xtensor-python/include/xtensor-python/pytensor.hpp at master · BeneSim/xtensor-python · GitHub

Repository navigation

History

100

101

102

103

104

105

106

107

108

109

110

111

112

113

114

115

116

117

118

119

120

121

122

123

124

125

126

127

128

129

130

131

132

133

134

135

136

137

138

139

140

141

142

143

144

145

146

147

148

149

150

151

152

153

154

155

156

157

158

159

160

161

162

163

164

165

166

167

168

169

170

171

172

173

174

175

176

177

178

179

180

181

182

183

184

185

186

187

188

189

190

191

192

193

194

195

196

197

198

199

200

201

202

203

204

205

206

207

208

209

210

211

212

213

214

215

216

217

218

219

220

221

222

223

224

225

226

227

228

229

230

231

232

233

234

235

236

237

238

239

240

241

242

243

244

245

246

247

248

249

250

251

252

253

254

255

256

257

258

259

260

261

262

263

264

265

266

267

268

269

270

271

272

273

274

275

276

277

278

279

280

281

282

283

284

285

286

287

288

289

290

291

292

293

294

295

296

297

298

299

300

301

302

303

304

305

306

307

308

309

310

311

312

313

314

315

316

317

318

319

320

321

322

323

324

325

326

327

328

329

330

331

332

333

334

335

336

337

338

339

340

341

342

343

344

345

346

347

348

349

350

351

352

353

354

355

356

357

358

359

360

361

362

363

364

365

366

367

368

369

370

371

372

373

374

375

376

377

378

379

380

381

382

383

384

385

386

387

388

389

390

391

392

393

394

395

396

397

398

399

400

401

402

403

404

405

406

407

408

409

410

411

412

413

414

415

416

417

418

419

420

421

422

423

424

425

426

427

428

429

430

431

432

433

434

435

436

437

438

439

440

441

442

443

444

445

446

447

448

449

450

451

452

453

454

455

456

457

458

459

460

461

462

463

464

465

466

467

468

469

470

471

472

473

474

475

476

477

478

479

480

481

482

483

484

485

486

487

488

489

490

491

492

493

494

495

496

497

498

499

500

501

502

503

504

505

506

507

/***************************************************************************

* *

* Distributed under the terms of the BSD 3-Clause License. *

* *

* The full license is in the file LICENSE, distributed with this software. *

****************************************************************************/

#ifndef PY_TENSOR_HPP

#define PY_TENSOR_HPP

#include <algorithm>

#include <array>

#include <cstddef>

#include "xtensor/xbuffer_adaptor.hpp"

#include "xtensor/xiterator.hpp"

#include "xtensor/xsemantic.hpp"

#include "xtensor/xutils.hpp"

#include "pycontainer.hpp"

#include "pystrides_adaptor.hpp"

#include "xtensor_type_caster_base.hpp"

namespace xt

{

template <class T, std::size_t N, layout_type L = layout_type::dynamic>

class pytensor;

}

namespace pybind11

{

namespace detail

{

template <class T, std::size_t N, xt::layout_type L>

struct handle_type_name<xt::pytensor<T, N, L>>

{

static PYBIND11_DESCR name()

{

return _("numpy.ndarray[") + npy_format_descriptor<T>::name() + _("]");

}

};

template <class T, std::size_t N, xt::layout_type L>

struct pyobject_caster<xt::pytensor<T, N, L>>

{

using type = xt::pytensor<T, N, L>;

bool load(handle src, bool convert)

{

if (!convert)

{

if (!xt::detail::check_array<T>(src))

{

return false;

}

value = type::ensure(src);

return static_cast<bool>(value);

}

static handle cast(const handle& src, return_value_policy, handle)

{

return src.inc_ref();

}

PYBIND11_TYPE_CASTER(type, handle_type_name<type>::name());

};

// Type caster for casting ndarray to xexpression<pytensor>

template <class T, std::size_t N, xt::layout_type L>

struct type_caster<xt::xexpression<xt::pytensor<T, N, L>>> : pyobject_caster<xt::pytensor<T, N, L>>

