#include "KaldiLib/Features.h" #include "KaldiLib/Labels.h" #include "KaldiLib/Common.h" #include "KaldiLib/UserInterface.h" #include <string> #define SNAME "TNET" extern "C" { #include "cwrapper.h" #include "string.h" #include "nerv/common.h" extern Matrix *nerv_matrix_host_float_create(long nrow, long ncol, Status *status); struct TNetFeatureRepo { TNet::FeatureRepository feature_repo; TNet::UserInterface ui; bool swap_features; int target_kind; int deriv_order; int* p_deriv_win_lenghts; int start_frm_ext; int end_frm_ext; char* cmn_path; char* cmn_file; const char* cmn_mask; char* cvn_path; char* cvn_file; const char* cvn_mask; const char* cvg_file; TNet::Matrix<float> feats_host; /* KaldiLib implementation */ }; TNetFeatureRepo *tnet_feature_repo_new(const char *p_script, const char *config, int context) { TNetFeatureRepo *repo = new TNetFeatureRepo(); repo->ui.ReadConfig(config); repo->swap_features = !repo->ui.GetBool(SNAME":NATURALREADORDER", TNet::IsBigEndian()); /* load defaults */ repo->target_kind = repo->ui.GetFeatureParams(&repo->deriv_order, &repo->p_deriv_win_lenghts, &repo->start_frm_ext, &repo->end_frm_ext, &repo->cmn_path, &repo->cmn_file, &repo->cmn_mask, &repo->cvn_path, &repo->cvn_file, &repo->cvn_mask, &repo->cvg_file, SNAME":", 0); repo->start_frm_ext = repo->end_frm_ext = context; repo->feature_repo.Init(repo->swap_features, repo->start_frm_ext, repo->end_frm_ext, repo->target_kind, repo->deriv_order, repo->p_deriv_win_lenghts, repo->cmn_path, repo->cmn_mask, repo->cvn_path, repo->cvn_mask, repo->cvg_file); repo->feature_repo.AddFileList(p_script); repo->feature_repo.Rewind(); return repo; } Matrix *tnet_feature_repo_read_utterance(TNetFeatureRepo *repo, lua_State *L, int debug) { Matrix *mat; /* nerv implementation */ repo->feature_repo.ReadFullMatrix(repo->feats_host); std::string utter_str = repo->feature_repo.Current().Logical(); repo->feats_host.CheckData(utter_str); int n = repo->feats_host.Rows(); int m = repo->feats_host.Cols(); Status status; mat = nerv_matrix_host_float_create(n, m, &status); NERV_LUA_CHECK_STATUS(L, status); size_t stride = mat->stride; if (debug) fprintf(stderr, "[tnet] feature: %s %d %d\n", utter_str.c_str(), n, m); for (int i = 0; i < n; i++) { float *row = repo->feats_host.pRowData(i); float *nerv_row = (float *)((char *)mat->data.f + i * stride); /* use memmove to copy the row, since KaldiLib uses compact storage */ memmove(nerv_row, row, sizeof(float) * m); } return mat; } void tnet_feature_repo_next(TNetFeatureRepo *repo) { repo->feature_repo.MoveNext(); } int tnet_feature_repo_is_end(TNetFeatureRepo *repo) { return repo->feature_repo.EndOfList(); } size_t tnet_feature_repo_current_samplerate(TNetFeatureRepo *repo) { return repo->feature_repo.CurrentHeader().mSamplePeriod; } const char *tnet_feature_repo_current_tag(TNetFeatureRepo *repo) { return repo->feature_repo.Current().Logical().c_str(); } void tnet_feature_repo_destroy(TNetFeatureRepo *repo) { if (repo->cmn_mask) free(repo->cmn_path); if (repo->cvn_mask) free(repo->cvn_path); free(repo->p_deriv_win_lenghts); delete repo; } struct TNetLabelRepo { TNet::LabelRepository label_repo; }; TNetLabelRepo *tnet_label_repo_new(const char *mlf, const char *fmt, const char *fmt_arg, const char *dir, const char *ext) { TNetLabelRepo *repo = new TNetLabelRepo(); repo->label_repo.InitExt(mlf, fmt, fmt_arg, dir, ext); /* repo->label_repo.Init(mlf, fmt_arg, dir, ext); */ return repo; } Matrix *tnet_label_repo_read_utterance(TNetLabelRepo *repo, size_t frames, size_t sample_rate, const char *tag, lua_State *L, int debug) { std::vector<TNet::Matrix<float> > labs_hosts; /* KaldiLib implementation */ Matrix *mat; repo->label_repo.GenDesiredMatrixExt(labs_hosts, frames, sample_rate, tag); int n = labs_hosts[0].Rows(); int m = labs_hosts[0].Cols(); Status status; mat = nerv_matrix_host_float_create(n, m, &status); NERV_LUA_CHECK_STATUS(L, status); size_t stride = mat->stride; if (debug) fprintf(stderr, "[tnet] label: %s %d %d\n", tag, n, m); for (int i = 0; i < n; i++) { float *row = labs_hosts[0].pRowData(i); float *nerv_row = (float *)((char *)mat->data.f + i * stride); /* use memmove to copy the row, since KaldiLib uses compact storage */ memmove(nerv_row, row, sizeof(float) * m); } return mat; } void tnet_label_repo_destroy(TNetLabelRepo *repo) { delete repo; } }