file->addModelicaSyntax(this);
+    }
+  }


Nitpick but you're using a cpp file to go with it, so why add the implementation inline in the header file?

jmarrec · 2026-01-06T10:29:55Z

src/modelica/ModelicaFileImpl.hpp

+  ModelicaSyntax(const ModelicaSyntax&) = delete;  // NOLINT
+  ModelicaSyntax(ModelicaSyntax&&) = default;      // NOLINT


(curious why NOLINT)

jmarrec · 2026-01-06T10:39:09Z

src/measure/ModelicaParameters.cpp

+  namespace detail {
+    class ModelicaParameter_Impl
+    {
+     public:
+      ModelicaParameter_Impl(std::string_view model, std::string_view key, std::string_view value) : m_model(model), m_key(key), m_value(value) {}
+
+      std::string model() const {
+        return m_model;
+      };
+
+      std::string key() const {
+        return m_key;
+      };
+
+      std::string value() const {
+        return m_value;
+      }
+
+      void setValue(std::string_view value) {
+        m_value = value;
+      }
+
+     private:
+      std::string m_model;
+      std::string m_key;
+      std::string m_value;
+    };


I find that weird to define the Impl fully inside the cpp file because it doesn't follow the usual convention in the repo.

jmarrec · 2026-01-06T10:46:43Z

src/modelica/ModelicaFile.hpp

Generally we have a single cpp file for both Impl and the public forward methods.

jmarrec · 2026-01-06T10:50:11Z

src/modelica/ModelicaFileImpl.cpp

+std::string ModelicaFileImpl::printTree() const {
+  std::stringstream ss;
+
+  const std::function<void(antlr4::tree::ParseTree*, antlr4::Parser*, int)> to_s = [&ss, &to_s](antlr4::tree::ParseTree* tree, antlr4::Parser* parser,


recursive std::function, fancy.

jmarrec · 2026-01-06T10:52:44Z

src/modelica/ModelicaFileImpl.cpp

+//std::vector<ComponentDeclaration> ComponentClause::componentDeclarations() {
+//  std::vector<ComponentDeclaration> result;
+//
+//  const auto component_declarations = antlr4::tree::xpath::XPath::findAll(ctx(), "//Component_declaration", parser());
+//
+//  for (auto* c : component_declarations) {
+//    auto* component_declaration_ctx = dynamic_cast<modelicaParser::Component_declarationContext*>(c);
+//    assert(component_declaration_ctx != nullptr);
+//    result.emplace_back(modelicaFile(), component_declaration_ctx);  // NOLINT
+//  }
+//
+//  return result;
+//}
+
+//std::string ComponentDeclaration::idFromCTX([[maybe_unused]] modelicaParser::Component_declarationContext* ctx) {
+//  // TODO
+//  return "";
+//}
+
+//std::vector<Declaration> ClassDefinition::declarations() {
+//  class_definition::GetDeclarationsVisitor v;
+//  v.visit(m_ctx);
+//  return v.getResult();
+//}
+//
+//void ClassDefinition::addDeclaration(const std::string& declarationText) {
+//  const auto el = m_ctx->class_specifier()->long_class_specifier()->composition()->element_list();
+//
+//  antlr4::ANTLRInputStream input(declarationText + ";");  // Ensure it's a complete statement
+//  modelicaLexer lexer(&input);
+//  antlr4::CommonTokenStream tempTokens(&lexer);
+//  modelicaParser tempParser(&tempTokens);
+//
+//  [[maybe_unused]] const auto* component = tempParser.component_clause();
+//
+//  // Attempt to parse as a component_clause (variable declaration)
+//  modelicaParser::Component_clauseContext* parsedDecl = nullptr;
+//}
+//
+//Declaration::Declaration(modelicaParser::DeclarationContext* ctx) : m_ctx(ctx) {}
+//
+//std::string Declaration::name() const {
+//  return m_ctx->IDENT()->getText();
+//}


what's up with the commented code?

jmarrec · 2026-01-06T10:55:25Z

src/workflow/OSWorkflow.cpp

-  if (workflowJSON.runOptions()->debug() && oriLogLevel > Debug) {
-    targetLogLevel = Debug;
-  }
+  LogLevel targetLogLevel = workflowJSON.runOptions()->debug() ? Debug : oriLogLevel;


Why?

Seems like you're no longer matching the comment above that says

If the user passed something like openstudio --loglevel Trace run --debug -w workflow.osw, we retain the Trace

kbenne added 2 commits December 15, 2025 21:00

Fix for ssl link on Apple

84223c0

kbenne mentioned this pull request Dec 16, 2025

Introduce first-class Modelica workflow and Measure support #5552

Closed

anchapin added the Pull Request - Ready for CI This pull request if finalized and is ready for continuous integration verification prior to merge. label Dec 16, 2025

anchapin merged commit b652977 into develop Dec 16, 2025
3 checks passed

anchapin deleted the issue-5356-modelica-measures-squashed branch December 16, 2025 15:07

jmarrec reviewed Jan 6, 2026

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Issue 5356 modelica measures squashed #5557

Issue 5356 modelica measures squashed #5557

kbenne commented Dec 16, 2025

Uh oh!

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

jmarrec Jan 6, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

		ModelicaSyntax(const ModelicaSyntax&) = delete; // NOLINT
		ModelicaSyntax(ModelicaSyntax&&) = default; // NOLINT

Issue 5356 modelica measures squashed #5557

Issue 5356 modelica measures squashed #5557

Conversation

kbenne commented Dec 16, 2025

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!

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!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants