first commit

.gitignore

@@ -0,0 +1,30 @@
+# -----------------------------
+# Environment variables
+# -----------------------------
+.env
+
+# -----------------------------
+# Python cache files
+# -----------------------------
+__pycache__/
+*.py[cod]
+*$py.class
+
+# -----------------------------
+# Virtual environments
+# -----------------------------
+myvenv/
+env/
+venv/
+ENV/
+
+# -----------------------------
+# VS Code / IDE
+# -----------------------------
+.vscode/
+.idea/
+
+# -----------------------------
+# Test Files
+# -----------------------------
+test.py

Dockerfile

@@ -0,0 +1,59 @@
+# =========================================================
+# BASE IMAGE
+# =========================================================
+
+FROM python:3.11-slim
+
+# =========================================================
+# ENV VARIABLES
+# =========================================================
+
+ENV PYTHONDONTWRITEBYTECODE=1
+ENV PYTHONUNBUFFERED=1
+
+# =========================================================
+# WORK DIRECTORY
+# =========================================================
+
+WORKDIR /app
+
+# =========================================================
+# INSTALL SYSTEM DEPENDENCIES
+# =========================================================
+
+RUN apt-get update && apt-get install -y \
+    gcc \
+    g++ \
+    && rm -rf /var/lib/apt/lists/*
+
+# =========================================================
+# COPY REQUIREMENTS
+# =========================================================
+
+COPY requirements.txt .
+
+# =========================================================
+# INSTALL PYTHON DEPENDENCIES
+# =========================================================
+
+RUN pip install --upgrade pip
+
+RUN pip install --no-cache-dir -r requirements.txt
+
+# =========================================================
+# COPY PROJECT FILES
+# =========================================================
+
+COPY . .
+
+# =========================================================
+# EXPOSE PORT
+# =========================================================
+
+EXPOSE 7860
+
+# =========================================================
+# START FASTAPI
+# =========================================================
+
+CMD ["uvicorn", "app.main:app", "--host", "0.0.0.0", "--port", "7860"]

app/core/config.py

@@ -0,0 +1,10 @@
+# app/core/config.py
+import os
+from dotenv import load_dotenv
+
+load_dotenv()
+
+HF_TOKEN = os.getenv("HF_TOKEN")
+GEMINI_API_KEY = os.getenv("GEMINI_API_KEY")
+QDRANT_API_KEY = os.getenv("QDRANT_API_KEY")
+QDRANT_URL = os.getenv("QDRANT_URL")

app/core/llm.py

@@ -0,0 +1,9 @@
+from openai import AsyncOpenAI
+from app.core.config import HF_TOKEN
+
+def get_llm():
+
+    return AsyncOpenAI(
+        base_url="https://router.huggingface.co/v1",
+        api_key=HF_TOKEN
+    )

app/main.py

@@ -0,0 +1,19 @@
+from fastapi import FastAPI
+
+from app.routes.predict import router as predict_router
+from app.routes.chat import router as chat_router
+
+app = FastAPI(
+    title="NeuroFlora API"
+)
+
+app.include_router(predict_router)
+app.include_router(chat_router)
+
+@app.get("/")
+
+async def root():
+
+    return {
+        "message": "NeuroFlora API Running"
+    }

app/memory/qdrant_memory.py

@@ -0,0 +1,93 @@
+from qdrant_client import QdrantClient
+from mem0 import Memory
+
+from app.core.config import *
+
+
+# =========================================================
+# MEM0 + QDRANT CONFIG
+# =========================================================
+
+config = {
+    "version": "v1.1",
+
+    "llm": {
+        "provider": "gemini",
+        "config": {
+            "api_key": GEMINI_API_KEY,
+            "model": "gemini-2.5-flash"
+        }
+    },
+
+    "embedder": {
+        "provider": "gemini",
+        "config": {
+            "api_key": GEMINI_API_KEY,
+            "model": "gemini-embedding-001"
+        }
+    },
+
+    "vector_store": {
+        "provider": "qdrant",
+        "config": {
+            "url": QDRANT_URL,
+            "api_key": QDRANT_API_KEY,
+            "collection_name": "mem0_memory",
+            "embedding_model_dims": 768
+        }
+    }
+}
+
+mem_client = Memory.from_config(config)
+
+# =========================================================
+# SEARCH MEMORY
+# =========================================================
+
+def search_memory(user_id, user_query):
+
+    try:
+
+        results = mem_client.search(
+            query=user_query,
+            filters={"user_id": user_id}
+        )
+
+        memories = [
+            f"ID:{mem.get('id')}\nMemory:{mem.get('memory')}"
+            for mem in results.get("results", [])
+        ]
+
+        return memories
+
+    except Exception as e:
+
+        print(f"[MEMORY SEARCH ERROR]: {e}")
+
+        return []
+
+# =========================================================
+# ADD MEMORY
+# =========================================================
+
+def add_memory(user_id, user_query, ai_response):
+
+    try:
+
+        mem_client.add(
+            user_id=user_id,
+            messages=[
+                {
+                    "role": "user",
+                    "content": user_query
+                },
+                {
+                    "role": "assistant",
+                    "content": ai_response
+                }
+            ]
+        )
+
+    except Exception as e:
+
+        print(f"[MEMORY ADD ERROR]: {e}")

app/model/model.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e4ad08428b471d099fe46c5422d4740ab6cfe11c7f97536faf4f9c73110cfa19
+size 243992671

