Installation
Novomodelo ships as the novomodelo command-line program, with pre-built binaries for the
platforms listed below, and as the novomodelo-python package for Python. Choose the
method that best fits your environment; to install the Python package, go to
Python Package.
Pre-built Binaries (Recommended)
Seção intitulada “Pre-built Binaries (Recommended)”No Rust toolchain or C compiler required.
Linux and macOS
Seção intitulada “Linux and macOS”curl --proto '=https' --tlsv1.2 -LsSf https://github.com/ons-ccee-epe/novomodelo/releases/latest/download/novomodelo-cli-installer.sh | shThe installer places the novomodelo binary in $CARGO_HOME/bin (typically
~/.cargo/bin). Add that directory to your PATH if it is not already present.
Windows (PowerShell)
Seção intitulada “Windows (PowerShell)”powershell -ExecutionPolicy Bypass -c "irm https://github.com/ons-ccee-epe/novomodelo/releases/latest/download/novomodelo-cli-installer.ps1 | iex"Supported Platforms
Seção intitulada “Supported Platforms”| Platform | Target Triple |
|---|---|
| macOS (Apple Silicon) | aarch64-apple-darwin |
| macOS (Intel) | x86_64-apple-darwin |
| Linux (x86-64) | x86_64-unknown-linux-gnu |
| Linux (ARM64) | aarch64-unknown-linux-gnu |
| Windows (x86-64) | x86_64-pc-windows-msvc |
The Linux archives, which the installer also downloads, are built on Ubuntu
24.04 and need glibc 2.38 or newer; ldd --version prints your glibc version.
They do not run on an older system such as Ubuntu 22.04, Debian 12 or RHEL 9;
there, install from crates.io instead, which builds against
the system’s own glibc.
You can also download individual archives directly from the GitHub Releases page.
Verify the Installation
Seção intitulada “Verify the Installation”novomodelo versionExpected output (exact versions and arch will vary):
novomodelo v0.21.0solver: HiGHS 1.13.1comm: localzstd: enabledarch: x86_64-linuxbuild: release (lto=thin)Python Package
Seção intitulada “Python Package”The novomodelo-python package installs the novomodelo Python module, which runs
studies and loads their results from Python. The wheels bundle the HiGHS solver
and depend on no other Python package.
pip install novomodelo-pythonThe package requires CPython 3.12 or newer; on Python 3.11 or older, pip
reports No matching distribution found for novomodelo-python. Each wheel is
built for the CPython stable ABI (abi3), so one wheel per platform serves every
supported Python version, and the release tests every wheel on its own platform
on Python 3.12, 3.13 and 3.14. PyPI publishes wheels for these platforms:
| Platform | Wheel platform tag |
|---|---|
| Linux (x86-64, glibc) | manylinux_2_34_x86_64 |
| Linux (ARM64, glibc) | manylinux_2_28_aarch64 |
| Linux (x86-64, musl) | musllinux_1_2_x86_64 |
| macOS (Apple Silicon) | macosx_11_0_arm64 |
| macOS (Intel) | macosx_10_12_x86_64 |
| Windows (x86-64) | win_amd64 |
pip selects the wheel that matches your system, so you do not choose a tag
yourself. On Linux with glibc (the C library most distributions use), the x86-64
wheel needs glibc 2.34 or newer (RHEL and Rocky Linux 9 or newer, Ubuntu 22.04
or newer, Debian 12 or newer) and the ARM64 wheel glibc 2.28 or newer;
ldd --version prints your glibc version. Alpine Linux and other musl systems
use the musllinux wheel.
Where no wheel fits, such as on Linux x86-64 with a glibc older than 2.34
(RHEL 8, Ubuntu 20.04), pip builds the package from the source distribution
(sdist) that PyPI publishes beside the wheels. The build needs a Rust toolchain
(1.88 or newer), CMake and a C++ compiler.
