Updated

Dockerfile

@@ -1,45 +0,0 @@
-# Use lightweight Python image
-FROM python:3.10-slim
-
-# Install system dependencies
-RUN apt-get update && apt-get install -y \
-    build-essential \
-    git \
-    curl \
-    && rm -rf /var/lib/apt/lists/*
-
-# Create non-root user
-RUN useradd -m appuser
-
-# Set work directory
-WORKDIR /app
-
-# Copy requirements first (for better caching)
-COPY requirements.txt .
-
-# Install Python dependencies
-RUN pip install --no-cache-dir -r requirements.txt
-
-# Copy project files
-COPY . .
-
-# Create Hugging Face cache dir and give full permissions
-RUN mkdir -p /app/huggingface_cache && chmod -R 777 /app/huggingface_cache
-
-# Change ownership of /app to appuser
-RUN chown -R appuser:appuser /app
-
-# Switch to non-root user
-USER appuser
-
-# Expose Hugging Face Spaces allowed port
-EXPOSE 7860
-
-# Hugging Face cache environment variables
-ENV HF_HOME=/app/huggingface_cache
-ENV HF_DATASETS_CACHE=/app/huggingface_cache
-ENV TORCH_HOME=/app/huggingface_cache
-
-
-# Start FastAPI app on port 7860
-CMD ["uvicorn", "app:app", "--host", "0.0.0.0", "--port", "7860"]

README.md

@@ -1,12 +0,0 @@
----
-title: AgriCopilot
-emoji: 📉
-colorFrom: purple
-colorTo: gray
-sdk: docker
-pinned: false
-license: apache-2.0
-short_description: AgriCopilot is an AgenticAI-powered super-app for farmers
----
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -1,80 +0,0 @@
-# app.py
-import os
-from fastapi import FastAPI
-from langchain.prompts import PromptTemplate
-from langchain_huggingface import HuggingFaceEndpoint
-from vector import query_vector
-
-app = FastAPI(title="AgriCopilot")
-
-# ==============================
-# ROOT HEALTH CHECK
-# ==============================
-@app.get("/")
-async def root():
-    return {"status": "AgriCopilot AI Backend is working perfectly"}
-
-# ==============================
-# MODELS PER ENDPOINT
-# ==============================
-
-# 1. Crop Doctor (Image/Text)
-crop_template = PromptTemplate(
-    input_variables=["symptoms"],
-    template="You are an agricultural crop doctor. A farmer reports: {symptoms}. Diagnose the most likely disease and suggest treatments in simple farmer-friendly language."
-)
-crop_llm = HuggingFaceEndpoint(repo_id="facebook/bart-large", task="text2text-generation")
-
-# 2. Multilingual Chat
-chat_template = PromptTemplate(
-    input_variables=["query"],
-    template="You are a multilingual AI assistant for farmers. Answer clearly in the same language as the user. Farmer says: {query}"
-)
-chat_llm = HuggingFaceEndpoint(repo_id="google/mt5-base", task="text2text-generation")
-
-# 3. Disaster Summarizer
-disaster_template = PromptTemplate(
-    input_variables=["report"],
-    template="You are an AI disaster assistant. Summarize the following report for farmers in simple steps: {report}"
-)
-disaster_llm = HuggingFaceEndpoint(repo_id="google/flan-t5-base", task="text2text-generation")
-
-# 4. Marketplace Recommendation
-market_template = PromptTemplate(
-    input_variables=["product"],
-    template="You are an agricultural marketplace recommender. Farmer wants to sell or buy: {product}. Suggest possible matches and advice."
-)
-market_llm = HuggingFaceEndpoint(repo_id="tiiuae/falcon-7b-instruct", task="text2text-generation")
-
-# ==============================
-# ENDPOINTS
-# ==============================
-
-@app.post("/crop-doctor")
-async def crop_doctor(symptoms: str):
-    prompt = crop_template.format(symptoms=symptoms)
-    response = crop_llm(prompt)
-    return {"diagnosis": response}
-
-@app.post("/multilingual-chat")
-async def multilingual_chat(query: str):
-    prompt = chat_template.format(query=query)
-    response = chat_llm(prompt)
-    return {"reply": response}
-
-@app.post("/disaster-summarizer")
-async def disaster_summarizer(report: str):
-    prompt = disaster_template.format(report=report)
-    response = disaster_llm(prompt)
-    return {"summary": response}
-
-@app.post("/marketplace")
-async def marketplace(product: str):
-    prompt = market_template.format(product=product)
-    response = market_llm(prompt)
-    return {"recommendation": response}
-
-@app.post("/vector-search")
-async def vector_search(query: str):
-    results = query_vector(query)
-    return {"results": results}

