Update app.py

app.py

@@ -1,9 +1,41 @@
+from share_btn import community_icon_html, loading_icon_html, share_js
+
+import os, subprocess
+import torch
+
+# def setup():
+#     install_cmds = [
+#         ['pip', 'install', 'ftfy', 'gradio', 'regex', 'tqdm', 'transformers==4.21.2', 'timm', 'fairscale', 'requests'],
+#         ['pip', 'install', 'open_clip_torch'],
+#         ['pip', 'install', '-e', 'git+https://github.com/pharmapsychotic/BLIP.git@lib#egg=blip'],
+#         ['git', 'clone', '-b', 'open-clip', 'https://github.com/pharmapsychotic/clip-interrogator.git']
+#     ]
+#     for cmd in install_cmds:
+#         print(subprocess.run(cmd, stdout=subprocess.PIPE).stdout.decode('utf-8'))
+
+# setup()
+
+# download cache files
+# print("Download preprocessed cache files...")
+# CACHE_URLS = [
+#     'https://huggingface.co/pharma/ci-preprocess/resolve/main/ViT-H-14_laion2b_s32b_b79k_artists.pkl',
+#     'https://huggingface.co/pharma/ci-preprocess/resolve/main/ViT-H-14_laion2b_s32b_b79k_flavors.pkl',
+#     'https://huggingface.co/pharma/ci-preprocess/resolve/main/ViT-H-14_laion2b_s32b_b79k_mediums.pkl',
+#     'https://huggingface.co/pharma/ci-preprocess/resolve/main/ViT-H-14_laion2b_s32b_b79k_movements.pkl',
+#     'https://huggingface.co/pharma/ci-preprocess/resolve/main/ViT-H-14_laion2b_s32b_b79k_trendings.pkl',
+# ]
+# os.makedirs('cache', exist_ok=True)
+# for url in CACHE_URLS:
+#     print(subprocess.run(['wget', url, '-P', 'cache'], stdout=subprocess.PIPE).stdout.decode('utf-8'))
+
+import sys
+sys.path.append('src/blip')
+sys.path.append('clip-interrogator')
+
 import gradio as gr
 from clip_interrogator import Config, Interrogator
-import torch
+import io 
 from PIL import Image
-import io
-
 config = Config()
 config.device = 'cuda' if torch.cuda.is_available() else 'cpu'
 config.blip_offload = False if torch.cuda.is_available() else True
@@ -13,31 +45,139 @@ config.blip_num_beams = 64
 
 ci = Interrogator(config)
 
-def inference(input_image, mode, best_max_flavors):
-    image_bytes = input_image.read()
-    image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
+def inference(input_images, mode, best_max_flavors):
+    # Process each image in the list and generate prompt results
+    prompt_results = []
+    for image_bytes in input_images:
+        image = Image.open(io.BytesIO(image_bytes)).convert('RGB')
+        if mode == 'best':
+            prompt_result = ci.interrogate(image, max_flavors=int(best_max_flavors))
+        elif mode == 'classic':
+            prompt_result = ci.interrogate_classic(image)
+        else:
+            prompt_result = ci.interrogate_fast(image)
+        prompt_results.append(prompt_result)  # Append text prompt result
+    return prompt_results
+
+
+title = """
+    <div style="text-align: center; max-width: 500px; margin: 0 auto;">
+        <div
+        style="
+            display: inline-flex;
+            align-items: center;
+            gap: 0.8rem;
+            font-size: 1.75rem;
+            margin-bottom: 10px;
+        "
+        >
+        <h1 style="font-weight: 600; margin-bottom: 7px;">
+            CLIP Interrogator 2.1
+        </h1>
+        </div>
+        <p style="margin-bottom: 10px;font-size: 94%;font-weight: 100;line-height: 1.5em;">
+        Want to figure out what a good prompt might be to create new images like an existing one? 
+        <br />The CLIP Interrogator is here to get you answers!
+        <br />This version is specialized for producing nice prompts for use with Stable Diffusion 2.0 using the ViT-H-14 OpenCLIP model!
+        </p>
+    </div>
+"""
+
+article = """
+<div style="text-align: center; max-width: 500px; margin: 0 auto;font-size: 94%;">
     
