PyPI - dfbench - Versions diffs - 0.2.0__tar.gz → 0.2.2__tar.gz - Mend

dfbench 0.2.0tar.gz → 0.2.2tar.gz

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.

Files changed (184) hide show

{dfbench-0.2.0 → dfbench-0.2.2}/.gitignore RENAMED Viewed

@@ -24,6 +24,11 @@ results/
 notebooks/
 .DS_Store
+runs/
+competition/
+thesis.typ
+training/
 # IDE related
 .vscode/
 .idea/

{dfbench-0.2.0 → dfbench-0.2.2}/PKG-INFO RENAMED Viewed

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: dfbench
-Version: 0.2.0
+Version: 0.2.2
 Summary: Differometor Benchmark - Optimization algorithms and problem definitions
 Project-URL: Homepage, https://github.com/artificial-scientist-lab/Differometor-Benchmark
 Project-URL: Repository, https://github.com/artificial-scientist-lab/Differometor-Benchmark
@@ -34,6 +34,7 @@ Requires-Dist: cmaes>=0.10.0; extra == 'all'
 Requires-Dist: evosax>=0.1.6; extra == 'all'
 Requires-Dist: evox>=1.3.0; extra == 'all'
 Requires-Dist: gpytorch>=1.14; extra == 'all'
+Requires-Dist: hebo>=0.3.2; extra == 'all'
 Requires-Dist: nevergrad>=1.0.12; extra == 'all'
 Requires-Dist: omads>=2408.0; extra == 'all'
 Requires-Dist: optax>=0.2.6; extra == 'all'
@@ -54,6 +55,7 @@ Provides-Extra: bo
 Requires-Dist: ax-platform>=1.2.4; extra == 'bo'
 Requires-Dist: botorch>=0.16.1; extra == 'bo'
 Requires-Dist: gpytorch>=1.14; extra == 'bo'
+Requires-Dist: hebo>=0.3.2; extra == 'bo'
 Requires-Dist: torch>=2.9.1; extra == 'bo'
 Provides-Extra: cuda12
 Requires-Dist: jax[cuda12]==0.9.0.1; extra == 'cuda12'
@@ -88,11 +90,20 @@ Provides-Extra: torch
 Requires-Dist: torch>=2.9.1; extra == 'torch'
 Description-Content-Type: text/markdown
+> [!IMPORTANT]
+> **BETA RELEASE:** This repository is currently in beta. The official competition will start on **July ~7**.
 # Differometor Benchmark
+[![image](https://img.shields.io/pypi/v/dfbench.svg)](https://pypi.python.org/pypi/dfbench)
+[![image](https://img.shields.io/pypi/l/dfbench)](https://pypi.python.org/pypi/dfbench)
+[![image](https://img.shields.io/pypi/pyversions/dfbench.svg)](https://pypi.python.org/pypi/dfbench)
+[![Ruff](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json)](https://github.com/astral-sh/ruff)
+[![Tests](https://img.shields.io/github/actions/workflow/status/artificial-scientist-lab/Differometor-Benchmark/tests.yaml?branch=main&label=tests&logo=github)](https://github.com/artificial-scientist-lab/Differometor-Benchmark/actions/workflows/tests.yaml)
 A benchmarking framework for optimization algorithms on gravitational-wave detector design problems, built on top of the [Differometor](https://github.com/artificial-scientist-lab/Differometor) simulator.
-> **For detailed documentation, see the [Wiki](docs/Home.md).**
+> **For detailed documentation, see the [Wiki](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki).**
 ## TL;DR (I want to try my own algorithm)
@@ -132,7 +143,7 @@ obj.plot_loss()
 obj.save_run_to_file("my_run.npz")
 ```
-**A loss below 0 means your solution beats the real Voyager detector's sensitivity.** (On `VoyagerProblem` without physical constraints — you might be burning mirrors.)
+**A loss below 0 means your solution beats the real Voyager detector's sensitivity.** (On `VoyagerProblem` without physical constraints, you might be burning mirrors.)
 ### Evaluation Methods
@@ -154,8 +165,8 @@ The problems are JAX-based and differentiable up to second order. Use whichever
 ```python
 from dfbench import t2j, j2t
-params_jax = t2j(params_torch)       # Torch → JAX
-losses_torch = j2t(obj.vmap_value(params_jax))  # JAX → Torch
+params_jax = t2j(params_torch)       # Torch -> JAX
+losses_torch = j2t(obj.vmap_value(params_jax))  # JAX -> Torch
 ```
 This adds negligible overhead compared to the objective function itself.
@@ -171,11 +182,11 @@ This adds negligible overhead compared to the objective function itself.
