Artyom Boyko commited on
Commit
00b941a
Β·
1 Parent(s): ea02e85

Made a header for the GUI

Browse files
Files changed (2) hide show
  1. app_client/app_client.py +1 -1
  2. app_srv/app_srv.py +13 -4
app_client/app_client.py CHANGED
@@ -19,7 +19,7 @@ A parameter of int type is set like this - 'parameter': 720. A parameter of the
19
 
20
  try:
21
  mcp_client = MCPClient(
22
- {"url": "https://artyomboyko-aura-ai-scan.hf.space/gradio_api/mcp/sse"}
23
  )
24
 
25
  tools = mcp_client.get_tools()
 
19
 
20
  try:
21
  mcp_client = MCPClient(
22
+ {"url": "https://artyomboyko-test-mcp-2.hf.space/gradio_api/mcp/sse"}
23
  )
24
 
25
  tools = mcp_client.get_tools()
app_srv/app_srv.py CHANGED
@@ -15,10 +15,19 @@ from model_api import get_device_and_dtype
15
 
16
  device, dtype = get_device_and_dtype()
17
 
18
- gui_header_element = """# πŸŽ₯ YouTube Comprehensive Video Analysis Tool
19
- YouTube AI Video Analyze - get frame-by-frame descriptions with timestamps and audio transcription.
20
- Since most MCP servers operate without a GUI, this version includes several tools to help you view JSON results, both raw and visually.
21
- This space can be run locally as it is a Docker container.
 
 
 
 
 
 
 
 
 
22
  """
23
 
24
  DEFAULT_PROMPT = """You are an expert at analyzing video, so pay close attention. Your main goal is to analyze the frame and find information in it to answer the MAIN QUESTION. Pay attention to details.
 
15
 
16
  device, dtype = get_device_and_dtype()
17
 
18
+ gui_header_element = """# πŸš€ Aura AI Scan: Deep Analysis of YouTube Videos
19
+
20
+ ([alpha version](https://en.wikipedia.org/wiki/Software_release_life_cycle))
21
+
22
+ Aura AI Scan is an MCP tool designed to deeply analyse YouTube videos, providing a frame-by-frame description with timestamps and full transcription of the audio track.
23
+
24
+ The following technologies were used in the implementation:
25
+ - [VSCode devcontainer](https://code.visualstudio.com/docs/devcontainers/containers), to simplify the development process
26
+ - [Docker](https://huggingface.co/docs/hub/spaces-sdks-docker), to simplify local deployment. This space is also implemented as [HuggingFace Docker Space](https://huggingface.co/docs/hub/spaces-sdks-docker).
27
+ - In the server part a full-fledged GUI is implemented allowing to view the final result in three different variants (visually, baudially and in JSON).
28
+ - Flexible settings (hint used during analysis, quality of analysed video, time interval between frames).
29
+
30
+
31
  """
32
 
33
  DEFAULT_PROMPT = """You are an expert at analyzing video, so pay close attention. Your main goal is to analyze the frame and find information in it to answer the MAIN QUESTION. Pay attention to details.