rok commented on code in PR #8510: URL: https://github.com/apache/arrow/pull/8510#discussion_r1124552242
########## cpp/src/arrow/extension/fixed_shape_tensor.cc: ########## @@ -0,0 +1,267 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +#include "arrow/extension/fixed_shape_tensor.h" + +#include "arrow/array/array_nested.h" +#include "arrow/array/array_primitive.h" +#include "arrow/json/rapidjson_defs.h" // IWYU pragma: keep +#include "arrow/tensor.h" +#include "arrow/util/logging.h" +#include "arrow/util/sort.h" + +#include <rapidjson/document.h> +#include <rapidjson/writer.h> + +namespace rj = arrow::rapidjson; + +namespace arrow { +namespace extension { + +bool FixedShapeTensorType::ExtensionEquals(const ExtensionType& other) const { + if (extension_name() != other.extension_name()) { + return false; + } + const auto& other_ext = static_cast<const FixedShapeTensorType&>(other); + bool equals = storage_type()->Equals(other_ext.storage_type()); + equals &= shape_ == other_ext.shape(); + equals &= permutation_ == other_ext.permutation(); + equals &= dim_names_ == other_ext.dim_names(); + return equals; +} + +std::string FixedShapeTensorType::Serialize() const { + rj::Document document; + document.SetObject(); + rj::Document::AllocatorType& allocator = document.GetAllocator(); + + rj::Value shape(rj::kArrayType); + for (auto v : shape_) { + shape.PushBack(v, allocator); + } + document.AddMember(rj::Value("shape", allocator), shape, allocator); + + if (!permutation_.empty()) { + rj::Value permutation(rj::kArrayType); + for (auto v : permutation_) { + permutation.PushBack(v, allocator); + } + document.AddMember(rj::Value("permutation", allocator), permutation, allocator); + } + + if (!dim_names_.empty()) { + rj::Value dim_names(rj::kArrayType); + for (std::string v : dim_names_) { + dim_names.PushBack(rj::Value{}.SetString(v.c_str(), allocator), allocator); + } + document.AddMember(rj::Value("dim_names", allocator), dim_names, allocator); + } + + rj::StringBuffer buffer; + rj::Writer<rj::StringBuffer> writer(buffer); + document.Accept(writer); + return buffer.GetString(); +} + +Result<std::shared_ptr<DataType>> FixedShapeTensorType::Deserialize( + std::shared_ptr<DataType> storage_type, const std::string& serialized_data) const { + if (storage_type->id() != Type::FIXED_SIZE_LIST) { + return Status::Invalid("Expected FixedSizeList storage type, got ", + storage_type->ToString()); + } + auto value_type = + internal::checked_pointer_cast<FixedSizeListType>(storage_type)->value_type(); + rj::Document document; + if (document.Parse(serialized_data.data(), serialized_data.length()).HasParseError() || + !document.HasMember("shape") || !document["shape"].IsArray()) { + return Status::Invalid("Invalid serialized JSON data: ", serialized_data); + } + + std::vector<int64_t> shape; + for (auto& x : document["shape"].GetArray()) { + shape.emplace_back(x.GetInt64()); + } + std::vector<int64_t> permutation; + if (document.HasMember("permutation")) { + for (auto& x : document["permutation"].GetArray()) { + permutation.emplace_back(x.GetInt64()); + } + if (shape.size() != permutation.size()) { + return Status::Invalid("Invalid permutation"); + } + } + std::vector<std::string> dim_names; + if (document.HasMember("dim_names")) { + for (auto& x : document["dim_names"].GetArray()) { + dim_names.emplace_back(x.GetString()); + } + if (shape.size() != dim_names.size()) { + return Status::Invalid("Invalid dim_names"); + } + } + + return