 Both constrained problems accept a `power_penalty_fn(value, threshold)` callable to control how power-constraint violations are penalized.  Built-in presets: `squashed_relu_penalty` (default), `relu_penalty`, `zero_penalty`. Feel free to try own ones. The penalty function can also be swapped after construction via `obj.set_penalty_fn(fn)` (before `obj.start_logging()`), which re-traces the objective so the change takes effect. The unconstrained Voyager problems reject this call because they have no power-constraint path.
-Constrained problems also expose aux diagnostics. `obj.value_aux(params)` and `obj.vmap_value_aux(batch)` return `(loss, aux)` where `aux` carries the loss decomposition, a physical `is_feasible` flag, per-constraint violations, and the raw per-group power arrays. Enable `save` tokens like `is_feasible`, `power_values`, or the `aux` alias and the standard `obj.value` / `obj.value_and_grad` loop records the enabled aux fields in the same forward pass, with no code change. See [Problems](docs/Problems.md) for the aux schema and [Objective API Reference](docs/Objective-API-Reference.md) for the full method set.
+Constrained problems also expose aux diagnostics. `obj.value_aux(params)` and `obj.vmap_value_aux(batch)` return `(loss, aux)` where `aux` carries the loss decomposition, a physical `is_feasible` flag, per-constraint violations, and the raw per-group power arrays. Enable `save` tokens like `is_feasible`, `power_values`, or the `aux` alias and the standard `obj.value` / `obj.value_and_grad` loop records the enabled aux fields in the same forward pass, with no code change. See [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) for the aux schema and [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) for the full method set.
 All problems also support `bounds_overrides` (e.g. `{"tuning": (0, 45)}`) to narrow default property bounds and `signal_floor` to floor detector signal magnitudes before sensitivity normalization. `signal_floor` defaults to `1e-20`. Use `problem.print_bounds()` to inspect effective bounds.
-See [Problems](docs/Problems.md) for details on loss computation, parameter meanings, and constraints.
+See [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) for details on loss computation, parameter meanings, and constraints.
 ---
@@ -212,7 +223,7 @@ pip install -e .                         # CPU-only editable install
 pip install -e ".[all,cuda13,analysis]"  # All backends, CUDA 13, and analysis extras
 ```
-See [Installation](docs/Installation.md) for GPU setup details and HPC notes.
+See [Installation](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Installation) for GPU setup details and HPC notes.
 ---
@@ -223,8 +234,8 @@ OptimizationAlgorithm.optimize()
          │
          ▼
    ┌───────────┐      records losses, params, grads, timestamps
-   │ Objective │ ──►  enforces time / eval budgets
-   └─────┬─────┘      bounded ↔ unbounded sigmoid transform
+   │ Objective │ ──>  enforces time / eval budgets
+   └─────┬─────┘      bounded <-> unbounded sigmoid transform
          │
          ▼
   ContinuousProblem        (VoyagerProblem, VoyagerTuningProblem, ConstrainedVoyagerProblem, UIFOProblem)
@@ -235,7 +246,7 @@ OptimizationAlgorithm.optimize()
 **Design Idea:** Algorithms never create their own `Objective`, they receive a pre-configured one. This lets the benchmark harness or user script control budgets, seeds, and history settings uniformly. The algorithm only has to implement its optimization logic.
-See [Architecture Overview](docs/Architecture-Overview.md) for full design details.
+See [Architecture Overview](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Architecture-Overview) for full design details.
 ---
@@ -247,15 +258,15 @@ src/dfbench/
 │   ├── problem.py        # ContinuousProblem ABC + ProblemSpec
 │   ├── algorithm.py       # OptimizationAlgorithm ABC + AlgorithmType enum
 │   ├── objective.py       # Objective wrapper (central piece)
-│   └── utils.py           # torch↔jax conversion, inverse sigmoid
+│   └── utils.py           # torch<->jax conversion, inverse sigmoid
 ├── algorithms/
 │   ├── derivative_free/   # OMADS, PDFO/Py-BOBYQA, NelderMead, Powell
-│   ├── global_search/     # SciPy BasinHopping, DualAnnealing
-│   ├── evolutionary/      # RandomSearch, EvoxPSO, EvoxES, Nevergrad, CMA family
+│   ├── global_search/     # RandomSearch, SciPy BasinHopping, DualAnnealing
+│   ├── evolutionary/      # EvoxPSO, EvoxES, Nevergrad, CMA family
 │   ├── gradient_based/
 │   │   ├── optax/         # 34 Optax-based optimizers (OptaxAdam, OptaxLAMB, ...)
-│   │   ├── scipy/         # 13 SciPy-based optimizers (BFGS, TNC, SLSQP, …)
-│   │   ├── custom_jax.py  # Native-JAX custom/hybrid batch (SGLD, ASAM, GD→L-BFGS, …)
+│   │   ├── scipy/         # 13 SciPy-based optimizers (BFGS, TNC, SLSQP, ...)
