solverpilot 0.1.0rc2__tar.gz → 0.2__tar.gz

solverpilot-0.2/BACKEND-CONTRACT-V0_1_0.json

@@ -0,0 +1,38 @@
+{
+  "milestone": "RELEASE_0_1",
+  "notes": [
+    "Backend IDs remain stable across the SolverPilot pre-GitHub cleanup.",
+    "Verification-only adapters are never admitted to the default solving registry.",
+    "Availability of optional public backends remains environment-dependent.",
+    "The legacy optimind Python package is not shipped in SolverPilot 0.1."
+  ],
+  "package_version": "0.1",
+  "production_baseline_policy": {
+    "convex_qp": "osqp-native when available; otherwise scipy-slsqp-qp-bridge for non-proof intents",
+    "learned_lp_performance_routing": false,
+    "lp": "scipy-highs-ds",
+    "milp": "scipy-highs-bridge"
+  },
+  "schema": "solverpilot.backend_contract.v1",
+  "stable_backend_ids": {
+    "base_runtime": [
+      "scipy-highs-ds",
+      "scipy-highs-ipm",
+      "scipy-highs-bridge",
+      "scipy-slsqp-qp-bridge"
+    ],
+    "optional_public": [
+      "highspy-native",
+      "osqp-native",
+      "pyscipopt-native",
+      "nlopt-slsqp-native"
+    ],
+    "verification_only": [
+      "casadi-osqp-bridge",
+      "casadi-highs-bridge",
+      "casadi-cbc-bridge",
+      "bundled-osqp-capi",
+      "bundled-highs-capi"
+    ]
+  }
+}

solverpilot-0.2/BACKEND-CONTRACT-V0_2_0.json

@@ -0,0 +1,42 @@
+{
+  "milestone": "RELEASE_0_2",
+  "notes": [
+    "Backend IDs remain stable across the SolverPilot pre-GitHub cleanup.",
+    "Verification-only adapters are never admitted to the default solving registry.",
+    "Availability of optional public backends remains environment-dependent.",
+    "The legacy optimind Python package is not shipped in SolverPilot 0.1.",
+    "Clarabel has a separate conic capability/conformance surface; it is not an LP/MILP registry entry."
+  ],
+  "package_version": "0.2",
+  "production_baseline_policy": {
+    "convex_qp": "osqp-native when available; otherwise scipy-slsqp-qp-bridge for non-proof intents",
+    "learned_lp_performance_routing": false,
+    "lp": "scipy-highs-ds",
+    "milp": "scipy-highs-bridge"
+  },
+  "schema": "solverpilot.backend_contract.v1",
+  "stable_backend_ids": {
+    "base_runtime": [
+      "scipy-highs-ds",
+      "scipy-highs-ipm",
+      "scipy-highs-bridge",
+      "scipy-slsqp-qp-bridge"
+    ],
+    "optional_public": [
+      "highspy-native",
+      "osqp-native",
+      "pyscipopt-native",
+      "nlopt-slsqp-native"
+    ],
+    "verification_only": [
+      "casadi-osqp-bridge",
+      "casadi-highs-bridge",
+      "casadi-cbc-bridge",
+      "bundled-osqp-capi",
+      "bundled-highs-capi"
+    ]
+  },
+  "optional_conic_backends": [
+    "clarabel-native"
+  ]
+}

{solverpilot-0.1.0rc2 → solverpilot-0.2}/CHANGELOG.md

@@ -1,5 +1,28 @@
 # Changelog
 
+## 0.2 — Runtime, modeling and direct conic upgrade
+
+- Serialize process start/close across concurrent batches to prevent the Windows child-handle race found during release qualification.
+
+- Cache backend version metadata and avoid discovery for explicit backend objects.
+- Eliminate duplicate QP convexity work and vectorize LP row conversion.
+- Expose absolute/relative validation tolerances and numerical LP/QP KKT checks.
+- Add supplied Farkas/recession certificate verification and named quality reports.
+- Add owned transactional reoptimization and serialize mutable solver operations.
+- Add bounded process batches with cancellation, deadlines and explicit POSIX memory limits.
+- Add indexed variables, explicit soft penalties, named diagnostics/results and linear lexicographic objectives.
+- Add direct Clarabel with SOC/RSOC/PSD/exponential/power and quadratic objectives.
+- Fix cloning of power/division/smooth expressions; retain historical API call compatibility.
+- Add numerical campaigns, repeated-solve RSS benchmarks and three public examples.
+- Extend exact-artifact qualification with Clarabel on all supported platforms.
+
+## 0.1 — Final release numbering
+
+- Promote the published 0.1.0rc2 code to the short final version 0.1.
+- Align package metadata, API snapshots, documentation, and release checks.
+- No algorithm, solver policy, public API, or performance changes.
+- Preserve historical release candidate evidence; qualify the new artifacts separately.
+
 ## 0.1.0rc2 — Correctness and trust hardening
 
