workspace: initial clean commit

2026-04-12 21:55:33 +03:00
commit 0fc86ac3ae
220 changed files with 31657 additions and 0 deletions
--- a/scripts/send_video_note.sh
+++ b/scripts/send_video_note.sh
@@ -0,0 +1,352 @@
+#!/usr/bin/env bash
+# send_video_note.sh — Generate TTS + lip-sync video (VEED Fabric 1.0) and send as Telegram video note
+#
+# Usage:
+#   ./send_video_note.sh "Текст для озвучки"
+#   ./send_video_note.sh "Текст" [VOICE_ID] [CHAT_ID]
+#
+# Dependencies:
+#   - curl, python3, base64
+#   - ffmpeg (path in FFMPEG_BIN)
+#   - ELEVENLABS_API_KEY in ~/.openclaw/.env
+#   - FAL_KEY in ~/.openclaw/.env (format: key_id:key_secret)
+#   - BOT_TOKEN read from ~/.openclaw/openclaw.json
+
+set -euo pipefail
+
+# ── Config ────────────────────────────────────────────────────────────────────
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+ENV_FILE="$HOME/.openclaw/.env"
+CONFIG_FILE="$HOME/.openclaw/openclaw.json"
+AVATAR="$HOME/.openclaw/workspace/data/avatars/stream-avatar.jpg"
+FFMPEG_BIN="/home/node/bin/ffmpeg-7.0.2-amd64-static/ffmpeg"
+FFPROBE_BIN="/home/node/bin/ffmpeg-7.0.2-amd64-static/ffprobe"
+DEFAULT_VOICE="TPIitICAZ8CqlGZ81AKm"
+DEFAULT_MODEL="eleven_multilingual_v2"
+VIDEO_SIZE=512        # Telegram video note max dimension
+MAX_DURATION=60       # Telegram video note max seconds
+FAL_RESOLUTION="720p" # VEED Fabric 1.0: 720p or 480p
+FAL_POLL_INTERVAL=5   # seconds between status polls
+FAL_MAX_WAIT=300      # max seconds to wait for fal.ai result
+TMP_DIR=$(mktemp -d /tmp/video_note_XXXXXX)
+LOG_TAG="[video_note]"
+
+# ── Cleanup on exit ───────────────────────────────────────────────────────────
+cleanup() {
+    rm -rf "$TMP_DIR"
+}
+trap cleanup EXIT
+
+# ── Logging ───────────────────────────────────────────────────────────────────
+log()  { echo "$LOG_TAG $*" >&2; }
+err()  { echo "$LOG_TAG ERROR: $*" >&2; }
+die()  { err "$*"; exit 1; }
+
+# ── Argument parsing ──────────────────────────────────────────────────────────
+TEXT="${1:-}"
+VOICE_ID="${2:-$DEFAULT_VOICE}"
+CHAT_ID="${3:-}"
+
+[[ -z "$TEXT" ]] && die "Usage: $0 \"текст для озвучки\" [voice_id] [chat_id]"
+
+# ── Load environment ──────────────────────────────────────────────────────────
+[[ -f "$ENV_FILE" ]] || die "Env file not found: $ENV_FILE"
+# shellcheck source=/dev/null
+set -a; source "$ENV_FILE"; set +a
+
+[[ -n "${ELEVENLABS_API_KEY:-}" ]] || die "ELEVENLABS_API_KEY not set in $ENV_FILE"
+[[ -n "${FAL_KEY:-}" ]] || die "FAL_KEY not set in $ENV_FILE"
+
+# ── Get bot token and chat ID from config ─────────────────────────────────────
+get_bot_token() {
+    python3 -c "
+import json, sys
+with open('$CONFIG_FILE') as f:
+    d = json.load(f)