+│   │   ├── custom_jax.py  # Native-JAX custom/hybrid batch (SGLD, ASAM, GD->L-BFGS, ...)
 │   │   └── *.py           # Custom-loop algorithms (AdamGD, LBFGSGD, SAGD, NAAdamGD, OptaxLBFGS)
 │   ├── surrogate_based/
 │   │   ├── botorch/       # BotorchBO, BotorchTuRBO, BotorchqNEI, BotorchqKG,
@@ -337,7 +348,7 @@ benchmark.print_summary(results)
 - `save_csv`: Writes a CSV with all metrics computed at evenly-spaced time points.
 - `save_run_data`: Persists raw loss/params/time histories to NPZ files for later re-evaluation.
-See [Benchmarking](docs/Benchmarking.md) for full configuration options and [Metrics Reference](docs/Metrics-Reference.md) for what gets computed.
+See [Benchmarking](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Benchmarking) for full configuration options and [Metrics Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Metrics-Reference) for what gets computed.
 ### Native-JAX Custom/Hybrid Batch
@@ -360,13 +371,13 @@ see `scripts/voyager_native_jax_custom_batch.py`.
 The interface is designed to make this as simple as possible. You write the optimization logic; `Objective` handles everything else (timing, logging, budget enforcement, file I/O).
-> **Full step-by-step tutorial:** [Implementing a New Algorithm](docs/Implementing-a-New-Algorithm.md)
+> **Full step-by-step tutorial:** [Implementing a New Algorithm](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Implementing-a-New-Algorithm)
 ### The Contract
 1. Subclass `OptimizationAlgorithm`
 2. Declare `algorithm_str` and `algorithm_type`
-3. Implement `optimize(objective, ...) → None`
+3. Implement `optimize(objective, ...) -> None`
 4. Use `Objective` for all function evaluations
 5. The `Objective` is mutated in place, thereby no return is needed
@@ -389,7 +400,7 @@ class MyAlgorithm(OptimizationAlgorithm):
     """My optimization algorithm."""
     algorithm_str = "my_algorithm"
-    algorithm_type = AlgorithmType.EVOLUTIONARY  # match one of the algorithms/ subfolders
+    algorithm_type = AlgorithmType.GLOBAL_SEARCH  # match one of the algorithms/ subfolders
     def __init__(self, batch_size: int = 50) -> None:
         """Algorithm-level meta-parameters that don't change between runs."""
@@ -495,7 +506,7 @@ obj.loss_history            # full loss history
 obj.time_steps              # elapsed time at each evaluation
 ```
-See [Objective API Reference](docs/Objective-API-Reference.md) for the complete interface.
+See [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) for the complete interface.
 ---
@@ -510,7 +521,6 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `BFGS`, `LBFGSB`, `NonlinearCG`, `NewtonCG` | Gradient | Classical SciPy gradient and quasi-Newton methods |
 | `TrustNCG`, `TrustKrylov`, `TrustConstr`, `Dogleg`, `SR1` | Gradient | Trust-region and constrained SciPy methods |
 | `TNC`, `SLSQP`, `COBYQA`, `COBYLA` | Gradient | Bounded physical-space SciPy solvers |
-| `RandomSearch` | Evolutionary | Unbiased baseline, no hyperparameters |
 | `EvoxPSO` | Evolutionary | Swarm intelligence, many variants (CLPSO, CSO, ...) |
 | `EvoxES` | Evolutionary | CMA-ES, OpenES, XNES, and more (EvoX backend) |
 | `PyCMACMAES` | Evolutionary | Vanilla CMA-ES (pycma backend) |
@@ -526,6 +536,7 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `OmadsMADS`, `OmadsOrthoMADS` | Derivative-Free | MADS / OrthoMADS direct search (OMADS) |
 | `PDFOUOBYQA`, `PDFONEWUOA`, `PDFOLINCOA`, `PyBOBYQA` | Derivative-Free | Powell-style trust-region DFO (PDFO + Py-BOBYQA) |
 | `NelderMead`, `Powell` | Derivative-Free | SciPy classical simplex / direction-set search |
+| `RandomSearch` | Global Search | Unbiased, derivative-free global search baseline, no hyperparameters |
 | `BasinHopping`, `DualAnnealing` | Global Search | SciPy stochastic global optimization |
 | `NevergradOnePlusOne`, `NevergradTBPSA`, `NevergradNGOpt` | Evolutionary | Nevergrad rugged-landscape baselines |
 | `BotorchBO` | Surrogate | Sample-efficient Bayesian Optimization |
@@ -538,7 +549,7 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `ReSTIR` | Surrogate | GPU-native kNN surrogate, scales to 100k+ candidates |
 | `VAESampling` | Generative | Latent-space compression + BO |
-See [Algorithms](docs/Algorithms.md) for hyperparameter details and usage examples.