Verify the Python Package
Seção intitulada “Verify the Python Package”python -c "import novomodelo; print(novomodelo.__version__); print(novomodelo.version_info())"0.21.0{'version': '0.21.0', 'solver': 'HiGHS 1.13.1', 'comm': 'local', 'zstd': 'enabled', 'arch': 'x86_64-linux', 'build': 'release'}If the first line shows the installed version and the second a dictionary, the
package works. version_info() reports the fields that novomodelo version prints;
comm is always
local, because the package runs a study in a single process. Next, run a study
from Python with the Python Quickstart.
From crates.io
Seção intitulada “From crates.io”cargo install novomodelo-cliRequires Rust 1.88+ and build prerequisites (see Build from Source below).
Installs to $CARGO_HOME/bin.
Build from Source
Seção intitulada “Build from Source”For contributors or unsupported platforms.
Prerequisites
Seção intitulada “Prerequisites”| Dependency | Minimum Version | Notes |
|---|---|---|
| Rust toolchain | 1.88 (stable) | Install via rustup |
| C compiler | any recent GCC or Clang | Required for the HiGHS LP solver |
| CMake | 3.15 | Required for the HiGHS build system |
| Git | any | Required for submodule initialization |
# Clone the repositorygit clone https://github.com/ons-ccee-epe/novomodelo.gitcd novomodelo
# Initialize HiGHS submodule (required for the solver backend)git submodule update --init --recursive
# Build the release binarycargo build --release -p novomodelo-cliThe binary is written to target/release/novomodelo. Optionally install to $CARGO_HOME/bin:
cargo install --path crates/novomodelo-cliVerify:
./target/release/novomodelo versionThe build sets no target features or target CPU, so on x86-64 a binary built from
source, cargo install included, uses the same floating-point kernel as the
pre-built archives of the same version and selects the same cuts. The bundled
HiGHS, CLP and qhull C and C++ sources compile without fused multiply-adds on
every target, so a C compiler whose default target has FMA changes neither the
LP solutions nor the fitted production planes. Setting -C target-feature or
-C target-cpu through RUSTFLAGS can change the kernel, and the run can then
select different, equally valid cuts.
Choosing the LP Backend
Seção intitulada “Choosing the LP Backend”Novomodelo supports two LP solver backends, selected at build time via Cargo features. Exactly one backend is compiled into any given binary.
| Backend | Feature flag | License | Notes |
|---|---|---|---|
| HiGHS | highs | MIT | Default. No extra steps required. |
| CLP | clp | EPL-2.0 | COIN-OR. Opt-in; requires the CLP/CoinUtils submodules. |
Default build (HiGHS)
Seção intitulada “Default build (HiGHS)”cargo build --release -p novomodelo-cliNo flags are needed. HiGHS is the default backend and the one shipped in pre-built binaries.
CLP build
Seção intitulada “CLP build”# Initialize the CLP and CoinUtils submodules firstgit submodule update --init --recursive
# Build with CLP, disabling the HiGHS defaultcargo build --release -p novomodelo-cli --no-default-features --features clpOn a CLP build, every training.solver/simulation.solver override field is
rejected at study setup — see
solver profile validation
for the full rejection rule and error messages.
Mutual exclusivity
Seção intitulada “Mutual exclusivity”The highs and clp features are mutually exclusive — exactly one LP backend
is compiled into a binary, and enabling both at once is a compile error. Because
highs is the default feature, selecting CLP requires --no-default-features
to suppress the default before --features clp is applied; a plain
--features clp leaves the highs default on and fails the build. Enabling
neither backend is also a compile error, so a backend is always chosen
explicitly. The default build (no extra flags) uses HiGHS.
Identifying the active backend
Seção intitulada “Identifying the active backend”The novomodelo version banner shows which backend is compiled in:
novomodelo v0.21.0solver: CLP 1.17.11comm: local...The solver and solver_version fields in each run’s output metadata record
the active backend identifier ("highs" or "clp") and its library version
string. These fields are written by both the CLI and the Python bindings.
Results are reproducible per binary: a build with the other backend is a different binary, so its results are not promised equal. See Determinism & Provenance.
Next Steps
Seção intitulada “Next Steps”- Quickstart — run a complete study end to end using the built-in
1dtoytemplate - Python Quickstart — run a study and read its results from Python
- Running Studies — validate, run, and inspect results for any case directory
- CLI Reference — complete flag and subcommand reference