Add overlays

animate.py

@@ -9,7 +9,7 @@ import numpy as np
 from frame_interpolation.eval import interpolator, util
 from huggingface_hub import snapshot_download
 from image_tools.sizes import resize_and_crop
-from moviepy.editor import CompositeVideoClip
+from moviepy.editor import CompositeVideoClip, ImageClip
 from moviepy.editor import VideoFileClip as vfc
 from PIL import Image
 
@@ -210,7 +210,9 @@ def get_video_frames(
     return frames, cv2_images
 
 
-def create_mp4_with_audio(frames, cv2_images, duration, audio, output_path):
+def create_mp4_with_audio(
+    frames, cv2_images, duration, audio, output_path, overlay_image
+):
     vid_output_dir = output_path.parent
     temp_vid_path = Path(vid_output_dir / "TEMP.mp4")
     mediapy.write_video(temp_vid_path, frames, fps=10)
@@ -234,6 +236,12 @@ def create_mp4_with_audio(frames, cv2_images, duration, audio, output_path):
 
     new_clip = CompositeVideoClip([clip1, clip2, clip3])
     new_clip.audio = audio  # Naviely append audio without considering the length of the video, could be a problem, no idea, but it works, so I'm not touching it
+    image = (
+        ImageClip(overlay_image).set_duration(duration).resize(0.5).set_pos("center")
+    )
+
     new_clip.set_duration(duration)
-    new_clip.write_videofile(output_path.as_posix(), audio_codec="aac")
+    # Now overlay the image with moviepy
+    final_clip = CompositeVideoClip([new_clip, image])
+    final_clip.write_videofile(output_path.as_posix(), audio_codec="aac")
     return output_path.as_posix()

app.py

@@ -1,4 +1,5 @@
 import os
+import pathlib
 import shutil
 import subprocess
 import sys
@@ -6,17 +7,14 @@ import uuid
 from pathlib import Path
 
 import gradio as gr
-from moviepy.editor import AudioFileClip
-
-
 import opengraph
-import pathlib
-
 import requests
+from moviepy.editor import AudioFileClip
 
 output_dir = Path("temp/").absolute()
 output_dir.mkdir(exist_ok=True, parents=True)
 
+
 class SpotifyApi:
     spotify_directory = Path(output_dir / "spotify")
 
@@ -52,6 +50,7 @@ class SpotifyApi:
         r = requests.get(image, allow_redirects=True)
         path = self.perma_output_path.with_suffix(".jpg").absolute()
         open(path, "wb").write(r.content)
+        return path
 
     def get_title(self) -> str:
         return self.opengraph["title"]
@@ -81,12 +80,15 @@ def process_inputs(
     if spotify_url:
         spotify = SpotifyApi(spotify_url)
         audio_input.path = spotify.download_episode()
+        spotify_image = spotify.download_image()
     images = get_stable_diffusion_images(prompt)
-    video = animate_images(images, audio_input)
+    video = animate_images(images, audio_input, spotify_image)
     return video
 
 
-def animate_images(image_paths: list[str], audio_input: AudioInput) -> str:
+def animate_images(
+    image_paths: list[str], audio_input: AudioInput, overlay_image_path: str
+) -> str:
     from animate import (  # Only import after git clone and when necessary takes loooong
         create_mp4_with_audio,
         get_video_frames,
@@ -103,7 +105,12 @@ def animate_images(image_paths: list[str], audio_input: AudioInput) -> str:
     video_frames, cv2_images = get_video_frames(image_paths, vid_output_dir)
     path = Path(vid_output_dir / "output_final.mp4")
     return create_mp4_with_audio(
-        video_frames, cv2_images, audio_clip.duration, audio_clip, path
+        video_frames,
+        cv2_images,
+        audio_clip.duration,
+        audio_clip,
+        path,
+        overlay_image_path.as_posix(),
     )
 
 
@@ -123,7 +130,15 @@ iface = gr.Interface(
         gr.Number(label="Run for (in seconds)"),
     ],
     outputs="video",
-    examples = [["A podcast clip", None, "https://open.spotify.com/episode/31u9tI8t5IFrdv3QhZtPHI",50,60]]
+    examples=[
+        [
+            "A podcast clip",
+            None,
+            "https://open.spotify.com/episode/31u9tI8t5IFrdv3QhZtPHI",
+            50,
+            60,
+        ]
+    ],
 )
 
 if __name__ == "__main__":
@@ -141,4 +156,4 @@ if __name__ == "__main__":
     os.chdir(
         output_dir
     )  # change working directory to output_dir because the hf spaces model has no option to specify output directory ¯\_(ツ)_/¯
-    iface.launch(share=True)
+    iface.launch()

test.ipynb

@@ -36,6 +36,29 @@
     "test_spotify_downloads()"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "id": "237e3903",
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/plain": [
+       "<moviepy.video.VideoClip.ImageClip at 0x7f5951ec8700>"
+      ]
+     },
+     "execution_count": 4,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "source": [
+    "from moviepy.editor import ImageClip\n",
+    "\n",
+    "ImageClip(img=\"/notebooks/stablepod/temp/1ae3a379.jpg\")"
+   ]
+  },
   {
    "cell_type": "code",
    "execution_count": 1,
@@ -68,7 +91,7 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3 (ipykernel)",
+   "display_name": "Python 3.9.6 64-bit",
    "language": "python",
    "name": "python3"
   },
@@ -83,6 +106,11 @@
    "nbconvert_exporter": "python",
    "pygments_lexer": "ipython3",
    "version": "3.9.13"
+  },
+  "vscode": {
+   "interpreter": {
+    "hash": "31f2aee4e71d21fbe5cf8b01ff0e069b9275f58929596ceb00d14d90e3e16cd6"
+   }
   }
  },
  "nbformat": 4,