+See [Algorithms](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Algorithms) for hyperparameter details and usage examples.
 ---
@@ -554,7 +565,7 @@ Reference implementations worth reading:
 - `gradient_based/adam_gd.py`: gradient-based pattern (custom loop)
 - `gradient_based/optax/adam.py`: Optax wrapper pattern (minimal subclass)
 - `gradient_based/scipy/_common.py`: shared SciPy wrapper, caching, and budget handling
-- `evolutionary/random_search.py`: simplest batched example
+- `global_search/random_search.py`: simplest batched example
 - `evolutionary/evox_es.py`: wrapping an external library (EvoX/PyTorch)
 - `evolutionary/pycma_cmaes.py`: wrapping pycma (ask/tell, restart strategies)
 - `evolutionary/jax_es.py`: native JAX ES without external library
@@ -568,13 +579,13 @@ For in-depth documentation beyond this README:
 | Page | Content |
 |------|---------|
-| [Architecture Overview](docs/Architecture-Overview.md) | Design, module map, data-flow diagrams |
-| [Objective API Reference](docs/Objective-API-Reference.md) | Complete `Objective` class reference |
-| [Problems](docs/Problems.md) | Loss computation, parameter meanings, constraints |
-| [Algorithms](docs/Algorithms.md) | All built-in algorithms with hyperparameters |
-| [Implementing a New Algorithm](docs/Implementing-a-New-Algorithm.md) | Full step-by-step contributor tutorial |
-| [Benchmarking](docs/Benchmarking.md) | Running benchmarks, saving/loading results |
-| [Metrics Reference](docs/Metrics-Reference.md) | Every benchmark metric explained |
-| [Utilities & Helpers](docs/Utilities-and-Helpers.md) | `t2j`/`j2t`, CLI config, inverse sigmoid |
-| [Installation](docs/Installation.md) | Environment setup, GPU support, HPC notes |
-| [FAQ](docs/FAQ.md) | Common pitfalls and troubleshooting |
+| [Architecture Overview](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Architecture-Overview) | Design, module map, data-flow diagrams |
+| [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) | Complete `Objective` class reference |
+| [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) | Loss computation, parameter meanings, constraints |
+| [Algorithms](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Algorithms) | All built-in algorithms with hyperparameters |
+| [Implementing a New Algorithm](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Implementing-a-New-Algorithm) | Full step-by-step contributor tutorial |
+| [Benchmarking](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Benchmarking) | Running benchmarks, saving/loading results |
+| [Metrics Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Metrics-Reference) | Every benchmark metric explained |
+| [Utilities & Helpers](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Utilities-and-Helpers) | `t2j`/`j2t`, CLI config, inverse sigmoid |
+| [Installation](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Installation) | Environment setup, GPU support, HPC notes |
+| [FAQ](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/FAQ) | Common pitfalls and troubleshooting |

{dfbench-0.2.0 → dfbench-0.2.2}/README.md RENAMED Viewed

@@ -1,8 +1,17 @@
+> [!IMPORTANT]
+> **BETA RELEASE:** This repository is currently in beta. The official competition will start on **July ~7**.
 # Differometor Benchmark
+[![image](https://img.shields.io/pypi/v/dfbench.svg)](https://pypi.python.org/pypi/dfbench)
+[![image](https://img.shields.io/pypi/l/dfbench)](https://pypi.python.org/pypi/dfbench)
+[![image](https://img.shields.io/pypi/pyversions/dfbench.svg)](https://pypi.python.org/pypi/dfbench)
+[![Ruff](https://img.shields.io/endpoint?url=https://raw.githubusercontent.com/astral-sh/ruff/main/assets/badge/v2.json)](https://github.com/astral-sh/ruff)
+[![Tests](https://img.shields.io/github/actions/workflow/status/artificial-scientist-lab/Differometor-Benchmark/tests.yaml?branch=main&label=tests&logo=github)](https://github.com/artificial-scientist-lab/Differometor-Benchmark/actions/workflows/tests.yaml)
 A benchmarking framework for optimization algorithms on gravitational-wave detector design problems, built on top of the [Differometor](https://github.com/artificial-scientist-lab/Differometor) simulator.
-> **For detailed documentation, see the [Wiki](docs/Home.md).**
+> **For detailed documentation, see the [Wiki](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki).**
 ## TL;DR (I want to try my own algorithm)
@@ -42,7 +51,7 @@ obj.plot_loss()
 obj.save_run_to_file("my_run.npz")
 ```
-**A loss below 0 means your solution beats the real Voyager detector's sensitivity.** (On `VoyagerProblem` without physical constraints — you might be burning mirrors.)
+**A loss below 0 means your solution beats the real Voyager detector's sensitivity.** (On `VoyagerProblem` without physical constraints, you might be burning mirrors.)