fixed_shape_tensor(value_type, shape, permutation, dim_names); +} + +std::shared_ptr<Array> FixedShapeTensorType::MakeArray( + std::shared_ptr<ArrayData> data) const { + return std::make_shared<ExtensionArray>(data); +} + +Result<std::shared_ptr<Array>> FixedShapeTensorType::MakeArray( + std::shared_ptr<Tensor> tensor) const { + auto permutation = internal::ArgSort(tensor->strides()); + std::reverse(permutation.begin(), permutation.end()); + if (permutation[0] != 0) { + return Status::Invalid( + "Only first-major tensors can be zero-copy converted to arrays"); + } + + auto cell_shape = tensor->shape(); + cell_shape.erase(cell_shape.begin()); + if (cell_shape != shape_) { + return Status::Invalid("Expected cell shape does not match input tensor shape"); + } + + permutation.erase(permutation.begin()); + for (auto& x : permutation) { + x--; + } + + auto ext_type = + fixed_shape_tensor(tensor->type(), cell_shape, permutation, tensor->dim_names()); + + std::shared_ptr<FixedSizeListArray> arr; + std::shared_ptr<Array> value_array; + switch (tensor->type_id()) { + case Type::UINT8: { + value_array = std::make_shared<UInt8Array>(tensor->size(), tensor->data()); + break; + } + case Type::INT8: { + value_array = std::make_shared<Int8Array>(tensor->size(), tensor->data()); + break; + } + case Type::UINT16: { + value_array = std::make_shared<UInt16Array>(tensor->size(), tensor->data()); + break; + } + case Type::INT16: { + value_array = std::make_shared<Int16Array>(tensor->size(), tensor->data()); + break; + } + case Type::UINT32: { + value_array = std::make_shared<UInt32Array>(tensor->size(), tensor->data()); + break; + } + case Type::INT32: { + value_array = std::make_shared<Int32Array>(tensor->size(), tensor->data()); + break; + } + case Type::UINT64: { + value_array = std::make_shared<Int64Array>(tensor->size(), tensor->data()); + break; + } + case Type::INT64: { + value_array = std::make_shared<Int64Array>(tensor->size(), tensor->data()); + break; + } + case Type::HALF_FLOAT: { + value_array = std::make_shared<HalfFloatArray>(tensor->size(), tensor->data()); + break; + } + case Type::FLOAT: { + value_array = std::make_shared<FloatArray>(tensor->size(), tensor->data()); + break; + } + case Type::DOUBLE: { + value_array = std::make_shared<DoubleArray>(tensor->size(), tensor->data()); + break; + } + default: { + return Status::NotImplemented("Unsupported tensor type: ", + tensor->type()->ToString()); + } + } + arr = std::make_shared<FixedSizeListArray>(ext_type->storage_type(), tensor->shape()[0], + value_array); + auto ext_data = arr->data(); + ext_data->type = ext_type; + return MakeArray(ext_data); +} + +Result<std::shared_ptr<Tensor>> FixedShapeTensorType::ToTensor( + std::shared_ptr<Array> arr) const { + // To convert an array of n dimensional tensors to a n+1 dimensional tensor we + // interpret the array's length as the first dimension the new tensor. Further, we + // define n+1 dimensional tensor's strides by front appending a new stride to the n + // dimensional tensor's strides. + + ARROW_DCHECK_EQ(arr->null_count(), 0) << "Null values not supported in tensors."; + auto ext_arr = internal::checked_pointer_cast<FixedSizeListArray>( + internal::checked_pointer_cast<ExtensionArray>(arr)->storage()); + + std::vector<int64_t> shape = shape_; + shape.insert(shape.begin(), 1, arr->length()); + + std::vector<int64_t> tensor_strides = strides(); + tensor_strides.insert(tensor_strides.begin(), 1, arr->length() * tensor_strides[0]); + + std::shared_ptr<Buffer> buffer = ext_arr->values()->data()->buffers[1]; + return *Tensor::Make(ext_arr->value_type(), buffer, shape, tensor_strides, dim_names()); +} + +std::shared_ptr<FixedShapeTensorType> fixed_shape_tensor( + const std::shared_ptr<DataType>& value_type, const std::vector<int64_t>& shape, + const std::vector<int64_t>& permutation, const std::vector<std::string>& dim_names) { + ARROW_CHECK(is_tensor_supported(value_type->id())); + + if (!permutation.empty()) { + ARROW_CHECK_EQ(shape.size(), permutation.size()) + << "permutation.size() == " << permutation.size() + << " must be empty or have the same length as shape.size() " << shape.size(); + } + if (!dim_names.empty()) { + ARROW_CHECK_EQ(shape.size(), dim_names.size()) Review Comment: I'm not sure how to assert for these in tests. `ASSERT_DEATH` would work but it seems it's not advised. We also can't change the return type of the function so we can't return `Status` or `Result`. ########## cpp/src/arrow/extension/fixed_shape_tensor_test.cc: ########## @@ -0,0 +1,293 @@ +// Licensed to the Apache Software Foundation (ASF) under one +// or more contributor license agreements. See the NOTICE file +// distributed with this work for additional information +// regarding copyright ownership. The ASF licenses this file +// to you under the Apache License, Version 2.0 (the +// "License"); you may not use this file except in compliance +// with the License. You may obtain a copy of the License at +// +// http://www.apache.org/licenses/LICENSE-2.0 +// +// Unless required by applicable law or agreed to in writing, +// software distributed under the License is distributed on an +// "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY +// KIND, either express or implied. See the License for the +// specific language governing permissions and limitations +// under the License. + +#include "arrow/extension/fixed_shape_tensor.h" + +#include "arrow/testing/matchers.h" + +#include "arrow/array/array_nested.h" +#include "arrow/array/array_primitive.h" +#include "arrow/io/memory.h" +#include "arrow/ipc/reader.h" +#include "arrow/ipc/writer.h" +#include "arrow/record_batch.h" +#include "arrow/tensor.h" +#include "arrow/testing/gtest_util.h" +#include "arrow/util/key_value_metadata.h" + +namespace arrow { + +using FixedShapeTensorType = extension::FixedShapeTensorType; +using extension::fixed_shape_tensor; + +class TestExtensionType : public ::testing::Test { + public: + void SetUp() override { + shape_ = {3, 3, 4}; + cell_shape_ = {3, 4}; + value_type_ = int64(); + cell_type_ = fixed_size_list(value_type_, 12); + dim_names_ = {"x", "y"}; + ext_type_ = fixed_shape_tensor(value_type_, cell_shape_, {}, dim_names_); + values_ = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, + 18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35}; + values_partial_ = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, + 12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23}; + shape_partial_ = {2, 3, 4}; + tensor_strides_ = {96, 32, 8}; + cell_strides_ = {32, 8}; + serialized_ = R"({"shape":[3,4],"dim_names":["x","y"]})"; + } + + protected: + std::vector<int64_t> shape_; + std::vector<int64_t> shape_partial_; + std::vector<int64_t> cell_shape_; + std::shared_ptr<DataType> value_type_; + std::shared_ptr<DataType> cell_type_; + std::vector<std::string> dim_names_; + std::shared_ptr<ExtensionType> ext_type_; + std::vector<int64_t> values_; + std::vector<int64_t> values_partial_; + std::vector<int64_t> tensor_strides_; + std::vector<int64_t> cell_strides_; + std::string