{

using Type = xt::xexpression<xt::pytensor<T, N, L>>;

operator Type&()

{

return this->value;

}

operator const Type&()

{

return this->value;

}

};

// Type caster for casting xt::xtensor to ndarray

template <class T, std::size_t N, xt::layout_type L>

struct type_caster<xt::xtensor<T, N, L>> : xtensor_type_caster_base<xt::xtensor<T, N, L>>

{

};

}

namespace xt

{

template <class T, std::size_t N, layout_type L>

struct xiterable_inner_types<pytensor<T, N, L>>

: xcontainer_iterable_types<pytensor<T, N, L>>

{

};

template <class T, std::size_t N, layout_type L>

struct xcontainer_inner_types<pytensor<T, N, L>>

{

using storage_type = xbuffer_adaptor<T*>;

using shape_type = std::array<npy_intp, N>;

using strides_type = shape_type;

using backstrides_type = shape_type;

using inner_shape_type = shape_type;

using inner_strides_type = strides_type;

using inner_backstrides_type = backstrides_type;

using temporary_type = pytensor<T, N, L>;

static constexpr layout_type layout = L;

};

/**

* @class pytensor

* @brief Multidimensional container providing the xtensor container semantics wrapping a numpy array.

* pytensor is similar to the xtensor container in that it has a static dimensionality.

* Unlike with the pyarray container, pytensor cannot be reshaped with a different number of dimensions

* and reshapes are not reflected on the Python side. However, pytensor has benefits compared to pyarray

* in terms of performances. pytensor shapes are stack-allocated which makes iteration upon pytensor

* faster than with pyarray.

* @tparam T The type of the element stored in the pyarray.

* @sa pyarray

template <class T, std::size_t N, layout_type L>

class pytensor : public pycontainer<pytensor<T, N, L>>,

public xcontainer_semantic<pytensor<T, N, L>>

{

public:

using self_type = pytensor<T, N, L>;

using semantic_base = xcontainer_semantic<self_type>;

using base_type = pycontainer<self_type>;

using storage_type = typename base_type::storage_type;

using value_type = typename base_type::value_type;

using reference = typename base_type::reference;

using const_reference = typename base_type::const_reference;

using pointer = typename base_type::pointer;

using size_type = typename base_type::size_type;

using shape_type = typename base_type::shape_type;

using strides_type = typename base_type::strides_type;

using backstrides_type = typename base_type::backstrides_type;

using inner_shape_type = typename base_type::inner_shape_type;

using inner_strides_type = typename base_type::inner_strides_type;

using inner_backstrides_type = typename base_type::inner_backstrides_type;

pytensor();

pytensor(nested_initializer_list_t<T, N> t);

pytensor(pybind11::handle h, pybind11::object::borrowed_t);

pytensor(pybind11::handle h, pybind11::object::stolen_t);

pytensor(const pybind11::object& o);

explicit pytensor(const shape_type& shape, layout_type l = layout_type::row_major);

explicit pytensor(const shape_type& shape, const_reference value, layout_type l = layout_type::row_major);

explicit pytensor(const shape_type& shape, const strides_type& strides, const_reference value);

explicit pytensor(const shape_type& shape, const strides_type& strides);

template <class S = shape_type>

static pytensor from_shape(S&& shape);

pytensor(const self_type& rhs);

self_type& operator=(const self_type& rhs);

pytensor(self_type&&) = default;

self_type& operator=(self_type&& e) = default;

template <class E>

pytensor(const xexpression<E>& e);

template <class E>

self_type& operator=(const xexpression<E>& e);

using base_type::begin;

using base_type::end;

static self_type ensure(pybind11::handle h);

static bool check_(pybind11::handle h);

private:

inner_shape_type m_shape;

inner_strides_type m_strides;

inner_backstrides_type m_backstrides;

storage_type m_storage;

void init_tensor(const shape_type& shape, const strides_type& strides);

void init_from_python();

inner_shape_type& shape_impl() noexcept;

const inner_shape_type& shape_impl() const noexcept;

inner_strides_type& strides_impl() noexcept;

const inner_strides_type& strides_impl() const noexcept;

inner_backstrides_type& backstrides_impl() noexcept;

const inner_backstrides_type& backstrides_impl() const noexcept;

storage_type& storage_impl() noexcept;

const storage_type& storage_impl() const noexcept;

friend class xcontainer<pytensor<T, N, L>>;

friend class pycontainer<pytensor<T, N, L>>;

};

/***************************

* pytensor implementation *

***************************/

/**

* @name Constructors

//@{

/**

* Allocates an uninitialized pytensor that holds 1 element.

