Support cross compiling for aarch64 (#52)

@@ -9,3 +9,4 @@ __pycache__
 dist/
 sherpa_onnx.egg-info/
 .DS_Store
+build-aarch64-linux-gnu

@@ -50,6 +50,12 @@ message(STATUS "SHERPA_ONNX_ENABLE_PYTHON ${SHERPA_ONNX_ENABLE_PYTHON}")
 set(CMAKE_CXX_STANDARD 14 CACHE STRING "The C++ version to be used.")
 set(CMAKE_CXX_EXTENSIONS OFF)
 
+include(CheckIncludeFileCXX)
+check_include_file_cxx(alsa/asoundlib.h SHERPA_ONNX_HAS_ALSA)
+if(SHERPA_ONNX_HAS_ALSA)
+  add_definitions(-DSHERPA_ONNX_ENABLE_ALSA=1)
+endif()
+
 list(APPEND CMAKE_MODULE_PATH ${CMAKE_SOURCE_DIR}/cmake/Modules)
 list(APPEND CMAKE_MODULE_PATH ${CMAKE_SOURCE_DIR}/cmake)
 

+#!/usr/bin/env bash
+
+if ! command -v aarch64-linux-gnu-gcc  &> /dev/null; then
+  echo "Please install a toolchain for cross-compiling."
+  echo "You can refer to: "
+  echo "  https://k2-fsa.github.io/sherpa/onnx/install/aarch64-embedded-linux.html"
+  echo "for help."
+  exit 1
+fi
+
+set -ex
+
+dir=build-aarch64-linux-gnu
+mkdir -p $dir
+cd $dir
+
+if [ ! -f alsa-lib/src/.libs/libasound.so ]; then
+  echo "Start to cross-compile alsa-lib"
+  if [ ! -d alsa-lib ]; then
+    git clone --depth 1 https://github.com/alsa-project/alsa-lib
+  fi
+  # If it shows:
+  #  ./gitcompile: line 79: libtoolize: command not found
+  # Please use:
+  #  sudo apt-get install libtool m4 automake
+  #
+  pushd alsa-lib
+  CC=aarch64-linux-gnu-gcc ./gitcompile --host=aarch64-linux-gnu
+  popd
+  echo "Finish cross-compiling alsa-lib"
+fi
+
+export CPLUS_INCLUDE_PATH=$PWD/alsa-lib/include:$CPLUS_INCLUDE_PATH
+export SHERPA_ONNX_ALSA_LIB_DIR=$PWD/alsa-lib/src/.libs
+
+cmake \
+  -DCMAKE_INSTALL_PREFIX=./install \
+  -DCMAKE_BUILD_TYPE=Release \
+  -DBUILD_SHARED_LIBS=OFF \
+  -DSHERPA_ONNX_ENABLE_TESTS=OFF \
+  -DSHERPA_ONNX_ENABLE_PYTHON=OFF \
+  -DCMAKE_TOOLCHAIN_FILE=../toolchains/aarch64-linux-gnu.toolchain.cmake \
+  ..
+
+make VERBOSE=1 -j4
+make install/strip
+
+# Enable it if only needed
+# cp -v $SHERPA_ONNX_ALSA_LIB_DIR/libasound.so* ./install/lib/

 function(download_onnxruntime)
   include(FetchContent)
 
