Remove conda from CI, use uv + system IPOPT

Replace conda-based test_nlp_solvers workflow with uv, installing IPOPT
via system packages (apt on Ubuntu, brew on macOS) instead of conda-forge.
Uncomment IPOPT optional dependency in pyproject.toml so uv sync --extra
IPOPT works. Update installation docs in CLAUDE.md and README.md.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: Transurgeon

.github/workflows/test_nlp_solvers.yml

@@ -10,37 +10,30 @@ on:
 jobs:
   run_nlp_tests:
     runs-on: ${{ matrix.os }}
-    defaults:
-      run:
-        shell: bash -l {0}
     strategy:
       fail-fast: false
       matrix:
         os: [ubuntu-latest, macos-latest]
     steps:
-      - uses: actions/setup-python@v6
-        with:
-          python-version: "3.12"
       - uses: actions/checkout@v5
         with:
           submodules: recursive
-      - name: Set up Conda
-        uses: conda-incubator/setup-miniconda@v3
+      - uses: astral-sh/setup-uv@v7
         with:
-          activate-environment: nlp-test
-          python-version: 3.12
-          channels: conda-forge
-          channel-priority: strict
-          auto-update-conda: true
-          miniforge-version: latest
-          use-mamba: true
-      - name: Install cyipopt via conda
-        run: |
-          conda install -y -c conda-forge cyipopt
-      - name: Install test dependencies
-        run: |
-          pip install -e .
-          pip install pytest hypothesis
+          python-version: "3.12"
+          enable-cache: true
+      - name: Install IPOPT (Ubuntu)
+        if: runner.os == 'Linux'
+        run: sudo apt-get update && sudo apt-get install -y coinor-libipopt-dev
+      - name: Install IPOPT (macOS)
+        if: runner.os == 'macOS'
+        run: brew install ipopt
+      - name: Install dependencies
+        run: uv sync --extra IPOPT --dev
+      - name: Print installed solvers
+        run: uv run python -c "import cvxpy; print(cvxpy.installed_solvers())"
       - name: Run nlp tests
-        run: |
-          pytest -s -v cvxpy/tests/nlp_tests/.
+        run: uv run pytest -s -v cvxpy/tests/nlp_tests/
+
+    env:
+      UV_SYSTEM_PYTHON: 1

CLAUDE.md

@@ -11,11 +11,16 @@ For theoretical foundation, see: [Disciplined Nonlinear Programming](https://web
 ## Build and Development Commands
 
 ```bash
-# Install IPOPT solver (required for NLP - use conda, NOT pip)
-conda install -c conda-forge cyipopt
+# Install IPOPT solver (required for NLP)
+# Ubuntu/Debian:
+sudo apt-get install coinor-libipopt-dev
+# macOS:
+brew install ipopt
+# Then: uv sync --extra IPOPT  (or: pip install cyipopt)
 
-# Install from source (development mode)
-pip install -e .
+# Install from source (development mode) — uv or pip
+uv sync --all-extras --dev   # preferred (uv.lock is committed)
+pip install -e .              # alternative
 
 # Install pre-commit hooks
 pip install pre-commit && pre-commit install
@@ -65,7 +70,7 @@ prob.solve(nlp=True, solver=cp.IPOPT, best_of=5)
 
 | Solver | License | Installation |
 |--------|---------|--------------|
-| [IPOPT](https://github.com/coin-or/Ipopt) | EPL-2.0 | `conda install -c conda-forge cyipopt` (conda only — pip install is unreliable) |
+| [IPOPT](https://github.com/coin-or/Ipopt) | EPL-2.0 | Install system IPOPT (`apt install coinor-libipopt-dev` / `brew install ipopt`), then `pip install cyipopt` |
 | [Knitro](https://www.artelys.com/solvers/knitro/) | Commercial | `pip install knitro` (requires license) |
 | [COPT](https://www.copt.de/) | Commercial | Requires license |
 | [Uno](https://github.com/cuter-testing/uno) | Open source | See Uno documentation |
@@ -246,6 +251,15 @@ A common pattern is to solve with both a DCP solver (CLARABEL) and an NLP solver
 
 Uses a custom build backend (`setup/build_meta.py`) that re-exports `setuptools.build_meta`. The `setup/` directory handles C extensions (cvxcore, sparsecholesky) and version management. Solver registration is in `cvxpy/reductions/solvers/defines.py` (`SOLVER_MAP_CONIC`, `SOLVER_MAP_QP`, `SOLVER_MAP_NLP`).
 
+## CI Checks
+
+PRs must pass these GitHub Actions workflows:
+- **pre-commit** — ruff linting, JSON schema validation (workflows/dependabot), actionlint, pyproject validation
+- **build** — full test suite on Ubuntu/macOS/Windows × Python 3.11–3.14, plus wheel builds
+- **test_nlp_solvers** — NLP solver tests (IPOPT via system package + uv, Knitro if available)
+- **test_optional_solvers** — optional solver integration tests (uses `uv sync --all-extras`)
+- **test_backends** — tests SCIPY and COO canonicalization backends
+
 ## Benchmarks
 
 Benchmarks use [Airspeed Velocity](https://asv.readthedocs.io/) and live in the `benchmarks/` directory. To run locally:

README.md

@@ -7,12 +7,19 @@ with some rules governing how the nonsmooth functions can be used. For details,
 ## Installation
 The installation consists of two steps.
 
-#### Step 1: Install IPOPT via Conda
-DNLP requires an NLP solver. The recommended solver is [Ipopt](https://coin-or.github.io/Ipopt/), which can be installed together with its Python interface [cyipopt](https://github.com/mechmotum/cyipopt):
+#### Step 1: Install IPOPT
+DNLP requires an NLP solver. The recommended solver is [Ipopt](https://coin-or.github.io/Ipopt/). First install the IPOPT system library, then install the Python interface [cyipopt](https://github.com/mechmotum/cyipopt):
 ```bash
-conda install -c conda-forge cyipopt
+# Ubuntu/Debian
+sudo apt-get install coinor-libipopt-dev
+
+# macOS
+brew install ipopt
+```
+Then install the Python interface:
+```bash
+pip install cyipopt
 ```
-Installing cyipopt via pip may lead to issues, so we strongly recommend using the Conda installation above, even if the rest of your environment uses pip.
 
 #### Step 2: Install DNLP
 DNLP is installed by cloning this repository and installing it locally:
@@ -56,7 +63,7 @@ print("Maximum eigenvalue:      " , np.max(eigenvalues))
 ## Supported Solvers
 | Solver | License | Installation |
 |--------|---------|--------------|
-| [IPOPT](https://github.com/coin-or/Ipopt) | EPL-2.0 | `conda install -c conda-forge cyipopt` |
+| [IPOPT](https://github.com/coin-or/Ipopt) | EPL-2.0 | Install system IPOPT (see above), then `pip install cyipopt` |
 | [Knitro](https://www.artelys.com/solvers/knitro/) | Commercial | `pip install knitro` (requires license) |
 
 ---

pyproject.toml

@@ -94,7 +94,7 @@ SCS = []
 XPRESS = ["xpress>=9.5"]
 DAQP = ["daqp"]
 KNITRO = ["knitro"]
-# IPOPT = ["cyipopt"]
+IPOPT = ["cyipopt"]
 testing = ["pytest", "hypothesis"]
 doc = ["sphinx",
        "sphinxcontrib.jquery",