 ### Evaluation Methods
@@ -64,8 +73,8 @@ The problems are JAX-based and differentiable up to second order. Use whichever
 ```python
 from dfbench import t2j, j2t
-params_jax = t2j(params_torch)       # Torch → JAX
-losses_torch = j2t(obj.vmap_value(params_jax))  # JAX → Torch
+params_jax = t2j(params_torch)       # Torch -> JAX
+losses_torch = j2t(obj.vmap_value(params_jax))  # JAX -> Torch
 ```
 This adds negligible overhead compared to the objective function itself.
@@ -81,11 +90,11 @@ This adds negligible overhead compared to the objective function itself.
 Both constrained problems accept a `power_penalty_fn(value, threshold)` callable to control how power-constraint violations are penalized.  Built-in presets: `squashed_relu_penalty` (default), `relu_penalty`, `zero_penalty`. Feel free to try own ones. The penalty function can also be swapped after construction via `obj.set_penalty_fn(fn)` (before `obj.start_logging()`), which re-traces the objective so the change takes effect. The unconstrained Voyager problems reject this call because they have no power-constraint path.
-Constrained problems also expose aux diagnostics. `obj.value_aux(params)` and `obj.vmap_value_aux(batch)` return `(loss, aux)` where `aux` carries the loss decomposition, a physical `is_feasible` flag, per-constraint violations, and the raw per-group power arrays. Enable `save` tokens like `is_feasible`, `power_values`, or the `aux` alias and the standard `obj.value` / `obj.value_and_grad` loop records the enabled aux fields in the same forward pass, with no code change. See [Problems](docs/Problems.md) for the aux schema and [Objective API Reference](docs/Objective-API-Reference.md) for the full method set.
+Constrained problems also expose aux diagnostics. `obj.value_aux(params)` and `obj.vmap_value_aux(batch)` return `(loss, aux)` where `aux` carries the loss decomposition, a physical `is_feasible` flag, per-constraint violations, and the raw per-group power arrays. Enable `save` tokens like `is_feasible`, `power_values`, or the `aux` alias and the standard `obj.value` / `obj.value_and_grad` loop records the enabled aux fields in the same forward pass, with no code change. See [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) for the aux schema and [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) for the full method set.
 All problems also support `bounds_overrides` (e.g. `{"tuning": (0, 45)}`) to narrow default property bounds and `signal_floor` to floor detector signal magnitudes before sensitivity normalization. `signal_floor` defaults to `1e-20`. Use `problem.print_bounds()` to inspect effective bounds.
-See [Problems](docs/Problems.md) for details on loss computation, parameter meanings, and constraints.
+See [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) for details on loss computation, parameter meanings, and constraints.
 ---
@@ -122,7 +131,7 @@ pip install -e .                         # CPU-only editable install
 pip install -e ".[all,cuda13,analysis]"  # All backends, CUDA 13, and analysis extras
 ```
-See [Installation](docs/Installation.md) for GPU setup details and HPC notes.
+See [Installation](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Installation) for GPU setup details and HPC notes.
 ---
@@ -133,8 +142,8 @@ OptimizationAlgorithm.optimize()
          │
          ▼
    ┌───────────┐      records losses, params, grads, timestamps
-   │ Objective │ ──►  enforces time / eval budgets
-   └─────┬─────┘      bounded ↔ unbounded sigmoid transform
+   │ Objective │ ──>  enforces time / eval budgets
+   └─────┬─────┘      bounded <-> unbounded sigmoid transform
          │
          ▼
   ContinuousProblem        (VoyagerProblem, VoyagerTuningProblem, ConstrainedVoyagerProblem, UIFOProblem)
@@ -145,7 +154,7 @@ OptimizationAlgorithm.optimize()
 **Design Idea:** Algorithms never create their own `Objective`, they receive a pre-configured one. This lets the benchmark harness or user script control budgets, seeds, and history settings uniformly. The algorithm only has to implement its optimization logic.
-See [Architecture Overview](docs/Architecture-Overview.md) for full design details.
+See [Architecture Overview](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Architecture-Overview) for full design details.
 ---
@@ -157,15 +166,15 @@ src/dfbench/
 │   ├── problem.py        # ContinuousProblem ABC + ProblemSpec
 │   ├── algorithm.py       # OptimizationAlgorithm ABC + AlgorithmType enum
 │   ├── objective.py       # Objective wrapper (central piece)
-│   └── utils.py           # torch↔jax conversion, inverse sigmoid
+│   └── utils.py           # torch<->jax conversion, inverse sigmoid
 ├── algorithms/
 │   ├── derivative_free/   # OMADS, PDFO/Py-BOBYQA, NelderMead, Powell
-│   ├── global_search/     # SciPy BasinHopping, DualAnnealing
-│   ├── evolutionary/      # RandomSearch, EvoxPSO, EvoxES, Nevergrad, CMA family
+│   ├── global_search/     # RandomSearch, SciPy BasinHopping, DualAnnealing
+│   ├── evolutionary/      # EvoxPSO, EvoxES, Nevergrad, CMA family
 │   ├── gradient_based/
 │   │   ├── optax/         # 34 Optax-based optimizers (OptaxAdam, OptaxLAMB, ...)