-  if(UNIX AND NOT APPLE)
-    # If you don't have access to the Internet,
-    # please pre-download onnxruntime
-    set(possible_file_locations
-      $ENV{HOME}/Downloads/onnxruntime-linux-x64-1.14.0.tgz
-      ${PROJECT_SOURCE_DIR}/onnxruntime-linux-x64-1.14.0.tgz
-      ${PROJECT_BINARY_DIR}/onnxruntime-linux-x64-1.14.0.tgz
-      /tmp/onnxruntime-linux-x64-1.14.0.tgz
-      /star-fj/fangjun/download/github/onnxruntime-linux-x64-1.14.0.tgz
-    )
-
-    set(onnxruntime_URL  "https://github.com/microsoft/onnxruntime/releases/download/v1.14.0/onnxruntime-linux-x64-1.14.0.tgz")
-    set(onnxruntime_HASH "SHA256=92bf534e5fa5820c8dffe9de2850f84ed2a1c063e47c659ce09e8c7938aa2090")
-    # After downloading, it contains:
-    #  ./lib/libonnxruntime.so.1.14.0
-    #  ./lib/libonnxruntime.so, which is a symlink to lib/libonnxruntime.so.1.14.0
-    #
-    # ./include
-    #    It contains all the needed header files
+  if(CMAKE_SYSTEM_NAME STREQUAL Linux)
+    if(CMAKE_SYSTEM_PROCESSOR STREQUAL aarch64)
+      # For embedded systems
+      set(possible_file_locations
+        $ENV{HOME}/Downloads/onnxruntime-linux-aarch64-1.14.0.tgz
+        ${PROJECT_SOURCE_DIR}/onnxruntime-linux-aarch64-1.14.0.tgz
+        ${PROJECT_BINARY_DIR}/onnxruntime-linux-aarch64-1.14.0.tgz
+        /tmp/onnxruntime-linux-aarch64-1.14.0.tgz
+        /star-fj/fangjun/download/github/onnxruntime-linux-aarch64-1.14.0.tgz
+      )
+      set(onnxruntime_URL  "https://github.com/microsoft/onnxruntime/releases/download/v1.14.0/onnxruntime-linux-aarch64-1.14.0.tgz")
+      set(onnxruntime_HASH "SHA256=9384d2e6e29fed693a4630303902392eead0c41bee5705ccac6d6d34a3d5db86")
+
+    else()
+      # If you don't have access to the Internet,
+      # please pre-download onnxruntime
+      set(possible_file_locations
+        $ENV{HOME}/Downloads/onnxruntime-linux-x64-1.14.0.tgz
+        ${PROJECT_SOURCE_DIR}/onnxruntime-linux-x64-1.14.0.tgz
+        ${PROJECT_BINARY_DIR}/onnxruntime-linux-x64-1.14.0.tgz
+        /tmp/onnxruntime-linux-x64-1.14.0.tgz
+        /star-fj/fangjun/download/github/onnxruntime-linux-x64-1.14.0.tgz
+      )
+
+      set(onnxruntime_URL  "https://github.com/microsoft/onnxruntime/releases/download/v1.14.0/onnxruntime-linux-x64-1.14.0.tgz")
+      set(onnxruntime_HASH "SHA256=92bf534e5fa5820c8dffe9de2850f84ed2a1c063e47c659ce09e8c7938aa2090")
+      # After downloading, it contains:
+      #  ./lib/libonnxruntime.so.1.14.0
+      #  ./lib/libonnxruntime.so, which is a symlink to lib/libonnxruntime.so.1.14.0
+      #
+      # ./include
+      #    It contains all the needed header files
+    endif()
   elseif(APPLE)
     # If you don't have access to the Internet,
     # please pre-download onnxruntime

@@ -11,6 +11,7 @@ add_library(sherpa-onnx-core
   online-transducer-model.cc
   online-zipformer-transducer-model.cc
   onnx-utils.cc
+  resample.cc
   symbol-table.cc
   text-utils.cc
   unbind.cc
@@ -32,6 +33,18 @@ endif()
 install(TARGETS sherpa-onnx-core DESTINATION lib)
 install(TARGETS sherpa-onnx DESTINATION bin)
 
+if(SHERPA_ONNX_HAS_ALSA)
+  add_executable(sherpa-onnx-alsa sherpa-onnx-alsa.cc alsa.cc)
+  target_link_libraries(sherpa-onnx-alsa PRIVATE sherpa-onnx-core)
+
+  if(DEFINED ENV{SHERPA_ONNX_ALSA_LIB_DIR})
+    target_link_libraries(sherpa-onnx-alsa PRIVATE -L$ENV{SHERPA_ONNX_ALSA_LIB_DIR} -lasound)
+  else()
+    target_link_libraries(sherpa-onnx-alsa PRIVATE asound)
+  endif()
+  install(TARGETS sherpa-onnx-alsa DESTINATION bin)
+endif()
+
 if(SHERPA_ONNX_ENABLE_TESTS)
   set(sherpa_onnx_test_srcs
     cat-test.cc

