Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -20,25 +20,22 @@ def save_temp_audio(audio_file):
|
|
| 20 |
"""
|
| 21 |
Speichert die hochgeladene Datei sicher mit korrekter Endung in einem temporären Verzeichnis.
|
| 22 |
"""
|
| 23 |
-
|
| 24 |
-
if isinstance(audio_file, str): # Gradio gibt es als NamedString zurück, wenn es sich um einen String handelt
|
| 25 |
-
# Hier können wir die Datei als binären Inhalt behandeln
|
| 26 |
ext = Path(audio_file).suffix
|
| 27 |
if ext.lower() not in allowed_audios:
|
| 28 |
ext = ".mp3" # Standard, falls Endung fehlt
|
| 29 |
temp_audio = Path(tempfile.mkdtemp()) / f"input{ext}"
|
| 30 |
with open(temp_audio, "wb") as f:
|
| 31 |
-
f.write(audio_file.encode())
|
| 32 |
return temp_audio
|
| 33 |
-
|
| 34 |
-
elif hasattr(audio_file, 'name'): # Wenn es sich um eine echte Datei handelt
|
| 35 |
ext = Path(audio_file.name).suffix
|
| 36 |
if ext.lower() not in allowed_audios:
|
| 37 |
-
ext = ".mp3"
|
| 38 |
temp_audio = Path(tempfile.mkdtemp()) / f"input{ext}"
|
| 39 |
-
audio_file.seek(0)
|
| 40 |
with open(temp_audio, "wb") as f:
|
| 41 |
-
shutil.copyfileobj(audio_file, f)
|
| 42 |
return temp_audio
|
| 43 |
else:
|
| 44 |
raise ValueError("Das übergebene Audio ist kein gültiges Dateiformat oder NamedString.")
|
|
@@ -66,7 +63,7 @@ def transcribe_audio(audio_file):
|
|
| 66 |
}
|
| 67 |
try:
|
| 68 |
response = requests.post(API_URL, json=payload)
|
| 69 |
-
response.raise_for_status()
|
| 70 |
except requests.RequestException as e:
|
| 71 |
return None, f"❌ API Fehler: {e}"
|
| 72 |
|
|
@@ -74,7 +71,7 @@ def transcribe_audio(audio_file):
|
|
| 74 |
text = result['choices'][0]['message']['content']
|
| 75 |
return text
|
| 76 |
|
| 77 |
-
def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_pos=0.5, fade_duration=0.7, font_size=60):
|
| 78 |
if not images:
|
| 79 |
return None, "❌ Keine Bilder ausgewählt"
|
| 80 |
|
|
@@ -82,7 +79,6 @@ def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_po
|
|
| 82 |
temp_dir = tempfile.mkdtemp()
|
| 83 |
clips = []
|
| 84 |
|
| 85 |
-
# Transkription, falls Audio vorhanden
|
| 86 |
if audio_file:
|
| 87 |
transcript, err = transcribe_audio(audio_file)
|
| 88 |
if err:
|
|
@@ -100,9 +96,8 @@ def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_po
|
|
| 100 |
texts = [""] * len(images)
|
| 101 |
temp_audio_file = None
|
| 102 |
|
| 103 |
-
# Einzelne Clips erstellen
|
| 104 |
for i, img_path in enumerate(images):
|
| 105 |
-
img_path = Path(img_path.name) #
|
| 106 |
clip_path = Path(temp_dir) / f"clip_{i}.mp4"
|
| 107 |
text = texts[i] if i < len(texts) else ""
|
| 108 |
|
|
@@ -113,8 +108,7 @@ def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_po
|
|
| 113 |
)
|
| 114 |
|
| 115 |
if text:
|
| 116 |
-
|
| 117 |
-
safe_text = shlex.quote(text) # Sichere Textverarbeitung
|
| 118 |
drawtext_filter = (
|
| 119 |
f",drawtext=text={safe_text}:fontcolor=white:fontsize={font_size}:borderw=2:"
|
| 120 |
f"x=(w-text_w)/2:y=(h-text_h)*{y_pos}:"
|
|
@@ -160,7 +154,6 @@ def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_po
|
|
| 160 |
except subprocess.CalledProcessError as e:
|
| 161 |
return None, f"❌ FFmpeg Concat Fehler:\n{e.stderr}"
|
| 162 |
|
| 163 |
-
# Audio hinzufügen, falls vorhanden
|
| 164 |
if temp_audio_file:
|
| 165 |
final_output = Path(temp_dir) / f"slideshow_audio_{uuid.uuid4().hex}.mp4"
|
| 166 |
cmd_audio = [
|
|
@@ -194,13 +187,15 @@ with gr.Blocks() as demo:
|
|
| 194 |
fade_input = gr.Number(value=0.7, label="Fade Dauer in Sekunden", precision=1)
|
| 195 |
ypos_input = gr.Slider(minimum=0.0, maximum=0.9, step=0.01, value=0.5, label="Y-Position für alle Texte (0=oben, 0.5=mitte, 0.9=unten)")
|
| 196 |
font_size_input = gr.Number(value=60, label="Textgröße (px)")
|
|
|
|
|
|
|
| 197 |
out_video = gr.Video(interactive=False, label="Generiertes Video")
|
| 198 |
status = gr.Textbox(interactive=False, label="Status")
|
| 199 |
|
| 200 |
btn = gr.Button("Video erstellen")
|
| 201 |
btn.click(
|
| 202 |
fn=generate_slideshow_with_audio,
|
| 203 |
-
inputs=[img_input, audio_input, duration_input, ypos_input, fade_input, font_size_input],
|
| 204 |
outputs=[out_video, status]
|
| 205 |
)
|
| 206 |
|
|
|
|
| 20 |
"""
|
| 21 |
Speichert die hochgeladene Datei sicher mit korrekter Endung in einem temporären Verzeichnis.