-│   │   ├── scipy/         # 13 SciPy-based optimizers (BFGS, TNC, SLSQP, …)
-│   │   ├── custom_jax.py  # Native-JAX custom/hybrid batch (SGLD, ASAM, GD→L-BFGS, …)
+│   │   ├── scipy/         # 13 SciPy-based optimizers (BFGS, TNC, SLSQP, ...)
+│   │   ├── custom_jax.py  # Native-JAX custom/hybrid batch (SGLD, ASAM, GD->L-BFGS, ...)
 │   │   └── *.py           # Custom-loop algorithms (AdamGD, LBFGSGD, SAGD, NAAdamGD, OptaxLBFGS)
 │   ├── surrogate_based/
 │   │   ├── botorch/       # BotorchBO, BotorchTuRBO, BotorchqNEI, BotorchqKG,
@@ -247,7 +256,7 @@ benchmark.print_summary(results)
 - `save_csv`: Writes a CSV with all metrics computed at evenly-spaced time points.
 - `save_run_data`: Persists raw loss/params/time histories to NPZ files for later re-evaluation.
-See [Benchmarking](docs/Benchmarking.md) for full configuration options and [Metrics Reference](docs/Metrics-Reference.md) for what gets computed.
+See [Benchmarking](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Benchmarking) for full configuration options and [Metrics Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Metrics-Reference) for what gets computed.
 ### Native-JAX Custom/Hybrid Batch
@@ -270,13 +279,13 @@ see `scripts/voyager_native_jax_custom_batch.py`.
 The interface is designed to make this as simple as possible. You write the optimization logic; `Objective` handles everything else (timing, logging, budget enforcement, file I/O).
-> **Full step-by-step tutorial:** [Implementing a New Algorithm](docs/Implementing-a-New-Algorithm.md)
+> **Full step-by-step tutorial:** [Implementing a New Algorithm](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Implementing-a-New-Algorithm)
 ### The Contract
 1. Subclass `OptimizationAlgorithm`
 2. Declare `algorithm_str` and `algorithm_type`
-3. Implement `optimize(objective, ...) → None`
+3. Implement `optimize(objective, ...) -> None`
 4. Use `Objective` for all function evaluations
 5. The `Objective` is mutated in place, thereby no return is needed
@@ -299,7 +308,7 @@ class MyAlgorithm(OptimizationAlgorithm):
     """My optimization algorithm."""
     algorithm_str = "my_algorithm"
-    algorithm_type = AlgorithmType.EVOLUTIONARY  # match one of the algorithms/ subfolders
+    algorithm_type = AlgorithmType.GLOBAL_SEARCH  # match one of the algorithms/ subfolders
     def __init__(self, batch_size: int = 50) -> None:
         """Algorithm-level meta-parameters that don't change between runs."""
@@ -405,7 +414,7 @@ obj.loss_history            # full loss history
 obj.time_steps              # elapsed time at each evaluation
 ```
-See [Objective API Reference](docs/Objective-API-Reference.md) for the complete interface.
+See [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) for the complete interface.
 ---
@@ -420,7 +429,6 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `BFGS`, `LBFGSB`, `NonlinearCG`, `NewtonCG` | Gradient | Classical SciPy gradient and quasi-Newton methods |
 | `TrustNCG`, `TrustKrylov`, `TrustConstr`, `Dogleg`, `SR1` | Gradient | Trust-region and constrained SciPy methods |
 | `TNC`, `SLSQP`, `COBYQA`, `COBYLA` | Gradient | Bounded physical-space SciPy solvers |
-| `RandomSearch` | Evolutionary | Unbiased baseline, no hyperparameters |
 | `EvoxPSO` | Evolutionary | Swarm intelligence, many variants (CLPSO, CSO, ...) |
 | `EvoxES` | Evolutionary | CMA-ES, OpenES, XNES, and more (EvoX backend) |
 | `PyCMACMAES` | Evolutionary | Vanilla CMA-ES (pycma backend) |
@@ -436,6 +444,7 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `OmadsMADS`, `OmadsOrthoMADS` | Derivative-Free | MADS / OrthoMADS direct search (OMADS) |
 | `PDFOUOBYQA`, `PDFONEWUOA`, `PDFOLINCOA`, `PyBOBYQA` | Derivative-Free | Powell-style trust-region DFO (PDFO + Py-BOBYQA) |
 | `NelderMead`, `Powell` | Derivative-Free | SciPy classical simplex / direction-set search |
+| `RandomSearch` | Global Search | Unbiased, derivative-free global search baseline, no hyperparameters |
 | `BasinHopping`, `DualAnnealing` | Global Search | SciPy stochastic global optimization |
 | `NevergradOnePlusOne`, `NevergradTBPSA`, `NevergradNGOpt` | Evolutionary | Nevergrad rugged-landscape baselines |
 | `BotorchBO` | Surrogate | Sample-efficient Bayesian Optimization |
@@ -448,7 +457,7 @@ See [Objective API Reference](docs/Objective-API-Reference.md) for the complete
 | `ReSTIR` | Surrogate | GPU-native kNN surrogate, scales to 100k+ candidates |
 | `VAESampling` | Generative | Latent-space compression + BO |
-See [Algorithms](docs/Algorithms.md) for hyperparameter details and usage examples.
