retriever.py

from smolagents import Tool
from langchain_community.retrievers import BM25Retriever
from langchain.docstore.document import Document

class VegetableInfoRetrieverTool(Tool):
    name = "vegetable_info_retriever"
    description = "Retrieve information about vegetables based on their name or characteristics."
    inputs = {
        "query": {
            "type": "string",
            "description": "The name or description of the vegetable you want information about."
        }
    }
    output_type = "string"

    def __init__(self, docs):
        self.is_initialized = False
        self.retriever = BM25Retriever.from_documents(docs)

    def forward(self, query: str):
        results = self.retriever.get_relevant_documents(query)
        if results:
            return "\n\n".join([doc.page_content for doc in results[:3]])
        else:
            return "No matching vegetable information found."


def load_vegetable_dataset():
    # Manually define the vegetable data
    vegetable_data = [
        {
            "name": "Sweet Potatoes",
            "type": "root",
            "description": "A starchy root vegetable with sweet flavor, rich in fiber and beta-carotene."
        },
        {
            "name": "Fresh Basil",
            "type": "herb",
            "description": "An aromatic green herb often used fresh in Mediterranean cuisine."
        },
        {
            "name": "Broccoli",
            "type": "cruciferous",
            "description": "Green vegetable with edible flowering heads and stalks, high in fiber and vitamin C."
        },
        {
            "name": "Celery",
            "type": "stem",
            "description": "Crisp green stalks often eaten raw or cooked, known for high water content and fiber."
        },
        {
            "name": "Lettuce",
            "type": "leaf",
            "description": "Leafy green used in salads, comes in various varieties like romaine or butterhead."
        },
    ]

    # Convert each entry to a Document
    docs = [
        Document(
            page_content="\n".join([
                f"Name: {veg['name']}",
                f"Type: {veg['type']}",
                f"Description: {veg['description']}"
            ]),
            metadata={"name": veg["name"]}
        )
        for veg in vegetable_data
    ]

    # Return the tool
    return VegetableInfoRetrieverTool(docs)