+// sherpa-onnx/csrc/sherpa-alsa.cc
+//
+// Copyright (c)  2022-2023  Xiaomi Corporation
+
+#ifdef SHERPA_ONNX_ENABLE_ALSA
+
+#include "sherpa-onnx/csrc/alsa.h"
+
+#include <algorithm>
+
+#include "alsa/asoundlib.h"
+
+namespace sherpa_onnx {
+
+void ToFloat(const std::vector<int16_t> &in, int32_t num_channels,
+             std::vector<float> *out) {
+  out->resize(in.size() / num_channels);
+
+  int32_t n = in.size();
+  for (int32_t i = 0, k = 0; i < n; i += num_channels, ++k) {
+    (*out)[k] = in[i] / 32768.;
+  }
+}
+
+Alsa::Alsa(const char *device_name) {
+  const char *kDeviceHelp = R"(
+Please use the command:
+
+  arecord -l
+
+to list all available devices. For instance, if the output is:
+
+**** List of CAPTURE Hardware Devices ****
+card 3: UACDemoV10 [UACDemoV1.0], device 0: USB Audio [USB Audio]
+  Subdevices: 1/1
+  Subdevice #0: subdevice #0
+
+and if you want to select card 3 and the device 0 on that card, please use:
+
+  hw:3,0
+
+  )";
+
+  int32_t err =
+      snd_pcm_open(&capture_handle_, device_name, SND_PCM_STREAM_CAPTURE, 0);
+  if (err) {
+    fprintf(stderr, "Unable to open: %s. %s\n", device_name, snd_strerror(err));
+    fprintf(stderr, "%s\n", kDeviceHelp);
+    exit(-1);
+  }
+
+  snd_pcm_hw_params_t *hw_params;
+  snd_pcm_hw_params_alloca(&hw_params);
+
+  err = snd_pcm_hw_params_any(capture_handle_, hw_params);
+  if (err) {
+    fprintf(stderr, "Failed to initialize hw_params: %s\n", snd_strerror(err));
+    exit(-1);
+  }
+
+  err = snd_pcm_hw_params_set_access(capture_handle_, hw_params,
+                                     SND_PCM_ACCESS_RW_INTERLEAVED);
+  if (err) {
+    fprintf(stderr, "Failed to set access type: %s\n", snd_strerror(err));
+    exit(-1);
+  }
+
+  err = snd_pcm_hw_params_set_format(capture_handle_, hw_params,
+                                     SND_PCM_FORMAT_S16_LE);
+  if (err) {
+    fprintf(stderr, "Failed to set format: %s\n", snd_strerror(err));
+    exit(-1);
+  }
+
+  // mono
+  err = snd_pcm_hw_params_set_channels(capture_handle_, hw_params, 1);
+  if (err) {
+    fprintf(stderr, "Failed to set number of channels to 1. %s\n",
+            snd_strerror(err));
+
+    err = snd_pcm_hw_params_set_channels(capture_handle_, hw_params, 2);
+    if (err) {
+      fprintf(stderr, "Failed to set number of channels to 2. %s\n",
+              snd_strerror(err));
+
+      exit(-1);
+    }
+    actual_channel_count_ = 2;
+    fprintf(stderr,
+            "Channel count is set to 2. Will use only 1 channel of it.\n");
+  }
+
+  uint32_t actual_sample_rate = expected_sample_rate_;
+
+  int32_t dir = 0;
+  err = snd_pcm_hw_params_set_rate_near(capture_handle_, hw_params,
+                                        &actual_sample_rate, &dir);
+  if (err) {
+    fprintf(stderr, "Failed to set sample rate to, %d: %s\n",
+            expected_sample_rate_, snd_strerror(err));
+    exit(-1);
+  }
+  actual_sample_rate_ = actual_sample_rate;
+
+  if (actual_sample_rate_ != expected_sample_rate_) {
+    fprintf(stderr, "Failed to set sample rate to %d\n", expected_sample_rate_);
+    fprintf(stderr, "Current sample rate is %d\n", actual_sample_rate_);
+    fprintf(stderr,
+            "Creating a resampler:\n"
+            "   in_sample_rate: %d\n"
+            "   output_sample_rate: %d\n",
+            actual_sample_rate_, expected_sample_rate_);
+
+    float min_freq = std::min(actual_sample_rate_, expected_sample_rate_);
+    float lowpass_cutoff = 0.99 * 0.5 * min_freq;
+
+    int32_t lowpass_filter_width = 6;
+    resampler_ = std::make_unique<LinearResample>(
+        actual_sample_rate_, expected_sample_rate_, lowpass_cutoff,
+        lowpass_filter_width);
+  } else {
+    fprintf(stderr, "Current sample rate: %d\n", actual_sample_rate_);
+  }
+
+  err = snd_pcm_hw_params(capture_handle_, hw_params);
+  if (err) {
+    fprintf(stderr, "Failed to set hw params: %s\n", snd_strerror(err));
+    exit(-1);
+  }
+
+  err = snd_pcm_prepare(capture_handle_);
+  if (err) {
+    fprintf(stderr, "Failed to prepare for recording: %s\n", snd_strerror(err));
+    exit(-1);
+  }
+
+  fprintf(stderr, "Recording started!\n");
+}
+
+Alsa::~Alsa() { snd_pcm_close(capture_handle_); }
+
+const std::vector<float> &Alsa::Read(int32_t num_samples) {
+  samples_.resize(num_samples * actual_channel_count_);
+
+  // count is in frames. Each frame contains actual_channel_count_ samples
+  int32_t count = snd_pcm_readi(capture_handle_, samples_.data(), num_samples);
+
+  samples_.resize(count * actual_channel_count_);
+
+  ToFloat(samples_, actual_channel_count_, &samples1_);
+
+  if (!resampler_) {
+    return samples1_;
+  }
+
+  resampler_->Resample(samples1_.data(), samples_.size(), false, &samples2_);
+  return samples2_;
+}
+
+}  // namespace sherpa_onnx
+
+#endif