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor()

: base_type()

{

m_shape = xtl::make_sequence<shape_type>(N, size_type(1));

m_strides = xtl::make_sequence<strides_type>(N, size_type(0));

init_tensor(m_shape, m_strides);

detail::default_initialize(m_storage);

}

/**

* Allocates a pytensor with a nested initializer list.

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(nested_initializer_list_t<T, N> t)

: base_type()

{

base_type::resize(xt::shape<shape_type>(t), layout_type::row_major);

nested_copy(m_storage.begin(), t);

}

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(pybind11::handle h, pybind11::object::borrowed_t b)

: base_type(h, b)

{

init_from_python();

}

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(pybind11::handle h, pybind11::object::stolen_t s)

: base_type(h, s)

{

init_from_python();

}

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const pybind11::object& o)

: base_type(o)

{

init_from_python();

}

/**

* Allocates an uninitialized pytensor with the specified shape and

* layout.

* @param shape the shape of the pytensor

* @param l the layout_type of the pytensor

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const shape_type& shape, layout_type l)

{

compute_strides(shape, l, m_strides);

init_tensor(shape, m_strides);

}

/**

* Allocates a pytensor with the specified shape and layout. Elements

* are initialized to the specified value.

* @param shape the shape of the pytensor

* @param value the value of the elements

* @param l the layout_type of the pytensor

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const shape_type& shape,

const_reference value,

layout_type l)

{

compute_strides(shape, l, m_strides);

init_tensor(shape, m_strides);

std::fill(m_storage.begin(), m_storage.end(), value);

}

/**

* Allocates an uninitialized pytensor with the specified shape and strides.

* Elements are initialized to the specified value.

* @param shape the shape of the pytensor

* @param strides the strides of the pytensor

* @param value the value of the elements

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const shape_type& shape,

const strides_type& strides,

const_reference value)

{

init_tensor(shape, strides);

std::fill(m_storage.begin(), m_storage.end(), value);

}

/**

* Allocates an uninitialized pytensor with the specified shape and strides.

* @param shape the shape of the pytensor

* @param strides the strides of the pytensor

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const shape_type& shape,

const strides_type& strides)

{

init_tensor(shape, strides);

}

/**

* Allocates and returns an pytensor with the specified shape.

* @param shape the shape of the pytensor

template <class T, std::size_t N, layout_type L>

template <class S>

inline pytensor<T, N, L> pytensor<T, N, L>::from_shape(S&& shape)

{

detail::check_dims<shape_type>::run(shape.size());

auto shp = xtl::forward_sequence<shape_type>(shape);

return self_type(shp);

}

//@}

/**

* @name Copy semantic

//@{

/**

* The copy constructor.

template <class T, std::size_t N, layout_type L>

inline pytensor<T, N, L>::pytensor(const self_type& rhs)

: base_type(), semantic_base(rhs)

{

init_tensor(rhs.shape(), rhs.strides());

std::copy(rhs.storage().cbegin(), rhs.storage().cend(), this->storage().begin());

}

/**

* The assignment operator.

