setting up the generation and processing of images and 3d assets

get_blender.py

@@ -1,63 +0,0 @@
-import os
-import shutil
-import subprocess
-import sys
-import tarfile
-import urllib.request
-
-def download_blender(url, destination):
-    urllib.request.urlretrieve(url, destination)
-
-def extract_archive(archive_file, destination):
-    with tarfile.open(archive_file, 'r:xz') as tar:
-        tar.extractall(destination)
-
-def remove_file(file_path):
-    os.remove(file_path)
-
-def move_folder(source, destination):
-    shutil.move(source, destination)
-
-def create_symbolic_link(source, target):
-    os.symlink(source, target)
-
-def install_packages(packages):
-    subprocess.run(['sudo', 'apt-get', 'update'])
-    subprocess.run(['sudo', 'apt-get', 'install'] + packages + ['-y'])
-
-def main():
-    blender_url = 'https://download.blender.org/release/Blender4.0/blender-4.0.2-linux-x64.tar.xz'
-    archive_file = 'blender-4.0.2-linux-x64.tar.xz'
-    extracted_folder = 'blender-4.0.2-linux-x64'
-    destination_folder = '/opt/blender-4.0.2'
-    symbolic_link = '/usr/local/bin/blender'
-    packages_to_install = ['libxxf86vm1', 'libgl1-mesa-glx', 'libegl-mesa0', 'libegl1']
-
-    # Download Blender
-    print("Downloading Blender...")
-    download_blender(blender_url, archive_file)
-
-    # Extract the downloaded archive
-    print("Extracting Blender archive...")
-    extract_archive(archive_file, '.')
-
-    # Remove the downloaded archive
-    print("Removing downloaded archive...")
-    remove_file(archive_file)
-
-    # Move the extracted folder to the desired location
-    print("Moving Blender folder...")
-    move_folder(extracted_folder, destination_folder)
-
-    # Create a symbolic link to the Blender executable
-    print("Creating symbolic link to Blender executable...")
-    create_symbolic_link(os.path.join(destination_folder, 'blender'), symbolic_link)
-
-    # Install required packages
-    print("Installing required packages...")
-    install_packages(packages_to_install)
-
-    print("Installation complete.")
-
-if __name__ == "__main__":
-    main()

main.py

@@ -0,0 +1,62 @@
+from langhchain_generate_components import maker_wikipedia_chain
+from utils import (
+    save_file, convert_obj_to_stl,
+    change_file_extension,
+)
+from mesh_utils import generate_mesh_images
+from gradio_client import Client
+
+
+def main():  
+    # the object to be generated
+    query = "A Microscope"
+    
+    # using a retriever we generat a list of Components
+    output = maker_wikipedia_chain.invoke(query)
+    
+    # the first item 
+    shap_e_sample = output['Material'][0]
+    
+    client = Client("hysts/Shap-E")
+    result = client.predict(
+            shap_e_sample,	# str  in 'Prompt' Textbox component
+            1621396601,	# float (numeric value between 0 and 2147483647) in 'Seed' Slider component
+            15,	# float (numeric value between 1 and 20) in 'Guidance scale' Slider component
+            64,	# float (numeric value between 2 and 100) in 'Number of inference steps' Slider component
+            api_name="/text-to-3d"
+    )
+    
+    saved_file_name = "sample.glb" 
+    # save to local machine
+    save_file(result,saved_file_name)
+    
+    stl_file_location = change_file_extension(
+        saved_file_name,
+        ".stl"
+    )
+    
+    # convert into a stl without the texture
+    # as it is easiest to handle
+    convert_obj_to_stl(
+        result, 
+        stl_file_location,
+        )
+    
+    # Need to generate screenshot for the item
+    
+    viewing_angles = [(30, 45), (60, 90), (45, 135)]
+    
+    generate_mesh_images(
+        stl_file_location, 
+        viewing_angles
+        )
+    
+    # These screenshots need to be given to GPT-V
+    # for feedback
+    
+    print(result)
+
+    x = 0
+    
+if __name__ == "__main__":
+    main()