+// sherpa-onnx/csrc/sherpa-alsa.h
+//
+// Copyright (c)  2022-2023  Xiaomi Corporation
+
+#ifndef SHERPA_ONNX_CSRC_ALSA_H_
+#define SHERPA_ONNX_CSRC_ALSA_H_
+
+#include <memory>
+#include <vector>
+
+#include "alsa/asoundlib.h"
+#include "sherpa-onnx/csrc/resample.h"
+
+namespace sherpa_onnx {
+
+class Alsa {
+ public:
+  explicit Alsa(const char *device_name);
+  ~Alsa();
+
+  // This is a blocking read.
+  //
+  // @param num_samples  Number of samples to read.
+  //
+  // The returned value is valid until the next call to Read().
+  const std::vector<float> &Read(int32_t num_samples);
+
+  int32_t GetExpectedSampleRate() const { return expected_sample_rate_; }
+  int32_t GetActualSampleRate() const { return actual_sample_rate_; }
+
+ private:
+  snd_pcm_t *capture_handle_;
+  int32_t expected_sample_rate_ = 16000;
+  int32_t actual_sample_rate_;
+
+  int32_t actual_channel_count_ = 1;
+
+  std::unique_ptr<LinearResample> resampler_;
+  std::vector<int16_t> samples_;  // directly from the microphone
+  std::vector<float> samples1_;   // normalized version of samples_
+  std::vector<float> samples2_;   // possibly resampled from samples1_
+};
+
+}  // namespace sherpa_onnx
+
+#endif  // SHERPA_ONNX_CSRC_ALSA_H_

+// sherpa-onnx/csrc/display.h
+//
+// Copyright (c)  2022-2023  Xiaomi Corporation
+
+#ifndef SHERPA_ONNX_CSRC_DISPLAY_H_
+#define SHERPA_ONNX_CSRC_DISPLAY_H_
+#include <stdio.h>
+
+#include <string>
+
+namespace sherpa_onnx {
+
+class Display {
+ public:
+  void Print(int32_t segment_id, const std::string &s) {
+#ifdef _MSC_VER
+    fprintf(stderr, "%d:%s\n", segment_id, s.c_str());
+    return;
+#endif
+    if (last_segment_ == segment_id) {
+      Clear();
+    } else {
+      if (last_segment_ != -1) {
+        fprintf(stderr, "\n\r");
+      }
+      last_segment_ = segment_id;
+      num_previous_lines_ = 0;
+    }
+
+    fprintf(stderr, "\r%d:", segment_id);
+
+    int32_t i = 0;
+    for (size_t n = 0; n < s.size();) {
+      if (s[n] > 0 && s[n] < 0x7f) {
+        fprintf(stderr, "%c", s[n]);
+        ++n;
+      } else {
+        // Each Chinese character occupies 3 bytes for UTF-8 encoding.
+        std::string tmp(s.begin() + n, s.begin() + n + 3);
+        fprintf(stderr, "%s", tmp.data());
+        n += 3;
+      }
+
+      ++i;
+      if (i >= max_word_per_line_ && n + 1 < s.size() &&
+          (s[n] == ' ' || s[n] < 0)) {
+        fprintf(stderr, "\n\r ");
+        ++num_previous_lines_;
+        i = 0;
+      }
+    }
+  }
+
+ private:
+  // Clear the output for the current segment
+  void Clear() {
+    ClearCurrentLine();
+    while (num_previous_lines_ > 0) {
+      GoUpOneLine();
+      ClearCurrentLine();
+      --num_previous_lines_;
+    }
+  }
+
+  // Clear the current line
+  void ClearCurrentLine() const { fprintf(stderr, "\33[2K\r"); }
+
+  // Move the cursor to the previous line
+  void GoUpOneLine() const { fprintf(stderr, "\033[1A\r"); }
+
+ private:
+  int32_t max_word_per_line_ = 60;
+  int32_t num_previous_lines_ = 0;
+  int32_t last_segment_ = -1;
+};
+
+}  // namespace sherpa_onnx
+
+#endif  // SHERPA_ONNX_CSRC_DISPLAY_H_