-    if mode == 'best':
-        prompt_result = ci.interrogate(image, max_flavors=int(best_max_flavors))
-    elif mode == 'classic':
-        prompt_result = ci.interrogate_classic(image)
-    else:
-        prompt_result = ci.interrogate_fast(image)
-        
-    return prompt_result
+    <p>
+    Server busy? You can also run on <a href="https://colab.research.google.com/github/pharmapsychotic/clip-interrogator/blob/open-clip/clip_interrogator.ipynb">Google Colab</a>
+    </p>
+    <p>
+    Has this been helpful to you? Follow Pharma on twitter 
+    <a href="https://twitter.com/pharmapsychotic">@pharmapsychotic</a> 
+    and check out more tools at his
+    <a href="https://pharmapsychotic.com/tools.html">Ai generative art tools list</a>
+    </p>
+</div>
+"""
+
+css = '''
+#col-container {width: width: 80%;; margin-left: auto; margin-right: auto;}
+a {text-decoration-line: underline; font-weight: 600;}
+.animate-spin {
+    animation: spin 1s linear infinite;
+}
+@keyframes spin {
+    from {
+        transform: rotate(0deg);
+    }
+    to {
+        transform: rotate(360deg);
+    }
+}
+#share-btn-container {
+    display: flex; padding-left: 0.5rem !important; padding-right: 0.5rem !important; background-color: #000000; justify-content: center; align-items: center; border-radius: 9999px !important; width: 13rem;
+}
+#share-btn {
+    all: initial; color: #ffffff;font-weight: 600; cursor:pointer; font-family: 'IBM Plex Sans', sans-serif; margin-left: 0.5rem !important; padding-top: 0.25rem !important; padding-bottom: 0.25rem !important;
+}
+#share-btn * {
+    all: unset;
+}
+#share-btn-container div:nth-child(-n+2){
+    width: auto !important;
+    min-height: 0px !important;
+}
+#share-btn-container .wrap {
+    display: none !important;
+}
+#gallery .caption-label {
+    font-size: 15px !important;
+    right: 0 !important;
+    max-width: 100% !important;
+    text-overflow: clip !important;
+    white-space: normal !important;
+    overflow: auto !important;
+    height: 20% !important;
+}
+
+#gallery .caption {
+    padding: var(--size-2) var(--size-3) !important;
+    text-overflow: clip !important;
+    white-space: normal !important; /* Allows the text to wrap */
+    color: var(--block-label-text-color) !important;
+    font-weight: var(--weight-semibold) !important;
+    text-align: center !important;
+    height: 100% !important;
+    font-size: 17px !important;
+}
+
+'''
 
 with gr.Blocks() as block:
     with gr.Column(elem_id="col-container"):
-        input_image = gr.Image(label="Upload Image")  # Using gr.Image for image upload
-        mode_input = gr.Radio(['classic', 'fast', 'best'], label='Select mode', default='fast')
-        flavor_input = gr.Slider(minimum=2, maximum=24, step=2, default=4, label='Best mode max flavors')
-        submit_btn = gr.Button("Submit")
-        output_text = gr.Textbox(label="Output", elem_id="output-txt")  # Output Textbox
-
-        def process_image():
-            prompt_result = inference(input_image, mode_input, flavor_input)
-            output_text = prompt_result
+        gr.HTML(title)
+
         
-        submit_btn.click(process_image)
+        input_images = gr.Files(label="Inputs", file_count="multiple", type='file', elem_id='inputs')
+
+# Create a function to display the uploaded images
+def display_uploaded_images(input_images):
+    image_display = []  # List to store images for display
+
+    for image_file in input_images:
+        # Get the path or value of the uploaded image
+        image_path = image_file.name if isinstance(image_file, gr.File) else None
+        image_value = image_file.getvalue() if isinstance(image_file, gr.File) else None
+        
+        if image_path:
+            # If using image path
+            image_display.append(gr.Image(image_path))
+        elif image_value:
+            # If using image value
+            image_display.append(gr.Image(image_bytes=image_value))
+    
+    return image_display
+
+# Display the uploaded images using gr.image
+uploaded_images_display = display_uploaded_images(input_images)
 
-block.queue(max_size=32, concurrency_count=10).launch(show_api=False)
+# Show the uploaded images in the Gradio interface
+gr.Interface(fn=inference, inputs=input_images, outputs=output_text, title="CLIP Interrogator 2.1", layout="vertical", examples=None).launch()