# Copyright (c) 2026 PaddlePaddle Authors. All Rights Reserved. # # Licensed 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. """ FFT操作单元测试 / FFT Operations Unit Tests 测试目标 / Test Target: paddle.fft 模块 (python/paddle/fft.py, 覆盖率约57.8%) 覆盖的模块 / Covered Modules: - paddle.fft: FFT变换相关API - 包括 fft, ifft, rfft, irfft, fft2, ifft2, fftn, ifftn等 作用 / Purpose: 覆盖fast Fourier transform (快速傅里叶变换)及相关逆变换的代码路径， 补充未被原有测试覆盖的频域变换函数。 """ import unittest import numpy as np import paddle import paddle.fft paddle.disable_static() class TestFFTBasic(unittest.TestCase): """测试基本FFT操作 / Test basic FFT operations""" def setUp(self): """初始化测试数据 / Initialize test data""" self.x_real = paddle.to_tensor(np.random.randn(8).astype('float32')) self.x_2d = paddle.to_tensor(np.random.randn(4, 8).astype('float32')) self.x_complex = paddle.to_tensor( (np.random.randn(8) + 1j * np.random.randn(8)).astype('complex64') ) def test_fft_1d(self): """测试1D FFT变换 / Test 1D FFT transform""" result = paddle.fft.fft(self.x_complex) self.assertEqual(result.shape, self.x_complex.shape) self.assertTrue(result.dtype == paddle.complex64) def test_fft_with_n(self): """测试带长度参数的FFT / Test FFT with n parameter""" result = paddle.fft.fft(self.x_complex, n=16) self.assertEqual(result.shape[-1], 16) def test_fft_with_axis(self): """测试沿指定轴的FFT / Test FFT along specified axis""" x = paddle.to_tensor( (np.random.randn(4, 8) + 1j * np.random.randn(4, 8)).astype( 'complex64' ) ) result = paddle.fft.fft(x, axis=0) self.assertEqual(result.shape, x.shape) result2 = paddle.fft.fft(x, axis=1) self.assertEqual(result2.shape, x.shape) def test_ifft_1d(self): """测试1D IFFT逆变换 / Test 1D IFFT inverse transform""" fft_result = paddle.fft.fft(self.x_complex) recovered = paddle.fft.ifft(fft_result) self.assertEqual(recovered.shape, self.x_complex.shape) def test_rfft_1d(self): """测试实数FFT / Test real FFT""" result = paddle.fft.rfft(self.x_real) # rfft output has n//2+1 complex elements expected_len = self.x_real.shape[-1] // 2 + 1 self.assertEqual(result.shape[-1], expected_len) self.assertTrue(result.dtype in [paddle.complex64, paddle.complex128]) def test_irfft_1d(self): """测试实数IFFT逆变换 / Test real IFFT inverse transform""" rfft_result = paddle.fft.rfft(self.x_real) recovered = paddle.fft.irfft(rfft_result) self.assertEqual(recovered.shape[-1], self.x_real.shape[-1]) self.assertTrue(recovered.dtype in [paddle.float32, paddle.float64]) def test_rfft_with_n(self): """测试带长度参数的实数FFT / Test real FFT with n""" result = paddle.fft.rfft(self.x_real, n=16) self.assertEqual(result.shape[-1], 16 // 2 + 1) class TestFFT2D(unittest.TestCase): """测试2D FFT操作 / Test 2D FFT operations""" def setUp(self): """初始化2D测试数据 / Initialize 2D test data""" self.x_real_2d = paddle.to_tensor( np.random.randn(4, 8).astype('float32') ) self.x_complex_2d = paddle.to_tensor( (np.random.randn(4, 8) + 1j * np.random.randn(4, 8)).astype( 'complex64' ) ) def test_fft2(self): """测试2D FFT / Test 2D FFT""" result = paddle.fft.fft2(self.x_complex_2d) self.assertEqual(result.shape, self.x_complex_2d.shape) def test_ifft2(self): """测试2D IFFT / Test 2D IFFT""" fft_result = paddle.fft.fft2(self.x_complex_2d) recovered = paddle.fft.ifft2(fft_result) self.assertEqual(recovered.shape, self.x_complex_2d.shape) def test_rfft2(self): """测试2D 实数FFT / Test 2D real FFT""" result = paddle.fft.rfft2(self.x_real_2d) # Last dim: n//2+1 expected_last = self.x_real_2d.shape[-1] // 2 + 1 self.assertEqual(result.shape[-1], expected_last) self.assertEqual(result.shape[-2], self.x_real_2d.shape[-2]) def