+/**
+ * Copyright     2013  Pegah Ghahremani
+ *               2014  IMSL, PKU-HKUST (author: Wei Shi)
+ *               2014  Yanqing Sun, Junjie Wang
+ *               2014  Johns Hopkins University (author: Daniel Povey)
+ * Copyright     2023  Xiaomi Corporation (authors: Fangjun Kuang)
+ *
+ * See LICENSE for clarification regarding multiple authors
+ *
+ * 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.
+ */
+// this file is copied and modified from
+// kaldi/src/feat/resample.cc
+
+#include "sherpa-onnx/csrc/resample.h"
+
+#include <assert.h>
+#include <math.h>
+#include <stdio.h>
+
+#include <cstdlib>
+#include <type_traits>
+
+#ifndef M_2PI
+#define M_2PI 6.283185307179586476925286766559005
+#endif
+
+#ifndef M_PI
+#define M_PI 3.1415926535897932384626433832795
+#endif
+
+namespace sherpa_onnx {
+
+template <class I>
+I Gcd(I m, I n) {
+  // this function is copied from kaldi/src/base/kaldi-math.h
+  if (m == 0 || n == 0) {
+    if (m == 0 && n == 0) {  // gcd not defined, as all integers are divisors.
+      fprintf(stderr, "Undefined GCD since m = 0, n = 0.");
+      exit(-1);
+    }
+    return (m == 0 ? (n > 0 ? n : -n) : (m > 0 ? m : -m));
+    // return absolute value of whichever is nonzero
+  }
+  // could use compile-time assertion
+  // but involves messing with complex template stuff.
+  static_assert(std::is_integral<I>::value, "");
+  while (1) {
+    m %= n;
+    if (m == 0) return (n > 0 ? n : -n);
+    n %= m;
+    if (n == 0) return (m > 0 ? m : -m);
+  }
+}
+
+/// Returns the least common multiple of two integers.  Will
+/// crash unless the inputs are positive.
+template <class I>
+I Lcm(I m, I n) {
+  // This function is copied from kaldi/src/base/kaldi-math.h
+  assert(m > 0 && n > 0);
+  I gcd = Gcd(m, n);
+  return gcd * (m / gcd) * (n / gcd);
+}
+
+static float DotProduct(const float *a, const float *b, int32_t n) {
+  float sum = 0;
+  for (int32_t i = 0; i != n; ++i) {
+    sum += a[i] * b[i];
+  }
+  return sum;
+}
+
+LinearResample::LinearResample(int32_t samp_rate_in_hz,
+                               int32_t samp_rate_out_hz, float filter_cutoff_hz,
+                               int32_t num_zeros)
+    : samp_rate_in_(samp_rate_in_hz),
+      samp_rate_out_(samp_rate_out_hz),
+      filter_cutoff_(filter_cutoff_hz),
+      num_zeros_(num_zeros) {
+  assert(samp_rate_in_hz > 0.0 && samp_rate_out_hz > 0.0 &&
+         filter_cutoff_hz > 0.0 && filter_cutoff_hz * 2 <= samp_rate_in_hz &&
+         filter_cutoff_hz * 2 <= samp_rate_out_hz && num_zeros > 0);
+
+  // base_freq is the frequency of the repeating unit, which is the gcd
+  // of the input frequencies.
+  int32_t base_freq = Gcd(samp_rate_in_, samp_rate_out_);
+  input_samples_in_unit_ = samp_rate_in_ / base_freq;
+  output_samples_in_unit_ = samp_rate_out_ / base_freq;
+
+  SetIndexesAndWeights();
+  Reset();
+}
+
+void LinearResample::SetIndexesAndWeights() {
+  first_index_.resize(output_samples_in_unit_);
+  weights_.resize(output_samples_in_unit_);
+
+  double window_width = num_zeros_ / (2.0 * filter_cutoff_);
+
+  for (int32_t i = 0; i < output_samples_in_unit_; i++) {
+    double output_t = i / static_cast<double>(samp_rate_out_);
+    double min_t = output_t - window_width, max_t = output_t + window_width;
+    // we do ceil on the min and floor on the max, because if we did it
+    // the other way around we would unnecessarily include indexes just
+    // outside the window, with zero coefficients.  It's possible
+    // if the arguments to the ceil and floor expressions are integers
+    // (e.g. if filter_cutoff_ has an exact ratio with the sample rates),
+    // that we unnecessarily include something with a zero coefficient,
+    // but this is only a slight efficiency issue.
+    int32_t min_input_index = ceil(min_t * samp_rate_in_),
+            max_input_index = floor(max_t * samp_rate_in_),
+            num_indices = max_input_index - min_input_index + 1;
+    first_index_[i] = min_input_index;
+    weights_[i].resize(num_indices);
+    for (int32_t j = 0; j < num_indices; j++) {
+      int32_t input_index = min_input_index + j;
+      double input_t = input_index / static_cast<double>(samp_rate_in_),
+             delta_t = input_t - output_t;
+      // sign of delta_t doesn't matter.
+      weights_[i][j] = FilterFunc(delta_t) / samp_rate_in_;
+    }
+  }
+}
+
+/** Here, t is a time in seconds representing an offset from
+    the center of the windowed filter function, and FilterFunction(t)
+    returns the windowed filter function, described
+    in the header as h(t) = f(t)g(t), evaluated at t.
+*/
+float LinearResample::FilterFunc(float t) const {
+  float window,  // raised-cosine (Hanning) window of width
+                 // num_zeros_/2*filter_cutoff_
+      filter;    // sinc filter function
+  if (fabs(t) < num_zeros_ / (2.0 * filter_cutoff_))
+    window = 0.5 * (1 + cos(M_2PI * filter_cutoff_ / num_zeros_ * t));
+  else
+    window = 0.0;  // outside support of window function
+  if (t != 0)
+    filter = sin(M_2PI * filter_cutoff_ * t) / (M_PI * t);
+  else
+    filter = 2 * filter_cutoff_;  // limit of the function at t = 0
+  return filter * window;
+}
+
+void LinearResample::Reset() {
+  input_sample_offset_ = 0;
+  output_sample_offset_ = 0;
+  input_remainder_.resize(0);
+}
+
+void LinearResample::Resample(const float *input, int32_t input_dim, bool flush,
+                              std::vector<float> *output) {
+  int64_t tot_input_samp = input_sample_offset_ + input_dim,
+          tot_output_samp = GetNumOutputSamples(tot_input_samp, flush);
+
+  assert(tot_output_samp >= output_sample_offset_);
+
+  output->resize(tot_output_samp - output_sample_offset_);
+
+  // samp_out is the index into the total output signal, not just the part
+  // of it we are producing here.
+  for (int64_t samp_out = output_sample_offset_; samp_out < tot_output_samp;
+       samp_out++) {
+    int64_t first_samp_in;
+    int32_t samp_out_wrapped;
+    GetIndexes(samp_out, &first_samp_in, &samp_out_wrapped);
+    const std::vector<float> &weights = weights_[samp_out_wrapped];
+    // first_input_index is the first index into "input" that we have a weight