+See [Algorithms](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Algorithms) for hyperparameter details and usage examples.
 ---
@@ -464,7 +473,7 @@ Reference implementations worth reading:
 - `gradient_based/adam_gd.py`: gradient-based pattern (custom loop)
 - `gradient_based/optax/adam.py`: Optax wrapper pattern (minimal subclass)
 - `gradient_based/scipy/_common.py`: shared SciPy wrapper, caching, and budget handling
-- `evolutionary/random_search.py`: simplest batched example
+- `global_search/random_search.py`: simplest batched example
 - `evolutionary/evox_es.py`: wrapping an external library (EvoX/PyTorch)
 - `evolutionary/pycma_cmaes.py`: wrapping pycma (ask/tell, restart strategies)
 - `evolutionary/jax_es.py`: native JAX ES without external library
@@ -478,13 +487,13 @@ For in-depth documentation beyond this README:
 | Page | Content |
 |------|---------|
-| [Architecture Overview](docs/Architecture-Overview.md) | Design, module map, data-flow diagrams |
-| [Objective API Reference](docs/Objective-API-Reference.md) | Complete `Objective` class reference |
-| [Problems](docs/Problems.md) | Loss computation, parameter meanings, constraints |
-| [Algorithms](docs/Algorithms.md) | All built-in algorithms with hyperparameters |
-| [Implementing a New Algorithm](docs/Implementing-a-New-Algorithm.md) | Full step-by-step contributor tutorial |
-| [Benchmarking](docs/Benchmarking.md) | Running benchmarks, saving/loading results |
-| [Metrics Reference](docs/Metrics-Reference.md) | Every benchmark metric explained |
-| [Utilities & Helpers](docs/Utilities-and-Helpers.md) | `t2j`/`j2t`, CLI config, inverse sigmoid |
-| [Installation](docs/Installation.md) | Environment setup, GPU support, HPC notes |
-| [FAQ](docs/FAQ.md) | Common pitfalls and troubleshooting |
+| [Architecture Overview](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Architecture-Overview) | Design, module map, data-flow diagrams |
+| [Objective API Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Objective-API-Reference) | Complete `Objective` class reference |
+| [Problems](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Problems) | Loss computation, parameter meanings, constraints |
+| [Algorithms](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Algorithms) | All built-in algorithms with hyperparameters |
+| [Implementing a New Algorithm](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Implementing-a-New-Algorithm) | Full step-by-step contributor tutorial |
+| [Benchmarking](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Benchmarking) | Running benchmarks, saving/loading results |
+| [Metrics Reference](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Metrics-Reference) | Every benchmark metric explained |
+| [Utilities & Helpers](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Utilities-and-Helpers) | `t2j`/`j2t`, CLI config, inverse sigmoid |
+| [Installation](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/Installation) | Environment setup, GPU support, HPC notes |
+| [FAQ](https://github.com/artificial-scientist-lab/Differometor-Benchmark/wiki/FAQ) | Common pitfalls and troubleshooting |

{dfbench-0.2.0 → dfbench-0.2.2}/docs/Algorithm-Status.md RENAMED Viewed

@@ -18,7 +18,7 @@
 ---
-## `EvoxES` — Multi-Algorithm Wrapper (on `main`)
+## `EvoxES`: Multi-Algorithm Wrapper (on `main`)
 `EvoxES` is a single class that wraps **12 distinct EvoX evolution strategies** selectable via its `variant` parameter.
 Each variant is a separate algorithm from the EvoX library.
@@ -38,7 +38,7 @@ Each variant is a separate algorithm from the EvoX library.
 | `GuidedES` | Guided Evolution Strategy | ES |
 | `ESMC` | Evolution Strategy with Monte Carlo | ES |
-## `EvoxPSO` — Multi-Algorithm Wrapper (on `main`)
+## `EvoxPSO`: Multi-Algorithm Wrapper (on `main`)
 `EvoxPSO` is a single class that wraps **7 distinct EvoX PSO variants** selectable via its `variant` parameter.
