Rename folder + remove coloring on omniparseragent printouts

computer_use_demo/gradio/.gitignore

@@ -0,0 +1 @@
+tmp/

computer_use_demo/gradio/colorful_text.py

@@ -1,17 +0,0 @@
-"""
-Define some colorful stuffs for better visualization in the chat.
-"""
-
-# Define the RGB colors for each letter
-colors = {
-    'S': 'rgb(106, 158, 210)',
-    'h': 'rgb(111, 163, 82)',
-    'o': 'rgb(209, 100, 94)',
-    'w': 'rgb(238, 171, 106)',
-    'U': 'rgb(0, 0, 0)',  
-    'I': 'rgb(0, 0, 0)',  
-}
-
-colorful_text_vlm = "**OmniParser Agent**"
-
-colorful_text_user = "**User**"

computer_use_demo/omniparserserver/omniparserserver.py

@@ -0,0 +1,51 @@
+'''
+python -m omniparserserver --som_model_path ../../weights/icon_detect_v1_5/model_v1_5.pt --caption_model_name florence2 --caption_model_path ../../weights/icon_caption_florence --device cuda --BOX_TRESHOLD 0.05
+'''
+
+import sys
+import os
+import time
+from fastapi import FastAPI
+from pydantic import BaseModel
+import argparse
+import uvicorn
+root_dir = os.path.dirname(os.path.dirname(os.path.dirname(os.path.abspath(__file__))))
+sys.path.append(root_dir)
+from util.omniparser import Omniparser
+
+def parse_arguments():
+    parser = argparse.ArgumentParser(description='Omniparser API')
+    parser.add_argument('--som_model_path', type=str, default='../../weights/icon_detect_v1_5/model_v1_5.pt', help='Path to the som model')
+    parser.add_argument('--caption_model_name', type=str, default='florence2', help='Name of the caption model')
+    parser.add_argument('--caption_model_path', type=str, default='../../weights/icon_caption_florence', help='Path to the caption model')
+    parser.add_argument('--device', type=str, default='cpu', help='Device to run the model')
+    parser.add_argument('--BOX_TRESHOLD', type=float, default=0.05, help='Threshold for box detection')
+    parser.add_argument('--host', type=str, default='0.0.0.0', help='Host for the API')
+    parser.add_argument('--port', type=int, default=8000, help='Port for the API')
+    args = parser.parse_args()
+    return args
+
+args = parse_arguments()
+config = vars(args)
+
+app = FastAPI()
+omniparser = Omniparser(config)
+
+class ParseRequest(BaseModel):
+    base64_image: str
+
+@app.post("/parse/")
+async def parse(parse_request: ParseRequest):
+    print('start parsing...')
+    start = time.time()
+    dino_labled_img, parsed_content_list = omniparser.parse(parse_request.base64_image)
+    latency = time.time() - start
+    print('time:', latency)
+    return {"som_image_base64": dino_labled_img, "parsed_content_list": parsed_content_list, 'latency': latency}
+
+@app.get("/probe/")
+async def root():
+    return {"message": "Omniparser API ready"}
+
+if __name__ == "__main__":
+    uvicorn.run("remote_request:app", host=args.host, port=args.port, reload=True)