Merge cb86fbd47d into 1174275b5b

.gitignore

@@ -2,4 +2,7 @@
 /venv
 /narration
 /frames/*
-!/frames/.gitkeep
+!/frames/.gitkeep
+
+# Mac-OS specific
+.DS_Store

README.md

@@ -20,11 +20,24 @@ Then, install the dependencies:
 
 Make a [Replicate](https://replicate.com), [OpenAI](https://beta.openai.com/), and [ElevenLabs](https://elevenlabs.io) account and set your tokens:
 
-```
-export OPENAI_API_KEY=<token>
-export ELEVENLABS_API_KEY=<eleven-token>
-```
+### Setting Up Environment Variables
 
+Instead of setting your tokens directly in the terminal, we'll use a `.env` file to manage them securely. Follow these steps:
+
+1. Create a file named `.env` in the root directory of your project.
+2. Add your API keys and voice ID to the `.env` file in the following format:
+
+   ```
+   OPENAI_API_KEY=your_openai_api_key
+   ELEVENLABS_API_KEY=your_elevenlabs_api_key
+   ELEVENLABS_VOICE_ID=your_elevenlabs_voice_id
+   ```
+
+   Replace `your_openai_api_key`, `your_elevenlabs_api_key`, and `your_elevenlabs_voice_id` with your actual keys and ID.
+
+3. The python-dotenv package (already included in `requirements.txt`) will load these variables automatically.
+
+**Note:** Ensure that `.env` is listed in your `.gitignore` file to keep your API keys secure.
 Make a new voice in Eleven and get the voice id of that voice using their [get voices](https://elevenlabs.io/docs/api-reference/voices) API, or by clicking the flask icon next to the voice in the VoiceLab tab.
 
 ```

capture.py

@@ -24,21 +24,21 @@ time.sleep(2)
 while True:
     ret, frame = cap.read()
     if ret:
-        # Convert the frame to a PIL image
-        pil_img = Image.fromarray(cv2.cvtColor(frame, cv2.COLOR_BGR2RGB))
-
-        # Resize the image
-        max_size = 250
-        ratio = max_size / max(pil_img.size)
-        new_size = tuple([int(x*ratio) for x in pil_img.size])
-        resized_img = pil_img.resize(new_size, Image.LANCZOS)
-
-        # Convert the PIL image back to an OpenCV image
-        frame = cv2.cvtColor(np.array(resized_img), cv2.COLOR_RGB2BGR)
+        # Resize the image before saving to improve performance
+        max_size = 400
+        height, width = frame.shape[:2]
+        if height > width:
+            new_height = max_size
+            new_width = int((max_size / height) * width)
+        else:
+            new_width = max_size
+            new_height = int((max_size / width) * height)
+        
+        frame = cv2.resize(frame, (new_width, new_height))
 
         # Save the frame as an image file
         print("📸 Say cheese! Saving frame.")
-        path = f"{folder}/frame.jpg"
+        path = os.path.join(frames_dir, "frame.jpg")
         cv2.imwrite(path, frame)
     else:
         print("Failed to capture image")

narrator.py

@@ -1,17 +1,33 @@
 import os
+from dotenv import load_dotenv
 from openai import OpenAI
 import base64
-import json
+# import json
 import time
-import simpleaudio as sa
+# import simpleaudio as sa
 import errno
-from elevenlabs import generate, play, set_api_key, voices
+from elevenlabs import play, Voice
+from elevenlabs.client import ElevenLabs
 
-client = OpenAI()
+# Load environment variables from a .env file
+load_dotenv()
 
-set_api_key(os.environ.get("ELEVENLABS_API_KEY"))
+# Initialize OpenAI and ElevenLabs clients
+clientOA = OpenAI()
+clientEL = ElevenLabs(
+  api_key=os.environ.get("ELEVENLABS_API_KEY")
+)
 
 def encode_image(image_path):
+    """
+    Encodes an image to base64.
+    
+    Args:
+        image_path (str): The path to the image file.
+        
+    Returns:
+        str: Base64 encoded string of the image.
+    """
     while True:
         try:
             with open(image_path, "rb") as image_file:
@@ -23,80 +39,111 @@ def encode_image(image_path):
             # File is being written to, wait a bit and retry
             time.sleep(0.1)
 
-
 def play_audio(text):
-    audio = generate(text, voice=os.environ.get("ELEVENLABS_VOICE_ID"))
+    """
+    Generates and plays audio from text using ElevenLabs.
+    
+    Args:
+        text (str): The text to be converted to speech.
+    """
+    # Generate audio from text
+    audio_generator = clientEL.generate(text=text, voice=Voice(voice_id=os.environ.get("ELEVENLABS_VOICE_ID")))
 