 - canonicalizes every accepted QP Hessian to one symmetric matrix before hashing, validation, or backend translation; pairwise symmetry checks can no longer be relaxed by an unrelated large coefficient;

solverpilot-0.2/KNOWN-LIMITATIONS.md

@@ -0,0 +1,33 @@
+# Known limitations — 0.2
+
+## Numerical evidence
+
+Primal validation is independent of backend termination. Continuous LP/QP duals from SciPy LP, native HiGHS and OSQP can additionally be checked for stationarity, complementarity and numerical gap. These are tolerance-qualified checks, not exact-arithmetic proofs. MILP/global nonlinear optimality is not proved by primal feasibility. Backend-only infeasible/unbounded statuses remain claims unless a certificate is checked.
+
+Per-call ValidationTolerances is supported by execute, solve, solve_production and execute_portfolio. The default feasibility floor is 1e-7 with relative tolerance 1e-9; the relative allowance can be disabled. Native solver tolerances must be configured separately. Nonfinite evidence cannot produce a verified certificate.
+
+## Conic and nonlinear scope
+
+Optional direct Clarabel supports convex quadratic objectives and affine SOC, rotated SOC, PSD, exponential and power cones. Runtime conformance and original-space validation gate its use. CasADi SuperSCS retains its restricted linear-objective SOC/RSOC scope; its generic PSD and quadratic-conic routes remain disabled. Generalized power cones are not implemented. Conic primal feasibility alone does not imply independently verified global optimality.
+
+Smooth NLP remains optional through CasADi/Ipopt and produces local candidates. Generic global NLP and arbitrary nonsmooth atoms are unsupported. MINLP remains the certified convex binary-discrete scope: general integer, nonconvex/global, nonlinear-equality, two-sided nonlinear and indicator+MINLP compositions remain unsupported. Existing proof flags retain their documented solver-certified scope.
+
+## Execution and ownership
+
+Batch accepts canonical LP/QP, uses spawn and requires a script main guard. Timeouts/cancellation terminate workers; unobserved incumbents cannot be recovered afterward. Native solver budgets may return incumbents before the outer deadline. Cleanup may add bounded latency. POSIX memory limits are RLIMIT_AS address-space caps, not RSS; Windows memory caps are rejected. Worker count does not limit internal solver threads.
+
+ReoptimizationSession owns a model clone and serializes update+solve. Session/PersistentSession serialize their public methods, but direct external model/backend mutations require caller coordination. Compiler caches are not shared mutable workspaces for uncoordinated modeling. Reuse is conditional on structure/settings and is not guaranteed faster. Session history can intentionally grow.
+
+CP-SAT remains process-isolated to avoid known HiGHS ABI collisions. Reference CP and exact TSP/VRP are exhaustive and only suitable for small oracle cases.
+
+## Modeling and records
+
+Soft helpers accept affine relations and require explicit penalty use in the objective. Lexicographic solving currently supports canonical LP/MILP with linear objectives; locks include numerical tolerances. Named conflict diagnostics support LP/MILP. Named values reject stale compiled models and retain IDs for duplicate labels.
+
+Learned routing remains disabled; OOD indicators are not calibrated probabilities. History is opt-in plaintext SQLite; user metadata can be sensitive. Extension registry mutation requires caller coordination. Development pickle caches are trusted artifacts, not a safe untrusted interchange format.
+
+## Qualification
+
+Cross-platform support requires exact-artifact qualification on Python 3.12–3.14. Upstream package availability alone is not integration evidence. Synthetic latency/RSS measurements do not establish industrial scalability or absence of long-term leaks. Historical reports retain their original versions and cannot qualify new artifacts.
+
+See [the 0.2 guide](docs/release/SOLVERPILOT-PUBLIC-DOCS-0.2.md) for APIs and examples.

{solverpilot-0.1.0rc2 → solverpilot-0.2}/MANIFEST.in

@@ -28,3 +28,13 @@ prune .github
 prune tools
 global-exclude __pycache__ *.py[cod] *.egg-info *.whl *.tar.gz *.zip
 recursive-include examples *.py *.md
+
+include PUBLIC-API-V0_1_0.json
+include BACKEND-CONTRACT-V0_1_0.json
+include docs/release/SOLVERPILOT-PUBLIC-DOCS-0.1.md
+include docs/release/README-CHECKLIST-0.1.md
+
+include PUBLIC-API-V0_2_0.json
+include BACKEND-CONTRACT-V0_2_0.json
+include docs/release/SOLVERPILOT-PUBLIC-DOCS-0.2.md
+include docs/release/README-CHECKLIST-0.2.md

{solverpilot-0.1.0rc2 → solverpilot-0.2}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: solverpilot
-Version: 0.1.0rc2
+Version: 0.2
 Summary: Trust-aware optimization runtime for mathematical and constraint programming in Python
 Author: Amirhossein Heidari Rashtabad
 Maintainer: Amirhossein Heidari Rashtabad
@@ -62,6 +62,8 @@ Requires-Dist: pyscipopt<7,>=6.2.1; extra == "native"
 Requires-Dist: nlopt<3,>=2.11; extra == "native"
 Provides-Extra: nlopt
 Requires-Dist: nlopt<3,>=2.11; extra == "nlopt"
+Provides-Extra: clarabel
+Requires-Dist: clarabel==0.11.1; extra == "clarabel"
 Provides-Extra: casadi
 Requires-Dist: casadi==3.7.2; extra == "casadi"
 Provides-Extra: conic
@@ -76,7 +78,7 @@ Dynamic: license-file
 
