Annotate Modelica Plots

Annotate Modelica Plots

This skill writes the spec-standard result-plot annotation into a Modelica .mo file so a model carries its own simulation plots:

annotation(Documentation(figures = {
  Figure(identifier = "response", title = "Mass response", preferred = true,
    plots = {
      Plot(title = "Position and velocity",
        curves = {Curve(x = time, y = s, legend = "position"),
                  Curve(x = time, y = v, legend = "velocity")},
        x = Axis(label = "time"), y = Axis(label = "states", unit = ""))})}));

It is a self-contained source transform: Claude assembles a figures spec (JSON) — from the user's request, or from variables discovered by simulating — and the Python engine renders it to valid Modelica and splices it into the class annotation. The engine is pure Python (standard library only); WSMKernelX is used only afterwards as a validation gate.

Wolfram System Modeler renders this standard annotation natively from the stored figure, picking up the Plot.title, Axis.label, and per-curve legend. No vendor-specific annotation is needed.

It is idempotent: a class that already has figures = is skipped unless you pass --force (which strips the old figures assignment and regenerates). By default it prints a dry-run diff; nothing is written until --write.

> --force is destructive — ask the user first. It deletes the class's existing figures > block before regenerating, so any hand-written or previously curated plot definitions there are > lost (a user may have tuned those by hand). Don't pass --force on a class that already carries > figures without first confirming with the user that discarding them is intended. Review the > dry-run diff (no --write) to see exactly what would be removed.

The figures spec (JSON)

Shaped after the spec records (MLS annotations.tex §figure-plot-properties). Only fields you set are emitted:

{
  "figures": [
    {
      "title": "Mass response", "identifier": "response", "preferred": true,
      "caption": "Damped oscillation of %[position](variable:s).",
      "plots": [
        {
          "title": "Position and velocity",
          "x": {"label": "time"},
          "y": {"label": "states", "unit": "", "scale": "Linear"},
          "curves": [
            {"x": "time", "y": "s", "legend": "position"},
            {"x": "time", "y": "v", "legend": "velocity"}
          ]
        }
      ]
    }
  ]
}

Rules the engine enforces (from the spec):

• Curve.x/Curve.y must be result-references: a scalar variable (s, mass.v), time, or

der(v[, n]) — not arbitrary expressions. A bad reference is a hard error. x defaults to "time"; set it to a variable for an X vs. Y plot (e.g. {"x": "s", "y": "v"} for a phase portrait).

• Figure.identifier must be unique in the spec; Plot.identifier unique within its figure.
• Axis scale is "Linear" (default), "Log", or {"Log": 10} to set the base.
• %{var} variable replacements and %[text](variable:ref) caption links are passed through; the

engine does not invent them.

To target a specific class in a multi-model file, either pass --class , or use a mapping spec {"Oscillator": {"figures": [...]}, "Plain": {"figures": [...]}}.

Running the annotator

The annotator lives in the PlotAnnotate/ package next to this file. Run it as a module from this skill's directory so the package is importable:

cd ""
python3 -m PlotAnnotate.main --file "" --analyze

Use python instead of python3 on Windows if that's what's on PATH. Only --file (and --spec, --vars-file) must be absolute or correctly-relative paths.

Cross-platform launcher (for simulate / validate steps)

> The simulate/validate steps below call the shared scripts/wsm_run.py and > scripts/plot_mat.py. Read [the shared-conventions appendix at the end of this file](#appendix-shared-conventions-for-the-modelica-skills) > (launcher resolution, shell rules, temp dir, JSON-array output, MSL 4.x notes) > before those steps. In a normal install the launcher is ../scripts/wsm_run.py > (` is the shared scripts/` folder); if that path doesn't exist, > see [Appendix → Locating the launcher](#locating-the-launcher).

Workflow

1. Identify the model file and target class

The user gives a .mo path. Pick a concrete instantiable model for any kernel step (Package.Model, not the package) — see [Appendix → Picking the model name](#picking-the-model-name).

2. Analyze — see classes and existing figure status

python3 -m PlotAnnotate.main --file "" --analyze

Lists every class, flagging has Documentation / HAS figures. Present this and confirm scope.

3. Choose what to plot

Pick curves that reveal emergent / non-obvious behavior — a dynamic response, a comparison, a phase shift, a limit case. Skip tautological curves that just re-trace a known input signal (if you'd see the same shape by plotting the input alone, it teaches nothing).

Curves default to x = time, but Curve.x can be any result variable, giving an X vs. Y plot. Reach for one when the relationship between two variables is the story: a phase portrait of an oscillator (limit cycles show as closed loops), a hysteresis loop, a transfer or force- displacement characteristic. Put an X vs. Y curve in its own Plot (not mixed with time-based curves — they'd share the x axis) and label both axes.