+    // for.
+    int32_t first_input_index =
+        static_cast<int32_t>(first_samp_in - input_sample_offset_);
+    float this_output;
+    if (first_input_index >= 0 &&
+        first_input_index + static_cast<int32_t>(weights.size()) <= input_dim) {
+      this_output =
+          DotProduct(input + first_input_index, weights.data(), weights.size());
+    } else {  // Handle edge cases.
+      this_output = 0.0;
+      for (int32_t i = 0; i < static_cast<int32_t>(weights.size()); i++) {
+        float weight = weights[i];
+        int32_t input_index = first_input_index + i;
+        if (input_index < 0 &&
+            static_cast<int32_t>(input_remainder_.size()) + input_index >= 0) {
+          this_output +=
+              weight * input_remainder_[input_remainder_.size() + input_index];
+        } else if (input_index >= 0 && input_index < input_dim) {
+          this_output += weight * input[input_index];
+        } else if (input_index >= input_dim) {
+          // We're past the end of the input and are adding zero; should only
+          // happen if the user specified flush == true, or else we would not
+          // be trying to output this sample.
+          assert(flush);
+        }
+      }
+    }
+    int32_t output_index =
+        static_cast<int32_t>(samp_out - output_sample_offset_);
+    (*output)[output_index] = this_output;
+  }
+
+  if (flush) {
+    Reset();  // Reset the internal state.
+  } else {
+    SetRemainder(input, input_dim);
+    input_sample_offset_ = tot_input_samp;
+    output_sample_offset_ = tot_output_samp;
+  }
+}
+
+int64_t LinearResample::GetNumOutputSamples(int64_t input_num_samp,
+                                            bool flush) const {
+  // For exact computation, we measure time in "ticks" of 1.0 / tick_freq,
+  // where tick_freq is the least common multiple of samp_rate_in_ and
+  // samp_rate_out_.
+  int32_t tick_freq = Lcm(samp_rate_in_, samp_rate_out_);
+  int32_t ticks_per_input_period = tick_freq / samp_rate_in_;
+
+  // work out the number of ticks in the time interval
+  // [ 0, input_num_samp/samp_rate_in_ ).
+  int64_t interval_length_in_ticks = input_num_samp * ticks_per_input_period;
+  if (!flush) {
+    float window_width = num_zeros_ / (2.0 * filter_cutoff_);
+    // To count the window-width in ticks we take the floor.  This
+    // is because since we're looking for the largest integer num-out-samp
+    // that fits in the interval, which is open on the right, a reduction
+    // in interval length of less than a tick will never make a difference.
+    // For example, the largest integer in the interval [ 0, 2 ) and the
+    // largest integer in the interval [ 0, 2 - 0.9 ) are the same (both one).
+    // So when we're subtracting the window-width we can ignore the fractional
+    // part.
+    int32_t window_width_ticks = floor(window_width * tick_freq);
+    // The time-period of the output that we can sample gets reduced
+    // by the window-width (which is actually the distance from the
+    // center to the edge of the windowing function) if we're not
+    // "flushing the output".
+    interval_length_in_ticks -= window_width_ticks;
+  }
+  if (interval_length_in_ticks <= 0) return 0;
+
+  int32_t ticks_per_output_period = tick_freq / samp_rate_out_;
+  // Get the last output-sample in the closed interval, i.e. replacing [ ) with
+  // [ ].  Note: integer division rounds down.  See
+  // http://en.wikipedia.org/wiki/Interval_(mathematics) for an explanation of
+  // the notation.
+  int64_t last_output_samp = interval_length_in_ticks / ticks_per_output_period;
+  // We need the last output-sample in the open interval, so if it takes us to
+  // the end of the interval exactly, subtract one.
+  if (last_output_samp * ticks_per_output_period == interval_length_in_ticks)
+    last_output_samp--;
+
+  // First output-sample index is zero, so the number of output samples
+  // is the last output-sample plus one.
+  int64_t num_output_samp = last_output_samp + 1;
+  return num_output_samp;
+}
+
+// inline
+void LinearResample::GetIndexes(int64_t samp_out, int64_t *first_samp_in,
+                                int32_t *samp_out_wrapped) const {
+  // A unit is the smallest nonzero amount of time that is an exact
+  // multiple of the input and output sample periods.  The unit index
+  // is the answer to "which numbered unit we are in".
+  int64_t unit_index = samp_out / output_samples_in_unit_;
+  // samp_out_wrapped is equal to samp_out % output_samples_in_unit_
+  *samp_out_wrapped =
+      static_cast<int32_t>(samp_out - unit_index * output_samples_in_unit_);
+  *first_samp_in =
+      first_index_[*samp_out_wrapped] + unit_index * input_samples_in_unit_;
+}
+
+void LinearResample::SetRemainder(const float *input, int32_t input_dim) {
+  std::vector<float> old_remainder(input_remainder_);
+  // max_remainder_needed is the width of the filter from side to side,
+  // measured in input samples.  you might think it should be half that,
+  // but you have to consider that you might be wanting to output samples
+  // that are "in the past" relative to the beginning of the latest
+  // input... anyway, storing more remainder than needed is not harmful.
+  int32_t max_remainder_needed =
+      ceil(samp_rate_in_ * num_zeros_ / filter_cutoff_);
+  input_remainder_.resize(max_remainder_needed);
+  for (int32_t index = -static_cast<int32_t>(input_remainder_.size());
+       index < 0; index++) {
+    // we interpret "index" as an offset from the end of "input" and
+    // from the end of input_remainder_.
+    int32_t input_index = index + input_dim;
+    if (input_index >= 0) {
+      input_remainder_[index + static_cast<int32_t>(input_remainder_.size())] =
+          input[input_index];
+    } else if (input_index + static_cast<int32_t>(old_remainder.size()) >= 0) {
+      input_remainder_[index + static_cast<int32_t>(input_remainder_.size())] =
+          old_remainder[input_index +
+                        static_cast<int32_t>(old_remainder.size())];
+      // else leave it at zero.
+    }
+  }
+}
+
+}  // namespace sherpa_onnx