app/model/predictor.py

@@ -0,0 +1,73 @@
+import torch
+import torch.nn as nn
+from torchvision import transforms
+from torchvision.models import efficientnet_v2_s
+from PIL import Image
+from collections import OrderedDict
+
+DEVICE = torch.device("cuda" if torch.cuda.is_available() else "cpu")
+
+MODEL_PATH = "app/model/model.pth"
+IMG_SIZE = 224
+
+checkpoint = torch.load(
+    MODEL_PATH,
+    map_location=DEVICE,
+    weights_only=False
+)
+
+class_names = checkpoint["class_names"]
+num_classes = len(class_names)
+
+model = efficientnet_v2_s(weights=None)
+
+in_features = model.classifier[1].in_features
+
+model.classifier = nn.Sequential(
+    nn.Dropout(p=0.3, inplace=True),
+    nn.Linear(in_features, num_classes),
+)
+
+state_dict = checkpoint["model_state_dict"]
+
+new_state_dict = OrderedDict()
+
+for k, v in state_dict.items():
+    name = k.replace("module.", "")
+    new_state_dict[name] = v
+
+if "n_averaged" in new_state_dict:
+    del new_state_dict["n_averaged"]
+
+model.load_state_dict(new_state_dict)
+
+model.to(DEVICE)
+model.eval()
+
+transform = transforms.Compose([
+    transforms.Resize((IMG_SIZE, IMG_SIZE)),
+    transforms.ToTensor(),
+    transforms.Normalize(
+        [0.485, 0.456, 0.406],
+        [0.229, 0.224, 0.225]
+    ),
+])
+
+def predict_image(image):
+
+    image = image.convert("RGB")
+
+    image_tensor = transform(image).unsqueeze(0).to(DEVICE)
+
+    with torch.no_grad():
+
+        outputs = model(image_tensor)
+
+        probabilities = torch.softmax(outputs, dim=1)
+
+        confidence, predicted = torch.max(probabilities, 1)
+
+    return {
+        "prediction": class_names[predicted.item()],
+        "confidence": round(confidence.item() * 100, 2)
+    }

app/routes/chat.py

@@ -0,0 +1,30 @@
+from fastapi import APIRouter
+from fastapi.responses import StreamingResponse
+from pydantic import BaseModel
+
+from app.services.chat_service import generate_response
+from app.utils.streaming import stream_response
+
+router = APIRouter()
+
+class ChatRequest(BaseModel):
+    user_id: str
+    message: str
+
+@router.post("/chat")
+
+async def chat(request: ChatRequest):
+
+    response = await generate_response(
+        request.user_id,
+        request.message
+    )
+
+    return StreamingResponse(
+        stream_response(
+            response,
+            request.user_id,
+            request.message
+        ),
+        media_type="text/plain"
+    )

app/routes/predict.py

@@ -0,0 +1,23 @@
+from fastapi import APIRouter, UploadFile, File
+from PIL import Image
+import io
+
+from app.model.predictor import predict_image
+
+router = APIRouter()
+
+@router.post("/predict")
+
+async def predict(
+    file: UploadFile = File(...)
+):
+
+    image_bytes = await file.read()
+
+    image = Image.open(
+        io.BytesIO(image_bytes)
+    )
+
+    result = predict_image(image)
+
+    return result

app/services/chat_service.py

@@ -0,0 +1,48 @@
+from app.core.llm import get_llm
+from app.memory.qdrant_memory import search_memory
+
+async def generate_response(
+    user_id,
+    user_query
+):
+
+    memories = search_memory(
+        user_id,
+        user_query
+    )
+
+    memory_context = "\n".join(memories)
+
+    SYSTEM_PROMPT = f"""
+    You are NeuroFlora,
+    an intelligent plant disease assistant.
+
+    Previous user memories:
+    {memory_context}
+
+    Help users with:
+    - plant diseases
+    - crop health
+    - farming guidance
+    - pesticide awareness
+    - plant care
+    """
+
+    client = get_llm()
+
+    response = await client.chat.completions.create(
+        model="meta-llama/Llama-3.1-70B-Instruct:scaleway",
+        messages=[
+            {
+                "role": "system",
+                "content": SYSTEM_PROMPT
+            },
+            {
+                "role": "user",
+                "content": user_query
+            }
+        ],
+        stream=True
+    )
+
+    return response

app/utils/streaming.py

@@ -0,0 +1,39 @@
+import asyncio
+import random
+from functools import partial
+from app.memory.qdrant_memory import add_memory
+
+async def _store_memory(user_id, user_query, full_response):
+    try:
+        loop = asyncio.get_event_loop()
+        await loop.run_in_executor(
+            None,  # uses default ThreadPoolExecutor
+            partial(add_memory, user_id, user_query, full_response)
+        )
+    except Exception as e:
+        print(f"[MEMORY STORE ERROR]: {e}")
+
+
+async def stream_response(response, user_id, user_query):
+    full_response = ""
+
+    try:
+        async for chunk in response:
+            delta = chunk.choices[0].delta
+
+            if delta and delta.content:
+                token = delta.content
+
+                for char in token:
+                    full_response += char
+                    yield char
+                    await asyncio.sleep(random.uniform(0.01, 0.02))
+
+    except Exception as e:
+        yield f"\n[ERROR]: {str(e)}"
+
+    finally:
+        if full_response:
+            asyncio.create_task(
+                _store_memory(user_id, user_query, full_response)
+            )

requirements.txt

@@ -0,0 +1,27 @@
+fastapi==0.115.0
+uvicorn==0.30.6
+
+torch==2.5.1
+torchvision==0.20.1
+
+pillow==10.4.0
+
+openai==1.51.0
+
+python-multipart==0.0.9
+
+qdrant-client==1.11.3
+
+mem0ai==0.1.48
+
+google-generativeai==0.8.3
+
+numpy==1.26.4
+
+pydantic==2.9.2
+
+httpx==0.27.2
+
+aiohttp==3.10.10
+
+dotenv