Mode A — user-specified. The user already knows the curves. Build the spec JSON directly from their request and skip to step 4.

Mode B — simulate-and-suggest. When the user doesn't name variables, discover them:

# simulate to produce a .mat (and, beside it, a .sim init file with the same stem)
python3 "/wsm_run.py" --mode simulate --model "" --name "" --timeout 180
# enumerate result variables (write working files next to the model, not into the skill dir)
python3 "/plot_mat.py" "/_wsm_simulate_temp/_res.mat" --list > "/vars.txt"
# get a starting spec (filters time/der/aux; one plot, x=time). Pass the .sim so protected
# variables are dropped — they are stored in the .mat but render blank in a stored figure.
python3 -m PlotAnnotate.main --file "" --vars-file "/vars.txt" \
  --sim-file "/_wsm_simulate_temp/_res.sim" --suggest --max-curves 8 > "/figs.json"

The .sim file sits next to the .mat with the same stem (swap the .mat extension for .sim). Passing --sim-file is what stops protected variables from being suggested — without it, a protected variable in the result is offered as a curve and then renders blank in System Modeler.

The suggestion is a starting point — refine figs.json: drop parameters/noise, split into multiple plots by physical quantity, set titles/legends/axis units. Passing --vars-file to the --annotate step also warns when a curve references a variable absent from the result; passing --sim-file there additionally warns when a curve references a protected variable.

3b. Title, caption, and reference alignment

Every figure should be self-identifying and self-explaining:

• Title — name the quantity shown, not "Plot 1" (e.g. "Refrigerant pressure"). The title is

how the user finds the plot in the model's figure list, so make it scannable.

• Caption — set a one-line caption on every Figure saying what it shows and what to look

for (e.g. "Pressure rises with the accumulation after the step"). It renders beneath the plot in System Modeler. Use %[label](variable:ref) to link a variable inside the caption.

• Default — mark exactly one figure preferred = true so it opens by default.
• Reference alignment — if the model implements a published reference (a paper, app note,

textbook) and that reference numbers its figures, name the model's figures to match it ("Fig. 6 - Stored refrigerant mass") and state the correspondence in the caption ("compare Fig. 6 of "). This makes a replication auditable at a glance — anyone can line the stored plot up against the source. When no reference is given, use your own clear, ordered scheme (e.g. "Evaporator 1 - Outlet vapour quality").

4. Preview — dry-run diff

python3 -m PlotAnnotate.main --file "" --class "" --spec "/figs.json" --annotate

Shows the unified diff without writing. --vars-file "/vars.txt" adds membership warnings, and --sim-file "/_wsm_simulate_temp/_res.sim" warns on any curve that references a protected variable. --force regenerates over an existing figures block.

5. Apply — write in place

python3 -m PlotAnnotate.main --file "" --class "" --spec "/figs.json" --annotate --write

Re-running without --force is a no-op for already-annotated classes (idempotent).

6. Validate — confirm the model still flattens (validate-modelica gate)

Annotations must not change the flatten result.

python3 "/wsm_run.py" --mode validate --model "" --name "" --timeout 90

Parse _wsm_validate_temp/validate.out.json (a JSON array — take [0]; field reference: [Appendix → Reading the JSON output](#reading-the-json-output)) and check status.flatten == "Pass". If it passed before annotating but fails after, that's a bug in the annotation, not the user's model — report it. Then remove the temp dir.

Notes and edge cases

• Pick an instantiable model for the kernel steps, not the package — use a nested model's full

dotted name. See [Appendix → Picking the model name](#picking-the-model-name).

• figures is inherited. A class's figures = its own plus those from base classes. Keep

Figure.identifiers unique across that whole collection (the engine checks within one spec; it can't see base-class figures, so choose distinct identifiers when extending).

• Result-references only. To plot a derived quantity, add a variable for it in the model and

reference that — Curve.y cannot be an expression like s + v.

• No protected variables in curves. System Modeler stores protected variables in the result

.mat (so plot_mat.py --list shows them), but a stored figure resolves curves against the public result tree only — a curve on a protected variable renders blank, and the model still flattens, so the validate gate won't catch it. The .sim file beside the .mat marks them (protected="true"); pass it as --sim-file to --suggest/--annotate so they are dropped and flagged. If you must show a protected quantity, expose it through a public variable in the model.

• X vs. Y (parametric) curves are supported. Curve.x defaults to time, but any

result-reference works (Curve(x = x, y = y)), and System Modeler renders the variable-vs-variable curve natively. Same rules as y: a result-reference (no expressions), and a protected variable renders blank on either axis. Don't mix time-based and X-vs-Y curves in one Plot — all curves share the plot's x axis (the engine warns on this); use separate plots. Label both axes, since the default "time" x-label no longer applies.

