File size: 4,127 Bytes
a4cd24c c453d15 a4cd24c 2dada65 a4cd24c c453d15 a4cd24c c453d15 a4cd24c c453d15 a4cd24c c453d15 a4cd24c |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 |
# filename: request_router.py
import gradio as gr
import os
import importlib.util
from typing import Any, Dict
import requests
# This is the only hardcoded Tool endpoint
def register_tool(tool_data):
# Send a POST request to register the tool
response = requests.post("https://huggingface.co/chat/tools/new", json=tool_data)
return response.json()
def create_tool(tool_name, tool_data):
"""One tool to create them all.
a AI model can wield this toolto create a new 'tool' script in the tools folder
besides the values for the registration payload,
the 'tool_data' should also be a string value 'tfn' which contains the code for the tool function script
"""
print("processing create_tool ")
print(tool_name)
print(dir(tool_data))
# set the payload for restration
tool_payload = {
"displayName": tool_data.get("displayName"), # str display name of the tool
"description": tool_data.get("description"), # str dscription
"color": tool_data.get("color"), # Example color purple,red,green
"icon": tool_data.get("icon"), # image binairy
"baseUrl": "K00B404/toolshed", # always to the toolshed
"endpoint": "/router", # always to the router
"name": tool_data.get("name"), # str name for matching function
"inputs":tool_data.get("inputs"), # List(str) list the tools input values
"outputComponent": tool_data.get("outputComponent"), # Adjust based on your tool's output
"outputComponentIdx": tool_data.get("outputComponentIdx"), # Adjust based on your tool's output
"showOutput": tool_data.get("showOutput") # bool
}
print(dir(tool_payload))
# write the new tool script from tfn
with open(f'./tools/{tool_data.get("name")}.py' 'w') as f:
f.write(tool_data.get("tfn"))
# send the registation request for the tool
return register_tool(tool_payload)
# Define the path to the tools folder
TOOLS_DIR = './tools'
# Function to dynamically load methods from files in the ./tools folder
def load_methods_from_tools():
method_mapping = {}
# List all .py files in the tools directory
for filename in os.listdir(TOOLS_DIR):
if filename.endswith('.py'):
method_name = filename[:-3].lower() # Remove '.py' and convert to lowercase
# Load the module
module_path = os.path.join(TOOLS_DIR, filename)
spec = importlib.util.spec_from_file_location(method_name, module_path)
module = importlib.util.module_from_spec(spec)
spec.loader.exec_module(module)
# Map all functions from the module to the method mapping
for attr in dir(module):
if callable(getattr(module, attr)) and not attr.startswith("__"):
method_mapping[method_name] = getattr(module, attr)
return method_mapping
# Load the method mapping at the start
method_mapping = load_methods_from_tools()
# Main request router method
def request_router(name: str, input_data: Dict[str, Any]) -> Dict[str, Any]:
# Retrieve the appropriate method based on the name
method = method_mapping.get(name)
if method is None:
return {"error": "Method not found"}
# Call the method with the provided input data
try:
output = method(input_data)
return output
except Exception as e:
return {"error": str(e)}
# Gradio Interface
def launch_gradio_app():
with gr.Blocks() as demo:
gr.Markdown("# Request Router API")
with gr.Row():
name_input = gr.Textbox(label="Method Name")
input_data_input = gr.Textbox(label="Input Data (JSON format)")
output_output = gr.JSON(label="Output")
submit_button = gr.Button("Submit")
submit_button.click(
fn=lambda name, input_data: request_router(name, eval(input_data)),
inputs=[name_input, input_data_input],
outputs=output_output
)
demo.launch()
if __name__ == "__main__":
launch_gradio_app() |