 # SolverPilot
 
-> **Pre-release:** `0.1.0rc2`. This is the second pre-public release candidate; it supersedes `0.1.0rc1` after correctness/trust hardening and packages the hardened Core + Track P P0–P9 platform together with the verified S2–S10 trust, evaluation, extension, TSP/VRP, and history layers. Public PyPI publication remains fail-closed until the exact-artifact external qualification workflow passes.
+> **Current version:** `0.2`. Adds configurable validation, numerical optimality checks, faster preparation, owned reoptimization sessions, bounded batch execution, modeling conveniences and a direct Clarabel conic adapter. Historical `0.1.0rc2` correctness/trust hardening remains in place.
 
 SolverPilot is a **trust-aware optimization runtime and modeling layer for Python**. It combines a small matrix-first solve API for LP/MILP/convex QP with a higher-level semantic modeling system that can compile into conic, smooth nonlinear, certified convex binary MINLP, and constraint-programming execution paths.
 
@@ -113,7 +115,7 @@ The extended APIs are canonically imported from submodules such as `solverpilot.
 
 ### Verified trust, evaluation, extension, and application layers
 
-The `0.1.0rc2` release candidate keeps the S2–S10 additive layers and adds correctness/trust hardening discovered during adversarial review of `0.1.0rc1`. The historical `0.0.40rc2` artifacts remain separate provenance and are not overwritten.
+The `0.2` release keeps the S2–S10 additive layers and adds correctness/trust hardening discovered during adversarial review of `0.1.0rc1`. The historical `0.0.40rc2` artifacts remain separate provenance and are not overwritten.
 
 - `solverpilot.io`: strict local JSON/CSV ingestion, bounded reads, explicit mapping, content hashes, and source provenance;
 - `solverpilot.evaluation`: direct-run oracle construction and decomposed quality/runtime/failure regret;
@@ -171,7 +173,7 @@ Optional integrations are installed separately. OR-Tools 9.15 provides Python 3.
 
 ### From PyPI
 
-After the first public release:
+Install the latest published release from PyPI:
 
 ```bash
 python -m pip install solverpilot
@@ -180,7 +182,7 @@ python -m pip install solverpilot
 ### From the repository before PyPI publication
 
 ```bash
-git clone <REPOSITORY_URL>
+git clone https://github.com/Amirhosseinheidari/SolverPilot.git
 cd solverpilot
 python -m pip install .
 ```
@@ -342,7 +344,7 @@ print(result.algorithm)
 print(result.globally_proven)
 ```
 
-`globally_proven=True` is emitted only when the certified scope, original-space validation, and bound-closure conditions all pass. In `0.1.0rc2` this is a **solver-certified proof under the compiler's convexity assumptions**, not an independently reconstructed end-to-end proof; inspect `result.proof_scope` and `result.independently_verified_global`. General integer, nonconvex, nonlinear-equality, and indicator+MINLP compositions remain fail-closed.
+`globally_proven=True` is emitted only when the certified scope, original-space validation, and bound-closure conditions all pass. In `0.2` this is a **solver-certified proof under the compiler's convexity assumptions**, not an independently reconstructed end-to-end proof; inspect `result.proof_scope` and `result.independently_verified_global`. General integer, nonconvex, nonlinear-equality, and indicator+MINLP compositions remain fail-closed.
 
 See [`examples/11_minlp_optional.py`](examples/11_minlp_optional.py).
 
@@ -464,6 +466,7 @@ A native patch is selected only when the exact granular capability and persisten
 | SCIP | `python -m pip install "solverpilot[scip]"` | PySCIPOpt adapter |
 | NLopt | `python -m pip install "solverpilot[nlopt]"` | NLopt SLSQP integration |
 | CasADi | `python -m pip install "solverpilot[casadi]"` | pinned CasADi 3.7.2 verification bridge |
+| Clarabel | `python -m pip install "solverpilot[clarabel]"` | direct convex conic solver, including PSD, exponential and power cones |
 | Conic | `python -m pip install "solverpilot[conic]"` | CasADi-based conic verification path |
 | NLP | `python -m pip install "solverpilot[nlp]"` | CasADi 3.7.2 + Ipopt path |
 | MINLP | `python -m pip install "solverpilot[minlp]"` | dependencies for certified binary MINLP orchestration |
@@ -552,3 +555,7 @@ Normal CI targets Python 3.12–3.14. The exact-artifact cross-platform release
 ## License
 
 SolverPilot is licensed under the [Apache License 2.0](LICENSE). Apache-2.0 permits commercial and closed-source use subject to its terms. It does not grant trademark rights.
+
+## Version 0.2 guide
+
+See [the 0.2 guide](docs/release/SOLVERPILOT-PUBLIC-DOCS-0.2.md) for new APIs, installation, numerical guarantees, batch limits and examples.