Compare commits

...

3 Commits

Author SHA1 Message Date
Aryan Rahar
0100afcdad
Merge a8923227a6 into edea8a9c3c 2025-09-08 06:32:48 +02:00
Siva Mahadevan
edea8a9c3c
whisper : prefer curl over wget in download scripts (#3409)
Some checks failed
Bindings Tests (Ruby) / ubuntu-22 (push) Has been cancelled
Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/main-cuda.Dockerfile platform:linux/amd64 tag:main-cuda]) (push) Has been cancelled
Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/main-intel.Dockerfile platform:linux/amd64 tag:main-intel]) (push) Has been cancelled
Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/main-musa.Dockerfile platform:linux/amd64 tag:main-musa]) (push) Has been cancelled
Publish Docker image / Push Docker image to Docker Hub (map[dockerfile:.devops/main.Dockerfile platform:linux/amd64 tag:main]) (push) Has been cancelled
Examples WASM / deploy-wasm-github-pages (push) Has been cancelled
On busybox-based systems like Alpine Linux, wget does not have
certain CLI flags such as '--no-config'. Thus, search for the
existence of 'curl' first in the PATH before wget. wget2 is
still the preferred download tool.
2025-09-08 06:32:19 +02:00
Aryan Rahar
a8923227a6 cli: extract validate_audio_extension(); clearer decode-failed hint; timings only on success 2025-09-03 13:57:47 +00:00
2 changed files with 53 additions and 8 deletions

View File

@ -1,3 +1,5 @@
#include <filesystem>
#include <algorithm>
#include "common.h"
#include "common-whisper.h"
@ -29,6 +31,37 @@ static void replace_all(std::string & s, const std::string & search, const std::
s.insert(pos, replace);
}
}
// helper: validate input file extension
static bool validate_audio_extension(const std::string & fname_inp) {
if (fname_inp == "-") return true; // allow stdin
std::string ext;
try {
ext = std::filesystem::path(fname_inp).extension().string();
std::transform(ext.begin(), ext.end(), ext.begin(),
[](unsigned char c){ return std::tolower(c); });
} catch (...) {
// if path parsing fails, let the decoder try anyway
return true;
}
auto ext_supported = [](const std::string &e) {
// keep in sync with usage text
return e == ".wav" || e == ".mp3" || e == ".flac" || e == ".ogg";
};
if (!ext.empty() && !ext_supported(ext)) {
fprintf(stderr,
"error: unsupported audio extension '%s' for '%s'.\n"
"supported: flac, mp3, ogg, wav.\n"
"hint: convert with ffmpeg, e.g.:\n"
" ffmpeg -i \"%s\" -ar 16000 -ac 1 -c:a pcm_s16le out.wav\n",
ext.c_str(), fname_inp.c_str(), fname_inp.c_str());
return false;
}
return true;
}
// command-line parameters
struct whisper_params {
@ -1051,8 +1084,13 @@ int main(int argc, char ** argv) {
}
}
bool processed_any = false;
for (int f = 0; f < (int) params.fname_inp.size(); ++f) {
const auto & fname_inp = params.fname_inp[f];
if (!validate_audio_extension(fname_inp)) {
continue;
}
struct fout_factory {
std::string fname_out;
const size_t basename_length;
@ -1105,10 +1143,15 @@ int main(int argc, char ** argv) {
std::vector<float> pcmf32; // mono-channel F32 PCM
std::vector<std::vector<float>> pcmf32s; // stereo-channel F32 PCM
if (!::read_audio_data(fname_inp, pcmf32, pcmf32s, params.diarize)) {
fprintf(stderr, "error: failed to read audio file '%s'\n", fname_inp.c_str());
continue;
}
if (!::read_audio_data(fname_inp, pcmf32, pcmf32s, params.diarize)) {
fprintf(stderr,
"error: failed to decode audio from '%s'.\n"
"Make sure the file is not corrupted and has one of: flac, mp3, ogg, wav.\n"
"If you still hit this, convert to a standard WAV with:\n"
" ffmpeg -i \"%s\" -ar 16000 -ac 1 -c:a pcm_s16le out.wav\n",
fname_inp.c_str(), fname_inp.c_str());
continue;
}
if (!whisper_is_multilingual(ctx)) {
if (params.language != "en" || params.translate) {
@ -1258,6 +1301,8 @@ int main(int argc, char ** argv) {
fprintf(stderr, "%s: failed to process audio\n", argv[0]);
return 10;
}
processed_any = true;
}
// output stuff
@ -1286,7 +1331,7 @@ int main(int argc, char ** argv) {
}
}
if (!params.no_prints) {
if (processed_any && !params.no_prints) {
whisper_print_timings(ctx);
}
whisper_free(ctx);

View File

@ -119,12 +119,12 @@ fi
if [ -x "$(command -v wget2)" ]; then
wget2 --no-config --progress bar -O ggml-"$model".bin $src/$pfx-"$model".bin
elif [ -x "$(command -v wget)" ]; then
wget --no-config --quiet --show-progress -O ggml-"$model".bin $src/$pfx-"$model".bin
elif [ -x "$(command -v curl)" ]; then
curl -L --output ggml-"$model".bin $src/$pfx-"$model".bin
elif [ -x "$(command -v wget)" ]; then
wget --no-config --quiet --show-progress -O ggml-"$model".bin $src/$pfx-"$model".bin
else
printf "Either wget or curl is required to download models.\n"
printf "Either wget2, curl, or wget is required to download models.\n"
exit 1
fi