+accounts = d.get('channels', {}).get('telegram', {}).get('accounts', {})
+for name in ['default', list(accounts.keys())[0] if accounts else None]:
+    if name and name in accounts:
+        tok = accounts[name].get('botToken', '')
+        if tok:
+            print(tok)
+            sys.exit(0)
+sys.exit(1)
+" 2>/dev/null
+}
+
+BOT_TOKEN="$(get_bot_token)" || die "Cannot read bot token from $CONFIG_FILE"
+[[ -n "$BOT_TOKEN" ]] || die "Bot token is empty"
+
+if [[ -z "$CHAT_ID" ]]; then
+    if [[ -n "${VOICE_TTS_TARGET:-}" ]]; then
+        CHAT_ID="${VOICE_TTS_TARGET#telegram:}"
+    fi
+fi
+[[ -n "$CHAT_ID" ]] || die "CHAT_ID not provided. Pass as arg 3 or set VOICE_TTS_TARGET=telegram:<id>"
+
+# ── Sanity checks ─────────────────────────────────────────────────────────────
+[[ -f "$AVATAR" ]]     || die "Avatar not found: $AVATAR"
+[[ -f "$FFMPEG_BIN" ]] || die "FFmpeg not found: $FFMPEG_BIN"
+command -v curl    >/dev/null 2>&1 || die "curl not found"
+command -v python3 >/dev/null 2>&1 || die "python3 not found"
+command -v base64  >/dev/null 2>&1 || die "base64 not found"
+
+# ── Step 1: Generate TTS via ElevenLabs ──────────────────────────────────────
+AUDIO_FILE="$TMP_DIR/tts.mp3"
+log "Step 1: Generating TTS (voice=$VOICE_ID)..."
+
+HTTP_CODE=$(curl -s \
+    -X POST "https://api.elevenlabs.io/v1/text-to-speech/${VOICE_ID}" \
+    -H "xi-api-key: $ELEVENLABS_API_KEY" \
+    -H "Content-Type: application/json" \
+    -d "$(python3 -c "
+import json, sys
+payload = {
+    'text': sys.argv[1],
+    'model_id': '$DEFAULT_MODEL',
+    'voice_settings': {'stability': 0.5, 'similarity_boost': 0.75}
+}
+print(json.dumps(payload))
+" "$TEXT")" \
+    -o "$AUDIO_FILE" \
+    -w "%{http_code}")
+
+if [[ "$HTTP_CODE" != "200" ]]; then
+    RESP_BODY=$(cat "$AUDIO_FILE" 2>/dev/null || echo "")
+    die "ElevenLabs TTS failed (HTTP $HTTP_CODE): $RESP_BODY"
+fi
+
+AUDIO_SIZE=$(stat -c%s "$AUDIO_FILE" 2>/dev/null || stat -f%z "$AUDIO_FILE")
+[[ "$AUDIO_SIZE" -gt 1024 ]] || die "TTS audio too small (${AUDIO_SIZE} bytes) — likely an error response"
+log "✓ TTS audio: $AUDIO_FILE (${AUDIO_SIZE} bytes)"
+
+# ── Step 2: Get audio duration ────────────────────────────────────────────────
+DURATION=$("$FFPROBE_BIN" -v quiet -show_entries format=duration \
+    -of default=noprint_wrappers=1:nokey=1 "$AUDIO_FILE" 2>/dev/null || echo "0")
+DURATION_INT=$(python3 -c "import math; print(min(int(math.ceil(float('$DURATION'))), $MAX_DURATION))")
+log "✓ Audio duration: ${DURATION}s (video capped at ${DURATION_INT}s)"
+
+# ── Step 3: Upload files to fal.ai storage ────────────────────────────────────
+# fal.ai requires publicly accessible URLs — upload via fal.ai storage
+log "Step 3: Uploading files to fal.ai storage..."
