Spaces:
Sleeping
Sleeping
File size: 3,130 Bytes
0ef40b0 f2e91a8 0ef40b0 |
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 |
import gradio as gr
from PIL import Image
import os
from dotenv import load_dotenv
from llama_index.multi_modal_llms.gemini import GeminiMultiModal
from llama_index.core.program import MultiModalLLMCompletionProgram
from pydantic import BaseModel
from llama_index.core.output_parsers import PydanticOutputParser
from llama_index.core.tools import FunctionTool
from llama_index.core import SimpleDirectoryReader
from duckduckgo_search import DDGS
from llama_index.core.agent import ReActAgent
from llama_index.llms.gemini import Gemini
load_dotenv()
token = os.getenv("GEMINI_API_KEY")
class PlantHealth(BaseModel):
plant_type: str
condition: str
symptoms: str
confidence: float
prompt_template_str = """Analyze the plant image and return:
- Plant type (e.g., tomato, rose)
- Condition (e.g., healthy, diseased)
- Symptoms (e.g., yellow spots, wilting)
- Confidence score (0.0 to 1.0)
"""
def analyze_plant_image(image_path: str) -> dict:
gemini_llm = GeminiMultiModal(model_name="gemini-2.0-flash")
llm_program = MultiModalLLMCompletionProgram.from_defaults(
output_parser=PydanticOutputParser(PlantHealth),
image_documents=[SimpleDirectoryReader(input_files=[image_path]).load_data()[0]],
prompt_template_str=prompt_template_str,
multi_modal_llm=gemini_llm,
)
return llm_program().dict()
image_tool = FunctionTool.from_defaults(
fn=analyze_plant_image,
name="analyze_plant_image",
description="Analyzes a plant image to identify type, condition, and symptoms."
)
def search_plant_info(query: str) -> str:
with DDGS() as ddgs:
results = ddgs.text(query, max_results=3)
return "\n".join([f"{r['title']}: {r['body']}" for r in results])
search_tool = FunctionTool.from_defaults(
fn=search_plant_info,
name="search_plant_info",
description="Searches the web for plant health and care information."
)
agent = ReActAgent.from_tools(
tools=[image_tool, search_tool],
llm=Gemini(model_name="gemini-2.0-flash"),
verbose=True
)
def plant_health_app(image, user_query):
image_path = "temp_image.jpg"
image.save(image_path)
# Query the ReAct agent
response = agent.chat(f"Analyze this plant image : {image_path}, give a recommendations to cure the issue and return the answer in points. User query: {user_query}")
# Parse response
recommendations = str(response)
output_text = recommendations
return output_text
iface = gr.Interface(
fn=plant_health_app,
inputs=[
gr.Image(type="pil", label="Upload Plant Image"),
gr.Textbox(label="Describe Symptoms or Ask a Question", placeholder="E.g., My tomato plant has yellow spots")
],
outputs=gr.Markdown(label="Diagnosis and Recommendations"),
title="🌿 Planthy -- Monitor Your Plant Health",
description="Upload a clear plant image and describe symptoms to get a diagnosis and care tips.",
examples=[
["tomato.jpeg", "What’s wrong with my tomato plant?"],
["rose.jpeg", "Is my rose plant healthy?"]
]
)
if __name__ == "__main__":
iface.launch() |