+/**
+ * Copyright     2013  Pegah Ghahremani
+ *               2014  IMSL, PKU-HKUST (author: Wei Shi)
+ *               2014  Yanqing Sun, Junjie Wang
+ *               2014  Johns Hopkins University (author: Daniel Povey)
+ * Copyright     2023  Xiaomi Corporation (authors: Fangjun Kuang)
+ *
+ * See LICENSE for clarification regarding multiple authors
+ *
+ * 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.
+ */
+// this file is copied and modified from
+// kaldi/src/feat/resample.h
+#ifndef SHERPA_ONNX_CSRC_RESAMPLE_H_
+#define SHERPA_ONNX_CSRC_RESAMPLE_H_
+
+#include <cstdint>
+#include <vector>
+
+namespace sherpa_onnx {
+
+/*
+   We require that the input and output sampling rate be specified as
+   integers, as this is an easy way to specify that their ratio be rational.
+*/
+
+class LinearResample {
+ public:
+  /// Constructor.  We make the input and output sample rates integers, because
+  /// we are going to need to find a common divisor.  This should just remind
+  /// you that they need to be integers.  The filter cutoff needs to be less
+  /// than samp_rate_in_hz/2 and less than samp_rate_out_hz/2.  num_zeros
+  /// controls the sharpness of the filter, more == sharper but less efficient.
+  /// We suggest around 4 to 10 for normal use.
+  LinearResample(int32_t samp_rate_in_hz, int32_t samp_rate_out_hz,
+                 float filter_cutoff_hz, int32_t num_zeros);
+
+  /// Calling the function Reset() resets the state of the object prior to
+  /// processing a new signal; it is only necessary if you have called
+  /// Resample(x, x_size, false, y) for some signal, leading to a remainder of
+  /// the signal being called, but then abandon processing the signal before
+  /// calling Resample(x, x_size, true, y) for the last piece.  Call it
+  /// unnecessarily between signals will not do any harm.
+  void Reset();
+
+  /// This function does the resampling.  If you call it with flush == true and
+  /// you have never called it with flush == false, it just resamples the input
+  /// signal (it resizes the output to a suitable number of samples).
+  ///
+  /// You can also use this function to process a signal a piece at a time.
+  /// suppose you break it into piece1, piece2, ... pieceN.  You can call
+  /// \code{.cc}
+  /// Resample(piece1, piece1_size, false, &output1);
+  /// Resample(piece2, piece2_size, false, &output2);
+  /// Resample(piece3, piece3_size, true, &output3);
+  /// \endcode
+  /// If you call it with flush == false, it won't output the last few samples
+  /// but will remember them, so that if you later give it a second piece of
+  /// the input signal it can process it correctly.
+  /// If your most recent call to the object was with flush == false, it will
+  /// have internal state; you can remove this by calling Reset().
+  /// Empty input is acceptable.
+  void Resample(const float *input, int32_t input_dim, bool flush,
+                std::vector<float> *output);
+
+  //// Return the input and output sampling rates (for checks, for example)
+  int32_t GetInputSamplingRate() const { return samp_rate_in_; }
+  int32_t GetOutputSamplingRate() const { return samp_rate_out_; }
+
+ private:
+  void SetIndexesAndWeights();
+
+  float FilterFunc(float) const;
+
+  /// This function outputs the number of output samples we will output
+  /// for a signal with "input_num_samp" input samples.  If flush == true,
+  /// we return the largest n such that
+  /// (n/samp_rate_out_) is in the interval [ 0, input_num_samp/samp_rate_in_ ),
+  /// and note that the interval is half-open.  If flush == false,
+  /// define window_width as num_zeros / (2.0 * filter_cutoff_);
+  /// we return the largest n such that (n/samp_rate_out_) is in the interval
+  /// [ 0, input_num_samp/samp_rate_in_ - window_width ).
+  int64_t GetNumOutputSamples(int64_t input_num_samp, bool flush) const;
+
+  /// Given an output-sample index, this function outputs to *first_samp_in the
+  /// first input-sample index that we have a weight on (may be negative),
+  /// and to *samp_out_wrapped the index into weights_ where we can get the
+  /// corresponding weights on the input.
+  inline void GetIndexes(int64_t samp_out, int64_t *first_samp_in,
+                         int32_t *samp_out_wrapped) const;
+
+  void SetRemainder(const float *input, int32_t input_dim);
+
+ private:
+  // The following variables are provided by the user.
+  int32_t samp_rate_in_;
+  int32_t samp_rate_out_;
+  float filter_cutoff_;
+  int32_t num_zeros_;
+
+  int32_t input_samples_in_unit_;  ///< The number of input samples in the
+                                   ///< smallest repeating unit: num_samp_in_ =
+                                   ///< samp_rate_in_hz / Gcd(samp_rate_in_hz,
+                                   ///< samp_rate_out_hz)
+
+  int32_t output_samples_in_unit_;  ///< The number of output samples in the
+                                    ///< smallest repeating unit: num_samp_out_
+                                    ///< = samp_rate_out_hz /
+                                    ///< Gcd(samp_rate_in_hz, samp_rate_out_hz)
+
+  /// The first input-sample index that we sum over, for this output-sample
+  /// index.  May be negative; any truncation at the beginning is handled
+  /// separately.  This is just for the first few output samples, but we can
+  /// extrapolate the correct input-sample index for arbitrary output samples.
+  std::vector<int32_t> first_index_;
+
+  /// Weights on the input samples, for this output-sample index.
+  std::vector<std::vector<float>> weights_;
+
+  // the following variables keep track of where we are in a particular signal,
+  // if it is being provided over multiple calls to Resample().
+
+  int64_t input_sample_offset_;   ///< The number of input samples we have
+                                  ///< already received for this signal
+                                  ///< (including anything in remainder_)
+  int64_t output_sample_offset_;  ///< The number of samples we have already
+                                  ///< output for this signal.
+  std::vector<float> input_remainder_;  ///< A small trailing part of the
+                                        ///< previously seen input signal.
+};
+
+}  // namespace sherpa_onnx
+
+#endif  // SHERPA_ONNX_CSRC_RESAMPLE_H_

