Spaces:
Sleeping
Sleeping
File size: 4,808 Bytes
a187358 |
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 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 |
import gradio as gr
import re
from transformers import pipeline
from query import get_liquefaction_at_point
# Load Qwen model pipeline
pipe = pipeline("text-generation", model="Qwen/Qwen3-0.6B")
def extract_coordinates(user_input):
"""Extract coordinates from user input using regex"""
# Look for patterns like "1511300, 5266130" or "x: 1511300, y: 5266130"
patterns = [
r'(\d{6,7})\s*,\s*(\d{6,7})', # 1511300, 5266130
r'x:\s*(\d{6,7})\s*,\s*y:\s*(\d{6,7})', # x: 1511300, y: 5266130
r'coordinates?\s*[:\-]?\s*(\d{6,7})\s*,\s*(\d{6,7})', # coordinates: 1511300, 5266130
]
for pattern in patterns:
match = re.search(pattern, user_input, re.IGNORECASE)
if match:
x = int(match.group(1))
y = int(match.group(2))
return x, y
return None, None
def agent(user_input):
# Try regex first (fast and reliable)
x, y = extract_coordinates(user_input)
# If regex fails, try LLM as fallback
if x is None or y is None:
try:
llm_response = pipe(
f"Extract NZTM coordinates from this text. Return only the numbers separated by comma: {user_input}",
max_new_tokens=20,
do_sample=False,
temperature=0.1
)[0]["generated_text"]
# Parse LLM response for coordinates
coords_match = re.search(r'(\d{6,7})\s*,\s*(\d{6,7})', llm_response)
if coords_match:
x, y = int(coords_match.group(1)), int(coords_match.group(2))
else:
return "Could not extract coordinates. Please provide coordinates in format like '1511300, 5266130' or 'x: 1511300, y: 5266130'"
except:
return "Could not extract coordinates. Please provide coordinates in format like '1511300, 5266130' or 'x: 1511300, y: 5266130'"
try:
result = get_liquefaction_at_point(x, y)
return f"Coordinates: ({x}, {y})\nLiquefaction Susceptibility: {result}"
except Exception as e:
return f"Error getting liquefaction data: {e}"
def predict(prompt):
# Use a simpler approach - just pass the prompt directly without messages format
completion = pipe(prompt, max_new_tokens=100, do_sample=True, temperature=0.7)[0]["generated_text"]
# Remove the original prompt from the response
if prompt in completion:
response = completion[len(prompt):].strip()
else:
response = completion
# Clean up any formatting issues
response = response.replace("\\n", "\n").replace("\\'", "'")
return response
# Create a tabbed interface
with gr.Blocks(title="Geo Agent Demo") as iface:
gr.Markdown("# Geo Agent Demo")
gr.Markdown("Enter NZTM coordinates to get liquefaction susceptibility data. The agent uses a Hugging Face model to extract coordinates and queries the Canterbury Liquefaction Susceptibility ArcGIS MapServer.")
with gr.Tab("Coordinate Analysis"):
coord_input = gr.Textbox(lines=2, label="Enter coordinates (e.g., '1511300, 5266130' or 'x: 1511300, y: 5266130')")
coord_output = gr.Textbox(label="Liquefaction Susceptibility Result", lines=4)
coord_button = gr.Button("Analyze Coordinates", variant="primary")
coord_button.click(fn=agent, inputs=coord_input, outputs=coord_output)
# Add example inputs
gr.Examples(
examples=[
["1511300, 5266130"],
["x: 1511300, y: 5266130"],
["What's the liquefaction at coordinates 1511300, 5266130?"],
["Can you check 1511300, 5266130 for me?"]
],
inputs=coord_input
)
with gr.Tab("Chat Assistant"):
chat_input = gr.Textbox(label="Type your message here:", placeholder="Hello, how can you help me?", lines=2)
chat_output = gr.Textbox(label="AI Response", lines=6)
chat_button = gr.Button("Send Message", variant="primary")
chat_button.click(fn=predict, inputs=chat_input, outputs=chat_output)
# Add example chat messages
gr.Examples(
examples=[
["Hello, what can you help me with?"],
["What is liquefaction susceptibility?"],
["How do I use this application?"],
["Tell me a joke"]
],
inputs=chat_input
)
if __name__ == "__main__":
iface.launch(
server_name="0.0.0.0", # Allow external connections
server_port=7860, # Standard Gradio port
share=False, # Set to True if you want a public link
show_error=True, # Show detailed errors
quiet=False # Show startup messages
) |