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::operator=(const self_type& rhs) -> self_type&

{

self_type tmp(rhs);

*this = std::move(tmp);

return *this;

}

//@}

/**

* @name Extended copy semantic

//@{

/**

* The extended copy constructor.

template <class T, std::size_t N, layout_type L>

template <class E>

inline pytensor<T, N, L>::pytensor(const xexpression<E>& e)

: base_type()

{

shape_type shape = xtl::forward_sequence<shape_type>(e.derived_cast().shape());

strides_type strides = xtl::make_sequence<strides_type>(N, size_type(0));

compute_strides(shape, layout_type::row_major, strides);

init_tensor(shape, strides);

semantic_base::assign(e);

}

/**

* The extended assignment operator.

template <class T, std::size_t N, layout_type L>

template <class E>

inline auto pytensor<T, N, L>::operator=(const xexpression<E>& e) -> self_type&

{

return semantic_base::operator=(e);

}

//@}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::ensure(pybind11::handle h) -> self_type

{

return base_type::ensure(h);

}

template <class T, std::size_t N, layout_type L>

inline bool pytensor<T, N, L>::check_(pybind11::handle h)

{

return base_type::check_(h);

}

template <class T, std::size_t N, layout_type L>

inline void pytensor<T, N, L>::init_tensor(const shape_type& shape, const strides_type& strides)

{

npy_intp python_strides[N];

std::transform(strides.begin(), strides.end(), python_strides,

[](auto v) { return sizeof(T) * v; });

int flags = NPY_ARRAY_ALIGNED;

if (!std::is_const<T>::value)

{

flags |= NPY_ARRAY_WRITEABLE;

}

auto dtype = pybind11::detail::npy_format_descriptor<T>::dtype();

auto tmp = pybind11::reinterpret_steal<pybind11::object>(

PyArray_NewFromDescr(&PyArray_Type, (PyArray_Descr*) dtype.release().ptr(), static_cast<int>(shape.size()),

const_cast<npy_intp*>(shape.data()), python_strides,

nullptr, flags, nullptr));

if (!tmp)

{

throw std::runtime_error("NumPy: unable to create ndarray");

}

this->m_ptr = tmp.release().ptr();

m_shape = shape;

m_strides = strides;

adapt_strides(m_shape, m_strides, m_backstrides);

m_storage = storage_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),

static_cast<size_type>(PyArray_SIZE(this->python_array())));

}

template <class T, std::size_t N, layout_type L>

inline void pytensor<T, N, L>::init_from_python()

{

if (PyArray_NDIM(this->python_array()) != N)

{

throw std::runtime_error("NumPy: ndarray has incorrect number of dimensions");

}

std::copy(PyArray_DIMS(this->python_array()), PyArray_DIMS(this->python_array()) + N, m_shape.begin());

std::transform(PyArray_STRIDES(this->python_array()), PyArray_STRIDES(this->python_array()) + N, m_strides.begin(),

[](auto v) { return v / sizeof(T); });

adapt_strides(m_shape, m_strides, m_backstrides);

if (L != layout_type::dynamic && !do_strides_match(m_shape, m_strides, L))

{

throw std::runtime_error("NumPy: passing container with bad strides for layout (is it a view?).");

}

m_storage = storage_type(reinterpret_cast<pointer>(PyArray_DATA(this->python_array())),

this->get_min_stride() * static_cast<size_type>(PyArray_SIZE(this->python_array())));

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::shape_impl() noexcept -> inner_shape_type&

{

return m_shape;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::shape_impl() const noexcept -> const inner_shape_type&

{

return m_shape;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::strides_impl() noexcept -> inner_strides_type&

{

return m_strides;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::strides_impl() const noexcept -> const inner_strides_type&

{

return m_strides;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::backstrides_impl() noexcept -> inner_backstrides_type&

{

return m_backstrides;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::backstrides_impl() const noexcept -> const inner_backstrides_type&

{

return m_backstrides;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::storage_impl() noexcept -> storage_type&

{

return m_storage;

}

template <class T, std::size_t N, layout_type L>

inline auto pytensor<T, N, L>::storage_impl() const noexcept -> const storage_type&

{

return m_storage;

}

#endif

xtensor-python/include/xtensor-python/pytensor.hpp at master · BeneSim/xtensor-python · GitHub

Repository navigation

Expand file tree

Breadcrumbs

pytensor.hpp

Latest commit

History

Breadcrumbs

pytensor.hpp

File metadata and controls