serialized_; +}; + +auto RoundtripBatch = [](const std::shared_ptr<RecordBatch>& batch, + std::shared_ptr<RecordBatch>* out) { + ASSERT_OK_AND_ASSIGN(auto out_stream, io::BufferOutputStream::Create()); + ASSERT_OK(ipc::WriteRecordBatchStream({batch}, ipc::IpcWriteOptions::Defaults(), + out_stream.get())); + + ASSERT_OK_AND_ASSIGN(auto complete_ipc_stream, out_stream->Finish()); + + io::BufferReader reader(complete_ipc_stream); + std::shared_ptr<RecordBatchReader> batch_reader; + ASSERT_OK_AND_ASSIGN(batch_reader, ipc::RecordBatchStreamReader::Open(&reader)); + ASSERT_OK(batch_reader->ReadNext(out)); +}; + +TEST_F(TestExtensionType, CheckDummyRegistration) { + // We need a dummy registration at runtime to allow for IPC deserialization + auto ext_type = fixed_shape_tensor(int64(), {}); + auto registered_type = GetExtensionType(ext_type->extension_name()); + ASSERT_TRUE(registered_type->Equals(*ext_type)); +} + +TEST_F(TestExtensionType, CreateExtensionType) { + auto exact_ext_type = internal::checked_pointer_cast<FixedShapeTensorType>(ext_type_); + + // Test ExtensionType methods + ASSERT_EQ(ext_type_->extension_name(), "arrow.fixed_shape_tensor"); + ASSERT_TRUE(ext_type_->Equals(*exact_ext_type)); + ASSERT_TRUE(ext_type_->storage_type()->Equals(*cell_type_)); + ASSERT_EQ(ext_type_->Serialize(), serialized_); + ASSERT_OK_AND_ASSIGN(auto ds, + ext_type_->Deserialize(ext_type_->storage_type(), serialized_)); + auto deserialized = std::reinterpret_pointer_cast<ExtensionType>(ds); + ASSERT_TRUE(deserialized->Equals(*ext_type_)); + + // Test FixedShapeTensorType methods + ASSERT_EQ(exact_ext_type->id(), Type::EXTENSION); + ASSERT_EQ(exact_ext_type->ndim(), cell_shape_.size()); + ASSERT_EQ(exact_ext_type->shape(), cell_shape_); + ASSERT_EQ(exact_ext_type->strides(), cell_strides_); + ASSERT_EQ(exact_ext_type->dim_names(), dim_names_); +} + +TEST_F(TestExtensionType, CreateFromArray) { + auto exact_ext_type = internal::checked_pointer_cast<FixedShapeTensorType>(ext_type_); + + std::vector<std::shared_ptr<Buffer>> buffers = {nullptr, Buffer::Wrap(values_)}; + auto arr_data = std::make_shared<ArrayData>(value_type_, values_.size(), buffers, 0, 0); + auto arr = std::make_shared<Int64Array>(arr_data); + EXPECT_OK_AND_ASSIGN(auto fsla_arr, FixedSizeListArray::FromArrays(arr, cell_type_)); + auto data = fsla_arr->data(); + data->type = ext_type_; + auto ext_arr = exact_ext_type->MakeArray(data); + ASSERT_EQ(ext_arr->length(), shape_[0]); + ASSERT_EQ(ext_arr->null_count(), 0); +} + +TEST_F(TestExtensionType, CreateFromTensor) { + std::vector<int64_t> column_major_strides = {8, 24, 72}; + std::vector<int64_t> neither_major_strides = {96, 8, 32}; + + ASSERT_OK_AND_ASSIGN(auto tensor, + Tensor::Make(value_type_, Buffer::Wrap(values_), shape_)); + + auto exact_ext_type = internal::checked_pointer_cast<FixedShapeTensorType>(ext_type_); + EXPECT_OK_AND_ASSIGN(auto ext_arr, exact_ext_type->MakeArray(tensor)); + + ASSERT_OK(ext_arr->ValidateFull()); + ASSERT_TRUE(tensor->is_row_major()); + ASSERT_EQ(tensor->strides(), tensor_strides_); + ASSERT_EQ(ext_arr->length(), shape_[0]); + + auto ext_type_2 = fixed_shape_tensor(int64(), {3, 4}, {0, 1}); + EXPECT_OK_AND_ASSIGN(auto ext_arr_2, ext_type_2->MakeArray(tensor)); + + ASSERT_OK_AND_ASSIGN( + auto column_major_tensor, + Tensor::Make(value_type_, Buffer::Wrap(values_), shape_, column_major_strides)); + auto ext_type_3 = fixed_shape_tensor(int64(), {3, 4}, {0, 