Script meetings/bin/meeting-start

Starts a new live transcription session.

./bin/meeting-start

Usage

• German: meeting-start --de

• English: meeting-start --en

• Auto: meeting-start --auto

--de and --auto do NOT work at this stage of the project. Only use --en.

meeting-start --en

Script

#!/usr/bin/env bash
set -euo pipefail

HERE="$(cd "$(dirname "$0")" && pwd)"
source "$HERE/../lib/paths.sh"
source "$HERE/../lib/whisper.sh"

# ------------------------------------------------------------
# Argument parsing
# ------------------------------------------------------------

LANG_MODE="en"
NO_AUDIO=0
NO_TRANSCRIPT=0

while [[ $# -gt 0 ]]; do
  case "$1" in
    --de)            LANG_MODE="de";    shift ;;
    --en)            LANG_MODE="en";    shift ;;
    --auto)          LANG_MODE="auto";  shift ;;
    --no-audio)      NO_AUDIO=1;        shift ;;
    --no-transcript) NO_TRANSCRIPT=1;   shift ;;
    *)
      echo "Usage: meeting-start [--de|--en|--auto] [--no-audio] [--no-transcript]"
      exit 2
      ;;
  esac
done

if [[ "$NO_AUDIO" -eq 1 && "$NO_TRANSCRIPT" -eq 1 ]]; then
  echo "❌ --no-audio and --no-transcript together: nothing to do." >&2
  exit 2
fi

# ------------------------------------------------------------
# Select model + language options for whisper-stream
# ------------------------------------------------------------

case "$LANG_MODE" in
  en)
    MODEL="$MODEL_EN"
    LANG_OPTS=(-l en)
    ;;
  de)
    MODEL="$MODEL_MULTI"
    LANG_OPTS=(-l de)
    ;;
  auto)
    MODEL="$MODEL_MULTI"
    LANG_OPTS=()
    ;;
esac

# ------------------------------------------------------------
# Audio setup: persistent sink + virtual microphone
# ------------------------------------------------------------

DISCORD_SINK="discord_sink"
DISCORD_MONITOR="${DISCORD_SINK}.monitor"
VIRTUAL_MIC="whisper_mic"

echo "📁 Session: $SESSION"
echo "🗣️  Language mode: $LANG_MODE"
echo "🎙️  Audio recording: $([ "$NO_AUDIO" -eq 1 ] && echo 'off (--no-audio)' || echo 'on')"
echo "📝 Live transcript: $([ "$NO_TRANSCRIPT" -eq 1 ] && echo 'off (--no-transcript)' || echo 'on')"
echo "🔧 Setting up audio routing…"
echo "   Sink:   $DISCORD_SINK"
echo "   Source: $VIRTUAL_MIC (master: $DISCORD_MONITOR)"
echo "----"

# 1) Ensure discord_sink exists
if ! pactl list short sinks | awk '{print $2}' | grep -qx "$DISCORD_SINK"; then
  pactl load-module module-null-sink \
    sink_name="$DISCORD_SINK" \
    sink_properties=device.description=DiscordSink >/dev/null
fi

# 2) Try to move an active sink-input (Discord) to discord_sink
#    This may be transient in silent channels → poll briefly.
moved="no"
for _ in {1..40}; do
  SID="$(pactl list short sink-inputs 2>/dev/null | awk 'NF{print $1}' | head -n1 || true)"
  if [[ -n "${SID:-}" ]]; then
    if pactl move-sink-input "$SID" "$DISCORD_SINK" 2>/dev/null; then
      moved="yes"
      break
    fi
  fi
  sleep 0.25
done

if [[ "$moved" != "yes" ]]; then
  echo "⚠️  Could not move a sink-input to $DISCORD_SINK."
  echo "    Make sure Legcord is connected to a voice channel, then re-run meeting-start."
fi

# 3) Remove existing whisper_mic remap-sources (idempotent)
for mid in $(pactl list short modules | awk '$0 ~ /module-remap-source/ && $0 ~ /source_name=whisper_mic/ {print $1}'); do
  pactl unload-module "$mid" >/dev/null 2>&1 || true
done

# 4) Create virtual microphone (mono; attempt 16 kHz)
REMAPPED_MID="$(pactl load-module module-remap-source \
  master="$DISCORD_MONITOR" \
  source_name="$VIRTUAL_MIC" \
  channels=1 \
  rate=16000 \
  master_channel_map=front-left \
  channel_map=mono \
  source_properties=device.description=WhisperMic16k)"

echo "✅ Audio ready (remap module id: $REMAPPED_MID)"
echo "----"

# ------------------------------------------------------------
# Start recording (ffmpeg) and/or transcription (whisper-stream)
# ------------------------------------------------------------

FFMPEG_PID=""
WHISPER_PID=""

# --- ffmpeg: record to WAV ---
if [[ "$NO_AUDIO" -eq 0 ]]; then
  if [[ -e "$AUDIO" ]]; then
    echo "❌ Refusing to overwrite existing file: $AUDIO" >&2
    echo "   (Session dir collision or leftover file)" >&2
    exit 1
  fi

  ffmpeg -nostdin -y -hide_banner -loglevel error \
    -f pulse -i "$VIRTUAL_MIC" \
    -ac 1 -ar 16000 \
    "$AUDIO" &
  FFMPEG_PID=$!
  echo "🎙️  Recording to: $AUDIO"
fi

# --- whisper-stream: live transcription ---
# whisper-stream reads directly from the capture device (-c 1 in STREAM_OPTS).
# Do NOT pass -f: for whisper-stream, -f is a text output file, not an audio input.
if [[ "$NO_TRANSCRIPT" -eq 0 ]]; then
  "$WHISPER" \
    -m "$MODEL" \
    "${LANG_OPTS[@]}" \
    "${STREAM_OPTS[@]}" \
    "${THREAD_OPTS[@]}" \
    | tee -a "$TRANSCRIPT" &
  WHISPER_PID=$!
  echo "📝 Live transcription to: $TRANSCRIPT"
fi

echo "➡️  Follow the transcript with: meeting-follow"
echo "⏹  Stop with: meeting-stop"
echo "----"

# Announce active session
echo "$SESSION" > "$CURRENT"

# ------------------------------------------------------------
# Metadata
# ------------------------------------------------------------

cat > "$META" <<EOF
SESSION=$SESSION
AUDIO=${AUDIO:-}
TRANSCRIPT=${TRANSCRIPT:-}
LANG_MODE=$LANG_MODE
MODEL=$MODEL
NO_AUDIO=$NO_AUDIO
NO_TRANSCRIPT=$NO_TRANSCRIPT
DISCORD_SINK=$DISCORD_SINK
DISCORD_MONITOR=$DISCORD_MONITOR
VIRTUAL_MIC=$VIRTUAL_MIC
REMAPPED_MID=$REMAPPED_MID
FFMPEG_PID=${FFMPEG_PID:-}
WHISPER_PID=${WHISPER_PID:-}
EOF

[[ -n "${FFMPEG_PID:-}" ]]  && echo "$FFMPEG_PID"  > "$SESSION/ffmpeg.pid"
[[ -n "${WHISPER_PID:-}" ]] && echo "$WHISPER_PID" > "$SESSION/whisper.pid"

echo "✅ meeting-start finished setup successfully."