@@ -52,10 +52,10 @@ Each variant is a separate algorithm from the EvoX library.
 | `SLPSOGS` | Social Learning PSO with Gaussian Sampling |
 | `SLPSOUS` | Social Learning PSO with Uniform Sampling |
-## `EvosaxES` — Multi-Algorithm Wrapper (on `algorithm/evo-algorithms`)
+## `EvosaxES`: Multi-Algorithm Wrapper (on `algorithm/evo-algorithms`)
 `EvosaxES` is a separate evosax-backed wrapper (pure JAX, JIT-friendly) that wraps **14 variants**.
-Coexists with `EvoxES` — different backend, class names use `evosax_` prefix.
+Coexists with `EvoxES` but uses a different backend; class names use `evosax_` prefix.
 | Variant | Full Name | Type |
 |---------|-----------|------|
@@ -129,7 +129,7 @@ Coexists with `EvoxES` — different backend, class names use `evosax_` prefix.
 |-----------|-----------|--------|--------|-------|
 | SGLD | `SGLDJAX` | Branch | `algorithm/jax-algorithms` | Stochastic Gradient Langevin Dynamics |
 | ASAM | `ASAMJAX` | Branch | `algorithm/jax-algorithms` | Adaptive Sharpness-Aware Minimization |
-| Adam→L-BFGS Switch | `AdamToLBFGSJAX` | Branch | `algorithm/jax-algorithms` | Hybrid warm-start strategy |
+| Adam->L-BFGS Switch | `AdamToLBFGSJAX` | Branch | `algorithm/jax-algorithms` | Hybrid warm-start strategy |
 | Entropy-SGD | `EntropySGDJAX` | Branch | `algorithm/jax-algorithms` | |
 | SGHMC | `SGHMCJAX` | Branch | `algorithm/jax-algorithms` | Stochastic Gradient Hamiltonian MC |
 | ARC | `ARCJAX` | Branch | `algorithm/jax-algorithms` | Adaptive Regularization with Cubics |
@@ -146,7 +146,6 @@ Coexists with `EvoxES` — different backend, class names use `evosax_` prefix.
 |-----------|-----------|--------|--------|-------|
 | EvoX ES (12 variants) | `EvoxES` | **Finished** | `main` | See variant table above |
 | EvoX PSO (7 variants) | `EvoxPSO` | **Finished** | `main` | See variant table above |
-| Random Search | `RandomSearch` | **Finished** | `main` | Baseline |
 | Evosax ES (14 variants) | `EvosaxES` | Branch | `algorithm/evo-algorithms` | See variant table above |
 | DE | `DE` | Branch | `algorithm/evo-algorithms` | Differential Evolution (custom impl) |
 | JADE | `JADE` | Branch | `algorithm/evo-algorithms` | Adaptive DE |
@@ -201,6 +200,7 @@ Coexists with `EvoxES` — different backend, class names use `evosax_` prefix.
 |-----------|-----------|--------|--------|-------|
 | Basin Hopping | `BasinHopping` | Branch | `algorithm/scipy-nongrad-algorithms` | SciPy stochastic global optimizer |
 | Dual Annealing | `DualAnnealing` | Branch | `algorithm/scipy-nongrad-algorithms` | SciPy stochastic global optimizer |
+| Random Search | `RandomSearch` | **Finished** | `main` | Baseline |
 ### Surrogate-Based / Bayesian Optimization
@@ -214,7 +214,7 @@ Coexists with `EvoxES` — different backend, class names use `evosax_` prefix.
 | GEBO | `GEBO` | Branch | `algorithm/bo-algorithms` | Gradient-Enhanced BO |
 | LineBO | `LineBO` | Branch | `algorithm/bo-algorithms` | Line-search BO |
 | qKG | `BotorchqKG` | Branch | `algorithm/bo-algorithms` | Knowledge Gradient |
-| qNEI | `BotorchqNEI` | Branch | `algorithm/bo-algorithms` | Noisy Expected Improvement |
+| qNEI | `BotorchQNEI` | Branch | `algorithm/bo-algorithms` | Noisy Expected Improvement |
 | REMBO | `REMBO` | Branch | `algorithm/bo-algorithms` | Random Embedding BO |
 | HEBO | `HEBO` | Branch | `algorithm/bo-algorithms` | Heteroscedastic BO |
 | SMAC | `SMAC` | Branch | `algorithm/bo-algorithms` | Sequential Model-based Config |

dfbench 0.2.0__tar.gz → 0.2.2__tar.gz

dfbench 0.2.0tar.gz → 0.2.2tar.gz