1}); + EXPECT_RAISES_WITH_MESSAGE_THAT( + Invalid, + testing::HasSubstr( + "Invalid: Only first-major tensors can be zero-copy converted to arrays"), + ext_type_3->MakeArray(column_major_tensor)); + ASSERT_THAT(ext_type_3->MakeArray(column_major_tensor), Raises(StatusCode::Invalid)); + + auto neither_major_tensor = std::make_shared<Tensor>(value_type_, Buffer::Wrap(values_), + shape_, neither_major_strides); + auto ext_type_4 = fixed_shape_tensor(int64(), {3, 4}, {1, 0}); + ASSERT_OK_AND_ASSIGN(auto ext_arr_4, ext_type_4->MakeArray(neither_major_tensor)); +} + +TEST_F(TestExtensionType, RoundtripTensor) { + ASSERT_OK_AND_ASSIGN(auto tensor, + Tensor::Make(value_type_, Buffer::Wrap(values_), shape_)); + auto exact_ext_type = internal::checked_pointer_cast<FixedShapeTensorType>(ext_type_); + EXPECT_OK_AND_ASSIGN(auto ext_arr, exact_ext_type->MakeArray(tensor)); + + EXPECT_OK_AND_ASSIGN(auto tensor_from_array, exact_ext_type->ToTensor(ext_arr)); + ASSERT_EQ(tensor_from_array->shape(), tensor->shape()); + ASSERT_EQ(tensor_from_array->strides(), tensor->strides()); + ASSERT_TRUE(tensor->Equals(*tensor_from_array)); +} + +TEST_F(TestExtensionType, SliceTensor) { + ASSERT_OK_AND_ASSIGN(auto tensor, + Tensor::Make(value_type_, Buffer::Wrap(values_), shape_)); + ASSERT_OK_AND_ASSIGN( + auto tensor_partial, + Tensor::Make(value_type_, Buffer::Wrap(values_partial_), shape_partial_)); + ASSERT_EQ(tensor->strides(), tensor_strides_); + ASSERT_EQ(tensor_partial->strides(), tensor_strides_); + auto ext_type = fixed_shape_tensor(value_type_, cell_shape_, {}, dim_names_); + auto exact_ext_type = internal::checked_pointer_cast<FixedShapeTensorType>(ext_type_); + + EXPECT_OK_AND_ASSIGN(auto ext_arr, exact_ext_type->MakeArray(tensor)); + EXPECT_OK_AND_ASSIGN(auto ext_arr_partial, exact_ext_type->MakeArray(tensor_partial)); + ASSERT_OK(ext_arr->ValidateFull()); + ASSERT_OK(ext_arr_partial->ValidateFull()); + + auto sliced = internal::checked_pointer_cast<ExtensionArray>(ext_arr->Slice(0, 2)); + auto partial = internal::checked_pointer_cast<ExtensionArray>(ext_arr_partial); + + ASSERT_TRUE(sliced->Equals(*partial)); + ASSERT_OK(sliced->ValidateFull()); + ASSERT_OK(partial->ValidateFull()); + ASSERT_TRUE(sliced->storage()->Equals(*partial->storage())); + ASSERT_EQ(sliced->length(), partial->length()); +} + +void CheckSerializationRoundtrip(const std::shared_ptr<ExtensionType>& ext_type) { + auto serialized = ext_type->Serialize(); + ASSERT_OK_AND_ASSIGN(auto deserialized, + ext_type->Deserialize(ext_type->storage_type(), serialized)); + ASSERT_TRUE(ext_type->Equals(*deserialized)); +} + +TEST_F(TestExtensionType, MetadataSerializationRoundtrip) { + CheckSerializationRoundtrip(fixed_shape_tensor(value_type_, {}, {}, {})); + CheckSerializationRoundtrip(fixed_shape_tensor(value_type_, {0}, {}, {})); + CheckSerializationRoundtrip(fixed_shape_tensor(value_type_, {1}, {0}, {"x"})); + CheckSerializationRoundtrip( + fixed_shape_tensor(value_type_, {256, 256, 3}, {0, 1, 2}, {"H", "W", "C"})); + CheckSerializationRoundtrip( + fixed_shape_tensor(value_type_, {256, 256, 3}, {2, 0, 1}, {"C", "H", "W"})); + + auto ext_type = fixed_shape_tensor(value_type_, cell_shape_, {0, 1}, dim_names_); + CheckSerializationRoundtrip(ext_type_); + + EXPECT_RAISES_WITH_MESSAGE_THAT( + Invalid, testing::HasSubstr("Invalid: Expected FixedSizeList storage type"), + ext_type->Deserialize(boolean(), serialized_)); Review Comment: Added. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