+
+upload_to_fal() {
+    local file_path="$1"
+    local mime_type="$2"
+    local file_name="$3"
+
+    # Upload via fal.ai storage endpoint (CDN v3)
+    local upload_response
+    upload_response=$(curl -s -w "\n%{http_code}" --max-time 30 \
+        -X POST "https://rest.fal.ai/storage/upload/initiate?storage_type=fal-cdn-v3" \
+        -H "Authorization: Key ${FAL_KEY}" \
+        -H "Content-Type: application/json" \
+        -d "{\"content_type\": \"${mime_type}\", \"file_name\": \"${file_name}\"}" 2>&1)
+
+    local http_code
+    http_code=$(echo "$upload_response" | tail -1)
+    local body
+    body=$(echo "$upload_response" | head -n -1)
+
+    if [[ "$http_code" != "200" ]]; then
+        # Check for balance exhaustion
+        if echo "$body" | grep -qi "balance\|billing\|locked"; then
+            err "fal.ai account issue (HTTP $http_code): balance exhausted or locked. Top up at fal.ai/dashboard/billing"
+        else
+            err "fal.ai storage initiate failed (HTTP $http_code): $body"
+        fi
+        return 1
+    fi
+
+    local upload_url
+    upload_url=$(echo "$body" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('upload_url',''))" 2>/dev/null)
+    local file_url
+    file_url=$(echo "$body" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('file_url',''))" 2>/dev/null)
+
+    if [[ -z "$upload_url" ]]; then
+        err "No upload_url in fal.ai response: $body"
+        return 1
+    fi
+
+    # Upload the actual file
+    local put_code
+    put_code=$(curl -s -o /dev/null -w "%{http_code}" \
+        -X PUT "$upload_url" \
+        -H "Content-Type: ${mime_type}" \
+        --data-binary "@${file_path}" 2>&1)
+
+    if [[ "$put_code" != "200" && "$put_code" != "204" ]]; then
+        err "fal.ai file upload failed (HTTP $put_code)"
+        return 1
+    fi
+
+    echo "$file_url"
+}
+
+# Upload avatar image
+log "  Uploading avatar image..."
+IMAGE_URL=$(upload_to_fal "$AVATAR" "image/jpeg" "avatar.jpg") || {
+    log "  fal.ai storage upload failed, trying base64 fallback..."
+    # Base64 fallback: encode image as data URI
+    IMAGE_B64=$(base64 -w 0 "$AVATAR")
+    IMAGE_URL="data:image/jpeg;base64,${IMAGE_B64}"
+    log "  Using base64 data URI for image ($(echo -n "$IMAGE_URL" | wc -c) chars)"
+}
+log "✓ Image URL ready"
+
+# Upload audio file
+log "  Uploading TTS audio..."
+AUDIO_URL=$(upload_to_fal "$AUDIO_FILE" "audio/mpeg" "tts.mp3") || {
+    log "  fal.ai storage upload failed, trying base64 fallback..."
+    AUDIO_B64=$(base64 -w 0 "$AUDIO_FILE")
+    AUDIO_URL="data:audio/mpeg;base64,${AUDIO_B64}"
+    log "  Using base64 data URI for audio ($(echo -n "$AUDIO_URL" | wc -c) chars)"
+}
+log "✓ Audio URL ready"
+
+# ── Step 4: Submit VEED Fabric 1.0 job ───────────────────────────────────────
+log "Step 4: Submitting VEED Fabric 1.0 lip sync job..."