• Units. A non-empty Axis.unit must be compatible with the plotted variable's unit, or the

model won't flatten; leave unit = "" to let the tool choose per-curve units.

• Idempotency / regen. Default runs never duplicate a figures block; use --force to

regenerate after editing the spec — but it discards the existing figures (hand-written included), so confirm with the user first (see the caution near the top).

• MSL dialect (this toolchain ships MSL 4.x). Annotations don't affect flattening, but if the

model uses 3.2 names it fails the gate for unrelated reasons — see [Appendix → MSL 4.x dialect](#msl-4x-dialect).

---

Appendix: shared conventions for the Modelica skills

> Shared by every Modelica skill that drives WSMKernelX through the > bundled launcher; inlined here at release time. For the CLI/option > reference, environment variables (WSM_HOME, WSM_VSDEVCMD), > install discovery, and the analysis scripts, see > [../scripts/README.md](../scripts/README.md).

Locating the launcher

` (used throughout the skills) is the shared scripts/` folder. Some installs symlink the skill directories without it, so resolve it in this order and use the first that exists:

1. $WSM_SKILLS_SCRIPTS (bash) or $env:WSM_SKILLS_SCRIPTS (PowerShell), if set.
2. ../scripts relative to the skill directory — in a normal install

../scripts/wsm_run.py already exists, so use that path directly; do not run a shell probe to "resolve" it.

1. The repo checkout you installed from, e.g. .../agentskills/scripts.
2. Last resort, search the home directory:

• PowerShell: Get-ChildItem $HOME -Recurse -Filter wsm_run.py -ErrorAction SilentlyContinue | Select-Object -First 1
• bash/zsh: find ~ -name wsm_run.py -path '*scripts*' 2>/dev/null | head -1

If only #4 finds it, the install is missing the scripts/ link — tell the user to run install.sh (or install.ps1) from the repo, which links scripts/ too.

Shell and Python

On Windows, use PowerShell. The Git-Bash/cygwin layer may be broken (even ls/find can be absent, giving a misleading "exit 127 / command not found"). Run wsm_run.py with python (not python3); those calls are single-line and shell-agnostic. For cleanup use Remove-Item -Recurse -Force, not rm -rf. On macOS/Linux any POSIX shell is fine and python3 is the usual name.

Let the launcher own .mos/.bat and paths

Do not hand-write .mos scripts, .bat files, or hardcode install/compiler paths. The bundled scripts/wsm_run.py handles every OS difference — it finds the System Modeler install and kernel binary (macOS / Windows / Linux), finds and loads the right MSL files, generates the .mos, and runs the kernel with a working compiler environment per platform (system clang/gcc on macOS/Linux; the Visual Studio dev environment via VsDevCmd.bat on Windows). See [../scripts/README.md](../scripts/README.md) for WSM_HOME, the Windows compiler prerequisites, and the full option table.

When the install or compiler isn't found

The launcher searches each OS's standard install locations. If it prints ERROR: Could not locate a Wolfram System Modeler installation, the install is in a non-standard place — ask the user for it and re-run with --wsm-home "" (or have them set WSM_HOME).

Building and simulating also need a C++ toolchain:

• Windows: Visual Studio Build Tools. The launcher locates VsDevCmd.bat

itself; if it reports the compiler environment is missing, pass --vsdevcmd "" (or set WSM_VSDEVCMD) and make sure Build Tools are installed.

• macOS: the Xcode command-line tools (xcode-select --install).
• Linux: gcc/g++.

Run python3 "/wsm_run.py" --mode info to see what the launcher discovered.

Temporary directories

The launcher works in a _wsm__temp/ directory next to the .mo file (_wsm_validate_temp/, _wsm_simulate_temp/, _wsm_diagnose_temp/) and leaves its outputs there. Tell the user, e.g.: "Working in temporary directory _wsm__temp/. This will be deleted afterwards." Pass --tempdir to reuse one directory across models in a session.

Clean up by removing the whole directory — use the user's shell:

rm -rf "/_wsm__temp"          # macOS / Linux
# PowerShell: Remove-Item -Recurse -Force "\_wsm__temp"

Picking the model name

• The user may provide a path to a .mo file, or you may already be working with

one in context.

• Extract the model name: the identifier after model on the first non-comment

line, e.g. model FooBar → FooBar. The filename does not always match the model name — parse the actual model/package declaration.

• For packages or nested models, use the top-level model name.
• Pick an instantiable model, not a package, for any kernel call. A package

cannot be validated or simulated ("Invalid insta

…

Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: WolframResearch
• Source: WolframResearch/system-modeler-ai-toolkit
• License: MIT
• Homepage: https://www.wolfram.com/system-modeler/

Install and usage instructions live in the source repository linked above.