|
| 22 |
"""
|
| 23 |
+
if isinstance(audio_file, str):
|
|
|
|
|
|
|
| 24 |
ext = Path(audio_file).suffix
|
| 25 |
if ext.lower() not in allowed_audios:
|
| 26 |
ext = ".mp3" # Standard, falls Endung fehlt
|
| 27 |
temp_audio = Path(tempfile.mkdtemp()) / f"input{ext}"
|
| 28 |
with open(temp_audio, "wb") as f:
|
| 29 |
+
f.write(audio_file.encode())
|
| 30 |
return temp_audio
|
| 31 |
+
elif hasattr(audio_file, 'name'):
|
|
|
|
| 32 |
ext = Path(audio_file.name).suffix
|
| 33 |
if ext.lower() not in allowed_audios:
|
| 34 |
+
ext = ".mp3"
|
| 35 |
temp_audio = Path(tempfile.mkdtemp()) / f"input{ext}"
|
| 36 |
+
audio_file.seek(0)
|
| 37 |
with open(temp_audio, "wb") as f:
|
| 38 |
+
shutil.copyfileobj(audio_file, f)
|
| 39 |
return temp_audio
|
| 40 |
else:
|
| 41 |
raise ValueError("Das übergebene Audio ist kein gültiges Dateiformat oder NamedString.")
|
|
|
|
| 63 |
}
|
| 64 |
try:
|
| 65 |
response = requests.post(API_URL, json=payload)
|
| 66 |
+
response.raise_for_status()
|
| 67 |
except requests.RequestException as e:
|
| 68 |
return None, f"❌ API Fehler: {e}"
|
| 69 |
|
|
|
|
| 71 |
text = result['choices'][0]['message']['content']
|
| 72 |
return text
|
| 73 |
|
| 74 |
+
def generate_slideshow_with_audio(images, audio_file, duration_per_image=3, y_pos=0.5, fade_duration=0.7, font_size=60, speed=1.0):
|
| 75 |
if not images:
|
| 76 |
return None, "❌ Keine Bilder ausgewählt"
|
| 77 |
|
|
|
|
| 79 |
temp_dir = tempfile.mkdtemp()
|
| 80 |
clips = []
|
| 81 |
|
|
|
|
| 82 |
if audio_file:
|
| 83 |
transcript, err = transcribe_audio(audio_file)
|
| 84 |
if err:
|
|
|
|
| 96 |
texts = [""] * len(images)
|
| 97 |
temp_audio_file = None
|
| 98 |
|
|
|
|
| 99 |
for i, img_path in enumerate(images):
|
| 100 |
+
img_path = Path(img_path.name) # Gradio liefert temporäre Dateipfade
|
| 101 |
clip_path = Path(temp_dir) / f"clip_{i}.mp4"
|
| 102 |
text = texts[i] if i < len(texts) else ""
|
| 103 |
|
|
|
|
| 108 |
)
|
| 109 |
|
| 110 |
if text:
|
| 111 |
+
safe_text = shlex.quote(text)
|
|
|
|
| 112 |
drawtext_filter = (
|
| 113 |
f",drawtext=text={safe_text}:fontcolor=white:fontsize={font_size}:borderw=2:"
|
| 114 |
f"x=(w-text_w)/2:y=(h-text_h)*{y_pos}:"
|
|
|
|
| 154 |
except subprocess.CalledProcessError as e:
|
| 155 |
return None, f"❌ FFmpeg Concat Fehler:\n{e.stderr}"
|
| 156 |
|
|
|
|
| 157 |
if temp_audio_file:
|
| 158 |
final_output = Path(temp_dir) / f"slideshow_audio_{uuid.uuid4().hex}.mp4"
|
| 159 |
cmd_audio = [
|
|
|
|
| 187 |
fade_input = gr.Number(value=0.7, label="Fade Dauer in Sekunden", precision=1)
|
| 188 |
ypos_input = gr.Slider(minimum=0.0, maximum=0.9, step=0.01, value=0.5, label="Y-Position für alle Texte (0=oben, 0.5=mitte, 0.9=unten)")
|
| 189 |
font_size_input = gr.Number(value=60, label="Textgröße (px)")
|
| 190 |
+
speed_input = gr.Slider(minimum=0.1, maximum=3.0, value=1.0, label="Geschwindigkeit der Texteinblendung")
|
| 191 |
+
|
| 192 |
out_video = gr.Video(interactive=False, label="Generiertes Video")
|
| 193 |
status = gr.Textbox(interactive=False, label="Status")
|
| 194 |
|
| 195 |
btn = gr.Button("Video erstellen")
|
| 196 |
btn.click(
|
| 197 |
fn=generate_slideshow_with_audio,
|
| 198 |
+
inputs=[img_input, audio_input, duration_input, ypos_input, fade_input, font_size_input, speed_input],
|
| 199 |
outputs=[out_video, status]
|
| 200 |
)
|
| 201 |
|