test_irfft2(self): """测试2D 实数IFFT / Test 2D real IFFT""" rfft_result = paddle.fft.rfft2(self.x_real_2d) recovered = paddle.fft.irfft2(rfft_result) self.assertEqual(recovered.shape[-1], self.x_real_2d.shape[-1]) self.assertEqual(recovered.shape[-2], self.x_real_2d.shape[-2]) def test_fft2_with_s(self): """测试带s参数的2D FFT / Test 2D FFT with s parameter""" result = paddle.fft.fft2(self.x_complex_2d, s=[8, 16]) self.assertEqual(result.shape[-2], 8) self.assertEqual(result.shape[-1], 16) class TestFFTND(unittest.TestCase): """测试N维FFT操作 / Test N-dimensional FFT operations""" def setUp(self): """初始化多维测试数据 / Initialize multi-dimensional test data""" self.x_real_3d = paddle.to_tensor( np.random.randn(2, 4, 8).astype('float32') ) self.x_complex_3d = paddle.to_tensor( (np.random.randn(2, 4, 8) + 1j * np.random.randn(2, 4, 8)).astype( 'complex64' ) ) def test_fftn(self): """测试N维FFT / Test N-dimensional FFT""" result = paddle.fft.fftn(self.x_complex_3d) self.assertEqual(result.shape, self.x_complex_3d.shape) def test_ifftn(self): """测试N维IFFT / Test N-dimensional IFFT""" fft_result = paddle.fft.fftn(self.x_complex_3d) recovered = paddle.fft.ifftn(fft_result) self.assertEqual(recovered.shape, self.x_complex_3d.shape) def test_rfftn(self): """测试N维实数FFT / Test N-dimensional real FFT""" result = paddle.fft.rfftn(self.x_real_3d) expected_last = self.x_real_3d.shape[-1] // 2 + 1 self.assertEqual(result.shape[-1], expected_last) def test_irfftn(self): """测试N维实数IFFT / Test N-dimensional real IFFT""" rfft_result = paddle.fft.rfftn(self.x_real_3d) recovered = paddle.fft.irfftn(rfft_result) self.assertEqual(recovered.shape[-1], self.x_real_3d.shape[-1]) def test_fftn_with_axes(self): """测试带axes参数的N维FFT / Test N-dimensional FFT with axes parameter""" result = paddle.fft.fftn(self.x_complex_3d, axes=[1, 2]) self.assertEqual(result.shape, self.x_complex_3d.shape) class TestFFTFreq(unittest.TestCase): """测试FFT频率函数 / Test FFT frequency functions""" def test_fftfreq(self): """测试FFT频率 / Test FFT frequency""" result = paddle.fft.fftfreq(8) self.assertEqual(result.shape[0], 8) def test_fftfreq_with_d(self): """测试带采样间隔的FFT频率 / Test FFT frequency with sample spacing""" result = paddle.fft.fftfreq(8, d=0.5) self.assertEqual(result.shape[0], 8) def test_rfftfreq(self): """测试实数FFT频率 / Test real FFT frequency""" result = paddle.fft.rfftfreq(8) self.assertEqual(result.shape[0], 8 // 2 + 1) def test_rfftfreq_with_d(self): """测试带采样间隔的实数FFT频率 / Test real FFT frequency with sample spacing""" result = paddle.fft.rfftfreq(8, d=0.5) self.assertEqual(result.shape[0], 8 // 2 + 1) def test_fftshift(self): """测试FFT频率移位 / Test FFT frequency shift""" x = paddle.to_tensor(np.fft.fftfreq(8).astype('float32')) result = paddle.fft.fftshift(x) self.assertEqual(result.shape, x.shape) def test_ifftshift(self): """测试IFFT频率移位 / Test IFFT frequency shift""" x = paddle.to_tensor( np.fft.fftshift(np.fft.fftfreq(8)).astype('float32') ) result = paddle.fft.ifftshift(x) self.assertEqual(result.shape, x.shape) def test_hfft(self): """测试Hermitian FFT / Test Hermitian FFT""" x = paddle.to_tensor( (np.random.randn(5) + 1j * np.random.randn(5)).astype('complex64') ) result = paddle.fft.hfft(x) # hfft output real, length 2*(n-1) self.assertTrue(result.dtype in [paddle.float32, paddle.float64]) def test_ihfft(self): """测试Hermitian IFFT / Test Hermitian IFFT""" x = paddle.to_tensor(np.random.randn(8).astype('float32')) result = paddle.fft.ihfft(x) self.assertTrue(result.dtype in [paddle.complex64, paddle.complex128]) if __name__ == '__main__': unittest.main()