@@ -10,9 +10,6 @@
 
 #include "sherpa-onnx/csrc/online-recognizer.h"
 #include "sherpa-onnx/csrc/online-stream.h"
-#include "sherpa-onnx/csrc/online-transducer-greedy-search-decoder.h"
-#include "sherpa-onnx/csrc/online-transducer-model-config.h"
-#include "sherpa-onnx/csrc/online-transducer-model.h"
 #include "sherpa-onnx/csrc/symbol-table.h"
 #include "sherpa-onnx/csrc/wave-reader.h"
 

+# Copied from https://github.com/Tencent/ncnn/blob/master/toolchains/aarch64-linux-gnu.toolchain.cmake
+
+set(CMAKE_SYSTEM_NAME Linux)
+set(CMAKE_SYSTEM_PROCESSOR aarch64)
+
+set(CMAKE_C_COMPILER "aarch64-linux-gnu-gcc")
+set(CMAKE_CXX_COMPILER "aarch64-linux-gnu-g++")
+
+set(CMAKE_FIND_ROOT_PATH_MODE_PROGRAM NEVER)
+set(CMAKE_FIND_ROOT_PATH_MODE_LIBRARY ONLY)
+set(CMAKE_FIND_ROOT_PATH_MODE_INCLUDE ONLY)
+
+set(CMAKE_C_FLAGS "-march=armv8-a")
+set(CMAKE_CXX_FLAGS "-march=armv8-a")
+
+# cache flags
+set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS}" CACHE STRING "c flags")
+set(CMAKE_CXX_FLAGS "${CMAKE_CXX_FLAGS}" CACHE STRING "c++ flags")