+
+FAL_PAYLOAD=$(python3 -c "
+import json, sys
+payload = {
+    'image_url': sys.argv[1],
+    'audio_url': sys.argv[2],
+    'resolution': '$FAL_RESOLUTION'
+}
+print(json.dumps(payload))
+" "$IMAGE_URL" "$AUDIO_URL")
+
+SUBMIT_RESPONSE=$(curl -s -w "\n%{http_code}" \
+    -X POST "https://queue.fal.run/veed/fabric-1.0" \
+    -H "Authorization: Key ${FAL_KEY}" \
+    -H "Content-Type: application/json" \
+    -d "$FAL_PAYLOAD" 2>&1)
+
+SUBMIT_HTTP=$(echo "$SUBMIT_RESPONSE" | tail -1)
+SUBMIT_BODY=$(echo "$SUBMIT_RESPONSE" | head -n -1)
+
+if [[ "$SUBMIT_HTTP" != "200" && "$SUBMIT_HTTP" != "201" ]]; then
+    if echo "$SUBMIT_BODY" | grep -qi "balance\|billing\|locked\|exhausted"; then
+        die "fal.ai account issue: balance exhausted or locked. Top up at https://fal.ai/dashboard/billing"
+    fi
+    die "fal.ai submit failed (HTTP $SUBMIT_HTTP): $SUBMIT_BODY"
+fi
+
+REQUEST_ID=$(echo "$SUBMIT_BODY" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('request_id',''))" 2>/dev/null)
+[[ -n "$REQUEST_ID" ]] || die "No request_id in fal.ai response: $SUBMIT_BODY"
+log "✓ Job submitted: request_id=$REQUEST_ID"
+
+# ── Step 5: Poll for VEED Fabric 1.0 result ──────────────────────────────────
+log "Step 5: Waiting for lip sync result (max ${FAL_MAX_WAIT}s)..."
+
+FAL_VIDEO_URL=""
+WAITED=0
+while [[ $WAITED -lt $FAL_MAX_WAIT ]]; do
+    sleep "$FAL_POLL_INTERVAL"
+    WAITED=$((WAITED + FAL_POLL_INTERVAL))
+
+    STATUS_RESPONSE=$(curl -s -w "\n%{http_code}" \
+        "https://queue.fal.run/veed/fabric-1.0/requests/${REQUEST_ID}/status" \
+        -H "Authorization: Key ${FAL_KEY}" 2>&1)
+
+    STATUS_HTTP=$(echo "$STATUS_RESPONSE" | tail -1)
+    STATUS_BODY=$(echo "$STATUS_RESPONSE" | head -n -1)
+
+    if [[ "$STATUS_HTTP" != "200" ]]; then
+        log "  Status poll returned HTTP $STATUS_HTTP — retrying..."
+        continue
+    fi
+
+    JOB_STATUS=$(echo "$STATUS_BODY" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('status',''))" 2>/dev/null || echo "")
+    log "  [${WAITED}s] Status: $JOB_STATUS"
+
+    if [[ "$JOB_STATUS" == "COMPLETED" ]]; then
+        # Fetch result
+        RESULT_RESPONSE=$(curl -s -w "\n%{http_code}" \
+            "https://queue.fal.run/veed/fabric-1.0/requests/${REQUEST_ID}" \
+            -H "Authorization: Key ${FAL_KEY}" 2>&1)
+
+        RESULT_HTTP=$(echo "$RESULT_RESPONSE" | tail -1)
+        RESULT_BODY=$(echo "$RESULT_RESPONSE" | head -n -1)
+
+        if [[ "$RESULT_HTTP" != "200" ]]; then
+            die "fal.ai result fetch failed (HTTP $RESULT_HTTP): $RESULT_BODY"
+        fi
+
+        FAL_VIDEO_URL=$(echo "$RESULT_BODY" | python3 -c "
+import json,sys
+d=json.load(sys.stdin)
+# Output schema: {\"video\": {\"url\": \"...\"}}
+video = d.get('video', {})
+print(video.get('url', ''))
+" 2>/dev/null)
+
+        [[ -n "$FAL_VIDEO_URL" ]] || die "No video URL in fal.ai result: $RESULT_BODY"
+        log "✓ Lip sync video ready: $FAL_VIDEO_URL"
+        break
+
+    elif [[ "$JOB_STATUS" == "FAILED" || "$JOB_STATUS" == "ERROR" ]]; then
+        ERROR_MSG=$(echo "$STATUS_BODY" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('error', d.get('detail', 'unknown')))" 2>/dev/null || echo "$STATUS_BODY")
+        die "fal.ai job failed: $ERROR_MSG"
+    fi
+    # IN_QUEUE or IN_PROGRESS — keep polling
+done
+
+[[ -n "$FAL_VIDEO_URL" ]] || die "Timeout: fal.ai did not complete within ${FAL_MAX_WAIT}s"
+
+# ── Step 6: Download lip sync video ──────────────────────────────────────────
+log "Step 6: Downloading lip sync video..."
