aboutsummaryrefslogtreecommitdiffstats
path: root/tools/stt.sh
diff options
context:
space:
mode:
Diffstat (limited to 'tools/stt.sh')
-rwxr-xr-xtools/stt.sh121
1 files changed, 121 insertions, 0 deletions
diff --git a/tools/stt.sh b/tools/stt.sh
new file mode 100755
index 0000000..13a1b5a
--- /dev/null
+++ b/tools/stt.sh
@@ -0,0 +1,121 @@
+#!/bin/bash
+
+# Function to print usage information
+usage() {
+ echo "Usage: $0 [record|transcription] <options>"
+ echo ""
+ echo "Actions:"
+ echo " record Record audio from a selected source"
+ echo " transcription Transcribe audio from a .wav file"
+ echo ""
+ echo "Options for 'record':"
+ echo " -s, --source Specify the audio source (required)"
+ echo ""
+ echo "Options for 'transcription':"
+ echo " -f, --file Specify the audio file to transcribe (required)"
+ echo " -l, --lang Specify the audio file language (default: en)"
+ exit 1
+}
+
+if [[ $# -eq 0 ]]; then
+ usage
+fi
+
+# Check for required environment variable
+if [[ -z "${LLM_API_KEY}" ]]; then
+ echo "The environment variable LLM_API_KEY is not set."
+ echo 'You can use the following command: export $(xargs < ../.env))'
+ exit 1
+fi
+
+ACTION=$1
+shift
+
+host=${STT_API_HOST:-"http://localhost:8001"}
+LANG="en" # Default language
+
+if [ "$ACTION" == "record" ]; then
+ if [ "$#" -eq 0 ]; then
+ echo "Error: Source is required for record action."
+ echo "Available sources:"
+ pactl list short sources | awk '{print $2}'
+ exit 1
+ fi
+
+ SOURCE=""
+ while [[ "$#" -gt 0 ]]; do
+ case $1 in
+ -s | --source)
+ SOURCE="$2"
+ shift
+ ;;
+ *)
+ echo "Unknown parameter passed: $1"
+ usage
+ ;;
+ esac
+ shift
+ done
+
+ # Validate the provided source
+ if ! pactl list short sources | awk '{print $2}' | grep -q "^$SOURCE$"; then
+ echo "Error: Invalid audio source. Available sources:"
+ pactl list short sources | awk '{print $2}'
+ exit 1
+ fi
+
+ timestamp=$(date +"%Y%m%d_%H%M%S")
+ filename="record_${timestamp}.wav"
+ echo "Start recording to ${filename} ; use CTRL+C to terminate."
+ parec -d "${SOURCE}" --file-format=wav "${filename}"
+elif [ "$ACTION" == "transcription" ]; then
+ if [ "$#" -eq 0 ]; then
+ echo "Error: File is required for transcription action."
+ usage
+ fi
+
+ FILE=""
+ while [[ "$#" -gt 0 ]]; do
+ case $1 in
+ -f | --file)
+ FILE="$2"
+ shift
+ ;;
+ -l | --lang)
+ LANG="$2"
+ shift
+ ;;
+ *)
+ echo "Unknown parameter passed: $1"
+ usage
+ ;;
+ esac
+ shift
+ done
+
+ if [ -z "$FILE" ]; then
+ echo "Error: File is required for transcription action."
+ usage
+ fi
+
+ # Check if the file exists
+ if [ ! -f "$FILE" ]; then
+ echo "Error: File '$FILE' does not exist."
+ exit 1
+ fi
+
+ # Ensure that curl is available
+ if ! command -v curl &>/dev/null; then
+ echo "curl is required for transcription but could not be found on your system. Please install it."
+ exit 1
+ fi
+
+ # Transcribe the specified file
+ echo "Transcribing file $FILE, be patient"
+ curl "${host}/v1/audio/transcriptions" -H "Authorization: Bearer ${LLM_API_KEY}" \
+ -F "file=@${FILE}" \
+ -F "stream=true" \
+ -F "language=${LANG}"
+else
+ usage
+fi