From eba3757572b04a2285ad77d5b5e38d6203237317 Mon Sep 17 00:00:00 2001 From: Hermes Agent Date: Wed, 17 Jun 2026 19:52:07 +0200 Subject: [PATCH] Server-seitige TTS-Sprachausgabe (Thorsten via Piper) MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit - Neuer /api/speak Endpunkt: Text โ†’ Piper TTS โ†’ WAV-Audio - Frontend speak() ruft jetzt Server-Endpunkt statt Web Speech API - Browser-TTS nur als Fallback wenn Server nicht erreichbar - ๐Ÿ”Š Button prominenter: BSN-blauer Hintergrund + sichtbarer Rand - Fix: SVG-Icons auf Beitrags-Detailseite statt Emojis - Fix: Datum bei Text-Only-Karten unten rechts (top:auto) --- app.py | 100 +++++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 83 insertions(+), 17 deletions(-) diff --git a/app.py b/app.py index 547287a..296e93f 100644 --- a/app.py +++ b/app.py @@ -1881,7 +1881,7 @@ def public_frontend(): .story-preview { display:none; } .text-only .story-preview { display:-webkit-box; -webkit-line-clamp:4; -webkit-box-orient:vertical; overflow:hidden; text-overflow:ellipsis; font-size:0.8rem; line-height:1.4; color:#444; margin-bottom:0.35rem; } .text-only .card-content { position:relative; padding:0.55rem 0.85rem 0.65rem; } - .text-only .story-time-overlay { position:absolute; top:0.45rem; right:0.6rem; left:auto; z-index:5; } + .text-only .story-time-overlay { position:absolute; top:auto; bottom:0.45rem; right:0.6rem; left:auto; z-index:5; } /* Frontend filter bar */ .frontend-filter { max-width:1100px; margin:1rem auto 0; padding:0 1.5rem; display:flex; gap:0.4rem; flex-wrap:wrap; align-items:center; } @@ -2059,7 +2059,7 @@ def public_frontend():
Hey! ๐Ÿ‘‹ Ich bin der BSN Community Assistant. Frag mich nach Spielen, Stรคnden oder was auf der SPIEL Essen los ist!
- +
@@ -2278,9 +2278,9 @@ def public_frontend(): const btn = document.createElement('button'); btn.innerHTML = '๐Ÿ”Š'; btn.title = 'Vorlesen'; - btn.style.cssText = 'background:none;border:none;cursor:pointer;font-size:0.85rem;opacity:0.5;padding:0 0 0.15rem 0;transition:opacity 0.15s;flex-shrink:0;'; - btn.onmouseenter = () => btn.style.opacity = '1'; - btn.onmouseleave = () => btn.style.opacity = '0.5'; + btn.style.cssText = 'background:rgba(32,34,138,0.08);border:1px solid rgba(32,34,138,0.2);border-radius:8px;cursor:pointer;font-size:0.9rem;padding:0.15rem 0.4rem;transition:all 0.15s;flex-shrink:0;'; + btn.onmouseenter = () => { btn.style.background = 'rgba(32,34,138,0.18)'; btn.style.borderColor = 'rgba(32,34,138,0.4)'; }; + btn.onmouseleave = () => { btn.style.background = 'rgba(32,34,138,0.08)'; btn.style.borderColor = 'rgba(32,34,138,0.2)'; }; btn.onclick = () => speak(text); wrapper.appendChild(btn); } @@ -2290,17 +2290,31 @@ def public_frontend(): return div; } - function speak(text) { - if (!window.speechSynthesis) return; - window.speechSynthesis.cancel(); - const u = new SpeechSynthesisUtterance(text); - u.lang = 'de-DE'; - u.rate = 0.95; - // Prefer German voice - const voices = speechSynthesis.getVoices(); - const de = voices.find(v => v.lang.startsWith('de')); - if (de) u.voice = de; - speechSynthesis.speak(u); + async function speak(text) { + try { + const r = await fetch('/api/speak', { + method:'POST', + headers:{'Content-Type':'application/json'}, + body:JSON.stringify({text:text}) + }); + if (!r.ok) throw new Error('TTS failed'); + const blob = await r.blob(); + const audio = new Audio(URL.createObjectURL(blob)); + audio.play(); + } catch(e) { + // Fallback: browser TTS + if (window.speechSynthesis) { + window.speechSynthesis.cancel(); + const u = new SpeechSynthesisUtterance(text); + u.lang = 'de-DE'; + u.rate = 0.95; + const voices = speechSynthesis.getVoices(); + const thorsten = voices.find(v => v.name === 'Thorsten'); + if (thorsten) u.voice = thorsten; + else { const de = voices.find(v => v.lang.startsWith('de')); if (de) u.voice = de; } + speechSynthesis.speak(u); + } + } } // Preload voices @@ -2485,6 +2499,51 @@ def api_chat(): return jsonify({'reply': reply}) +@app.route('/api/speak', methods=['POST']) +def api_speak(): + """Generate TTS audio via Piper (Thorsten voice) and return as WAV.""" + import subprocess + import tempfile + data = request.get_json(force=True, silent=True) or {} + text = (data.get('text') or '').strip() + if not text or len(text) > 2000: + return jsonify({'error': 'Text required, max 2000 chars'}), 400 + + model = '/home/hermes/.local/share/piper-voices/de_DE-thorsten-medium.onnx' + if not os.path.exists(model): + model = '/home/hermes/.hermes/cache/piper-voices/de_DE-thorsten-medium.onnx' + if not os.path.exists(model): + return jsonify({'error': 'Thorsten voice not found'}), 500 + + with tempfile.NamedTemporaryFile(suffix='.wav', delete=False) as tmp: + tmp_path = tmp.name + + try: + env = os.environ.copy() + env['LD_LIBRARY_PATH'] = '/home/hermes/.local/lib:' + env.get('LD_LIBRARY_PATH', '') + env['ESPEAK_DATA_PATH'] = '/home/hermes/.local/share/espeak-ng-data' + result = subprocess.run( + ['/home/hermes/.local/bin/piper', '--model', model, '--output_file', tmp_path], + input=text.encode('utf-8'), + capture_output=True, + timeout=30, + env=env, + ) + if result.returncode != 0: + return jsonify({'error': 'TTS generation failed'}), 500 + + from flask import send_file + response = send_file(tmp_path, mimetype='audio/wav', as_attachment=False) + response.headers['X-Content-Type-Options'] = 'nosniff' + return response + except Exception as e: + return jsonify({'error': f'TTS failed: {str(e)}'}), 500 + finally: + # Clean up temp file after response is sent + import threading + threading.Timer(5.0, lambda: os.unlink(tmp_path) if os.path.exists(tmp_path) else None).start() + + @app.route('/api/submit', methods=['POST']) def api_submit(): """Public submission endpoint โ€” text + optional media upload.""" @@ -2638,7 +2697,14 @@ def public_detail(sub_id: int): except Exception: pass - type_emoji = {"text": "๐Ÿ’ฌ", "image": "๐Ÿ–ผ๏ธ", "audio": "๐ŸŽค", "video": "๐ŸŽฌ", "document": "๐Ÿ“Ž"}.get(row["type"], "๐Ÿ“Œ") + _detail_svg_icons = { + "text": '', + "image": '', + "audio": '', + "video": '', + "document": '', + } + type_emoji = _detail_svg_icons.get(row["type"], _detail_svg_icons["document"]) name = (row["publish_name"] or row["sender_name"] or "Anonym").strip() if not name or name.lower() in ("unknown", "anonym"): name = "Anonym"