+RAW_VIDEO="$TMP_DIR/lipsync_raw.mp4"
+
+DL_HTTP=$(curl -s -L \
+    -o "$RAW_VIDEO" \
+    -w "%{http_code}" \
+    "$FAL_VIDEO_URL" 2>&1)
+
+[[ "$DL_HTTP" == "200" ]] || die "Video download failed (HTTP $DL_HTTP)"
+RAW_SIZE=$(stat -c%s "$RAW_VIDEO" 2>/dev/null || stat -f%z "$RAW_VIDEO")
+[[ "$RAW_SIZE" -gt 10240 ]] || die "Downloaded video too small (${RAW_SIZE} bytes)"
+log "✓ Downloaded: $RAW_VIDEO (${RAW_SIZE} bytes)"
+
+# ── Step 7: Convert to square 1:1 format for Telegram video note ─────────────
+log "Step 7: Converting to ${VIDEO_SIZE}x${VIDEO_SIZE} square for Telegram..."
+VIDEO_FILE="$TMP_DIR/video_note.mp4"
+
+"$FFMPEG_BIN" -y \
+    -i "$RAW_VIDEO" \
+    -vf "scale=${VIDEO_SIZE}:${VIDEO_SIZE}:force_original_aspect_ratio=increase,crop=${VIDEO_SIZE}:${VIDEO_SIZE}" \
+    -c:v libx264 -preset fast -crf 23 \
+    -c:a aac -b:a 64k \
+    -t "$DURATION_INT" \
+    -movflags +faststart \
+    "$VIDEO_FILE" 2>/dev/null
+
+[[ -f "$VIDEO_FILE" ]] || die "FFmpeg square conversion failed"
+VIDEO_SIZE_BYTES=$(stat -c%s "$VIDEO_FILE" 2>/dev/null || stat -f%z "$VIDEO_FILE")
+log "✓ Final video: $VIDEO_FILE (${VIDEO_SIZE_BYTES} bytes, ${DURATION_INT}s)"
+
+# Telegram video note limit: 50MB
+MAX_VIDEO_BYTES=$((50 * 1024 * 1024))
+[[ "$VIDEO_SIZE_BYTES" -lt "$MAX_VIDEO_BYTES" ]] || die "Video too large for Telegram (${VIDEO_SIZE_BYTES} > 50MB)"
+
+# ── Step 8: Send as Telegram video note (кружочек) ────────────────────────────
+log "Step 8: Sending video note to Telegram (chat=$CHAT_ID)..."
+
+RESPONSE=$(curl -s \
+    -X POST "https://api.telegram.org/bot${BOT_TOKEN}/sendVideoNote" \
+    -F "chat_id=${CHAT_ID}" \
+    -F "video_note=@${VIDEO_FILE}" \
+    -F "length=${VIDEO_SIZE}" \
+    -F "duration=${DURATION_INT}" 2>&1)
+
+SUCCESS=$(echo "$RESPONSE" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('ok', False))" 2>/dev/null || echo "false")
+
+if [[ "$SUCCESS" == "True" ]]; then
+    MSG_ID=$(echo "$RESPONSE" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d['result']['message_id'])" 2>/dev/null || echo "?")
+    log "✅ Video note sent! Message ID: $MSG_ID"
+    echo "$MSG_ID"
+else
+    TG_ERR=$(echo "$RESPONSE" | python3 -c "import json,sys; d=json.load(sys.stdin); print(d.get('description','unknown error'))" 2>/dev/null || echo "$RESPONSE")
+    die "Telegram API error: $TG_ERR"
+fi