+    # Create a unique directory for storing the audio file
     unique_id = base64.urlsafe_b64encode(os.urandom(30)).decode("utf-8").rstrip("=")
     dir_path = os.path.join("narration", unique_id)
     os.makedirs(dir_path, exist_ok=True)
     file_path = os.path.join(dir_path, "audio.wav")
 
+    # Gather audio data from generator
+    audio_bytes = b''.join(audio_generator)
+
+    # Save audio to file
     with open(file_path, "wb") as f:
-        f.write(audio)
-
-    play(audio)
+        f.write(audio_bytes)
 
+    # Play the generated audio
+    play(audio_bytes)
 
 def generate_new_line(base64_image):
+    """
+    Generates a new line of messages for the OpenAI API call.
+    
+    Args:
+        base64_image (str): Base64 encoded string of the image.
+        
+    Returns:
+        list: A list of messages to be sent to the OpenAI API.
+    """
     return [
         {
             "role": "user",
             "content": [
-                {"type": "text", "text": "Describe this image"},
+                {"type": "text", "text": "Describe this image as if you are Sir David Attenborough narrating a nature documentary about homo sapiens."},
                 {
                     "type": "image_url",
-                    "image_url": f"data:image/jpeg;base64,{base64_image}",
+                    "image_url": {
+                        "url": f"data:image/jpeg;base64,{base64_image}",
+                    },
                 },
             ],
         },
     ]
 
-
 def analyze_image(base64_image, script):
-    response = client.chat.completions.create(
-        model="gpt-4-vision-preview",
+    """
+    Analyzes an image using OpenAI's language model.
+    
+    Args:
+        base64_image (str): Base64 encoded string of the image.
+        script (list): List of previous messages to maintain context.
+        
+    Returns:
+        str: The response text from OpenAI.
+    """
+    response = clientOA.chat.completions.create(
+        model="gpt-4o-mini",
         messages=[
             {
                 "role": "system",
                 "content": """
                 You are Sir David Attenborough. Narrate the picture of the human as if it is a nature documentary.
-                Make it snarky and funny. Don't repeat yourself. Make it short. If I do anything remotely interesting, make a big deal about it!
+                Be accurate, snarky, and funny. Describe what the human is actually doing. Make it short and concise, within 3 sentences. If the human is doing something remotely interesting, make a big deal about it!
                 """,
             },
         ]
         + script
         + generate_new_line(base64_image),
-        max_tokens=500,
+        max_tokens=150,
+        temperature=0.7,
     )
     response_text = response.choices[0].message.content
     return response_text
 
-
 def main():
     script = []
 
     while True:
-        # path to your image
+        # Path to your image
         image_path = os.path.join(os.getcwd(), "./frames/frame.jpg")
 
-        # getting the base64 encoding
+        # Get the base64 encoding of the image
         base64_image = encode_image(image_path)
 
-        # analyze posture
+        # Analyze the image and generate a narration
         print("👀 David is watching...")
         analysis = analyze_image(base64_image, script=script)
 
+        # Print and play the narration
         print("🎙️ David says:")
         print(analysis)
-
         play_audio(analysis)
 
-        script = script + [{"role": "assistant", "content": analysis}]
-
-        # wait for 5 seconds
-        time.sleep(5)
+        # Append the analysis to the script for context in future requests
+        script.append({"role": "assistant", "content": analysis})
 
+        # wait for 3 seconds
+        time.sleep(3)
 
 if __name__ == "__main__":
-    main()
+    main()

requirements.txt

@@ -6,7 +6,7 @@ certifi==2023.7.22
 charset-normalizer==3.3.2
 decorator==5.1.1
 distro==1.8.0
-elevenlabs==0.2.26
+elevenlabs==1.5.0
 exceptiongroup==1.1.3
 executing==2.0.1
 h11==0.14.0
@@ -28,6 +28,7 @@ pure-eval==0.2.2
 pydantic==2.4.2
 pydantic_core==2.10.1
 Pygments==2.16.1
+python-dotenv==1.0.0
 requests==2.31.0
 simpleaudio==1.0.4
 six==1.16.0