prepare_data.py

@@ -1,51 +0,0 @@
-# prepare_data.py
-import os
-import kagglehub
-import pandas as pd
-from datasets import load_dataset
-
-os.makedirs("datasets", exist_ok=True)
-
-# -----------------------
-# 1. PlantVillage (Kaggle)
-# -----------------------
-print("Downloading PlantVillage dataset...")
-pv_path = kagglehub.dataset_download("dittakavinikhita/plant-disease-prediction-disease-and-healthy")
-
-# Pick the metadata CSV if available
-for file in os.listdir(pv_path):
-    if file.endswith(".csv"):
-        src = os.path.join(pv_path, file)
-        dst = "datasets/plant_disease.csv"
-        pd.read_csv(src).to_csv(dst, index=False)
-        print("✅ Saved PlantVillage ->", dst)
-
-# -----------------------
-# 2. AfriQA (Hugging Face)
-# -----------------------
-print("Downloading AfriQA dataset...")
-afriqa = load_dataset("masakhane/afriqa")
-afriqa_df = pd.DataFrame(afriqa["train"])
-
-# Merge question + answer into one text column
-afriqa_df["text"] = "Q: " + afriqa_df["question"].astype(str) + " A: " + afriqa_df["answer"].astype(str)
-afriqa_df[["text"]].to_csv("datasets/afriqa.csv", index=False)
-print("✅ Saved AfriQA -> datasets/afriqa.csv")
-
-# -----------------------
-# 3. CrisisNLP (Hugging Face)
-# -----------------------
-print("Downloading CrisisNLP dataset...")
-crisis = load_dataset("QCRI/CrisisBench-all-lang")
-crisis_df = pd.DataFrame(crisis["train"])
-
-# Pick relevant columns (tweet_text, label, etc.)
-if "tweet_text" in crisis_df.columns:
-    crisis_df["text"] = crisis_df["tweet_text"].astype(str)
-else:
-    crisis_df["text"] = crisis_df.astype(str).agg(" ".join, axis=1)
-
-crisis_df[["text"]].to_csv("datasets/crisis.csv", index=False)
-print("✅ Saved CrisisNLP -> datasets/crisis.csv")
-
-print("🎉 All datasets prepared in /datasets")

requirements.txt

@@ -1,12 +0,0 @@
-fastapi
-uvicorn
-langchain
-faiss-cpu
-huggingface-hub
-sentence-transformers
-langchain_community
-langchain
-langchain-huggingface
-kagglehub
-pandas
-datasets

vector.py

@@ -1,64 +0,0 @@
-# vector.py
-import os
-import glob
-import pandas as pd
-from langchain_community.vectorstores import FAISS
-from langchain_community.embeddings import HuggingFaceEmbeddings
-
-# ----------------- CONFIG -----------------
-VECTOR_PATH = "faiss_index"
-EMBEDDING_MODEL = os.getenv("HF_EMBEDDING_MODEL", "sentence-transformers/all-MiniLM-L6-v2")
-HF_CACHE_DIR = os.getenv("HF_CACHE_DIR", "/app/huggingface_cache")
-
-# Ensure cache dir exists
-os.makedirs(HF_CACHE_DIR, exist_ok=True)
-
-# ----------------- EMBEDDINGS -----------------
-embeddings = HuggingFaceEmbeddings(
-    model_name=EMBEDDING_MODEL,
-    cache_folder=HF_CACHE_DIR
-)
-
-# ----------------- VECTOR STORE -----------------
-def build_vectorstore():
-    """Build FAISS index from CSV datasets."""
-    texts = []
-    for file in glob.glob("datasets/*.csv"):
-        try:
-            df = pd.read_csv(file)
-            if "text" in df.columns:
-                texts.extend(df["text"].dropna().astype(str).tolist())
-            else:
-                # fallback: join all columns into one string
-                for _, row in df.iterrows():
-                    texts.append(" ".join(map(str, row.values)))
-            print(f"✅ Loaded {len(df)} rows from {file}")
-        except Exception as e:
-            print(f"⚠️ Skipping {file}, error: {e}")
-
-    if not texts:
-        texts = ["AgriCopilot initialized knowledge base."]
-
-    vectorstore = FAISS.from_texts(texts, embeddings)
-    vectorstore.save_local(VECTOR_PATH)
-    print("🎉 Vectorstore built with", len(texts), "documents")
-    return vectorstore
-
-def load_vector_store():
-    """Load FAISS index if available, else build new one."""
-    if os.path.exists(VECTOR_PATH):
-        return FAISS.load_local(
-            VECTOR_PATH,
-            embeddings,
-            allow_dangerous_deserialization=True
-        )
-    else:
-        return build_vectorstore()
-
-vectorstore = load_vector_store()
-
-# ----------------- QUERY -----------------
-def query_vector(query: str, k: int = 3):
-    """Perform similarity search on FAISS index."""
-    docs = vectorstore.similarity_search(query, k=k)
-    return [d.page_content for d in docs]