setup.sh

@@ -1,20 +0,0 @@
-# download belnder 
-wget https://download.blender.org/release/Blender4.0/blender-4.0.2-linux-x64.tar.xz
-
-# Extract the downloaded archive:
-
-tar xf blender-4.0.2-linux-x64.tar.xz
-
-rm -fr blender-4.0.2-linux-x64.tar.xz
-
-# Move the extracted folder to the desired location (for example, /opt):
-sudo mv blender-4.0.2-linux-x64 /opt/blender-4.0.2
-
-# Create a symbolic link to the Blender executable:
-sudo ln -s /opt/blender-4.0.2/blender /usr/local/bin/blender
-
-sudo apt-get update
-sudo apt-get install libxxf86vm1 -y
-sudo apt-get install libgl1-mesa-glx -y
-sudo apt-get install libegl-mesa0 -y
-sudo apt-get install libegl1 -y

utils.py

@@ -0,0 +1,44 @@
+import shutil
+import trimesh
+import os
+
+def save_file(input_file, output_file):
+    """
+    Copy a file from input location to output location.
+
+    Args:
+    input_file (str): Path to the input file.
+    output_file (str): Path to the output file.
+
+    Returns:
+    bool: True if the file is successfully saved, False otherwise.
+    """
+    try:
+        shutil.copy(input_file, output_file)
+        return True
+    except Exception as e:
+        print(f"Error: {e}")
+        return False
+    
+
+def convert_obj_to_stl(input_file: str, output_file: str):
+    # Load the OBJ file
+    mesh = trimesh.load(input_file)
+
+    # Export as STL
+    mesh.export(output_file)
+    
+def change_file_extension(file_path: str, new_extension: str) -> str:
+    """
+    Change the extension of a file path.
+
+    Args:
+        file_path (str): The original file path.
+        new_extension (str): The new file extension (without the dot).
+
+    Returns:
+        str: The modified file path with the new extension.
+    """
+    base_path, _ = os.path.splitext(file_path)
+    new_file_path = base_path + '.' + new_extension
+    return new_file_path

vision_model.py

@@ -0,0 +1,84 @@
+import base64
+from openai import OpenAI
+from typing import List, Dict, Any
+from dotenv import load_dotenv
+import os
+
+load_dotenv()
+
+# source
+# https://platform.openai.com/docs/guides/vision?lang=python
+def analyze_images(
+    images: List[str],
+    prompt: str,
+    # api_key: str,
+    model: str = "gpt-4-vision-preview",
+    max_tokens: int = 300
+    ) -> Dict[str, Any]:
+    """
+    Analyze multiple images using OpenAI's vision model.
+
+    Args:
+        images (List[str]): List of URLs and/or local paths to the image files.
+        prompt (str): Prompt message for the AI model.
+        api_key (str): Your OpenAI API key.
+        model (str, optional): Name of the vision model to use. Defaults to "gpt-4-vision-preview".
+        max_tokens (int, optional): Maximum number of tokens for the response. Defaults to 300.
+
+    Returns:
+        dict: JSON response from the API.
+    """
+    client = OpenAI()
+    messages = [{
+        "role": "user",
+        "content": [{"type": "text", "text": prompt}]
+    }]
+
+    for image in images:
+        if image.startswith("http://") or image.startswith("https://"):
+            # Image is a URL
+            messages.append({
+                "role": "user",
+                "content": [{"type": "image_url", "image_url": {"url": image}}]
+            })
+        else:
+            # Image is a local path
+            with open(image, "rb") as image_file:
+                base64_image = base64.b64encode(image_file.read()).decode('utf-8')
+            messages.append({
+                "role": "user",
+                "content": [{"type": "image_url", "image_url": {"url": f"data:image/jpeg;base64,{base64_image}"}}]
+            })
+
+    response = client.chat.completions.create(
+        model=model,
+        messages=messages,
+        max_tokens=max_tokens
+    )
+    return response.choices[0]
+
+def main():
+    api_key = os.getenv("OPENAI_API_KEY")
+    images = [
+        "/workspaces/Maker-Tech-Tree/mesh_1.png",
+        "/workspaces/Maker-Tech-Tree/mesh_2.png",
+        "/workspaces/Maker-Tech-Tree/mesh_3.png",
+    ]
+    prompt = "I am creating an 3d model of a Glass lenses for refracting light,\
+        using a text-to-3d model\
+        Do these images look correct?\
+        If not please make a suggesttion on how to improve the text input\
+        As this response will be used in a pipeline please only output a new \
+        potential prompt or output nothing, \
+        Please keep the prompt to 5 25 words to not confuse the model"
+    
+    response = analyze_images(
+        images, 
+        prompt, 
+        # api_key,
+        )
+    
+    print(response)
+
+if __name__ == "__main__":
+    main()