MamMates / ml-model-serving

Serving Services for All ML Model

ml-model-serving

Serving Services for All ML Model using FastAPI and Docker.

ML Model 🤖

Here is the list of ML Model that we use in this project:

• Food Classification
• Food Rating
• Food Price
• Food Recommendation

Features 💡

Using ML Model Serving, you can:

• Get prediction from our ML model for predicting food category, rating, and price.
• Get food recommendation from our ML model by given user_id.
• Get status and metadata from our ML model.

Prerequisites 📋

• Python 3.10 or higher
• FastAPI 0.104.1 or higher
• Docker 24.0.7 or higher
• Docker Compose 2.15.1 or higher

Usage ✨

Actually, if you already have Docker and Docker Compose, you just need the compose.yml file.

• You can run this app with Docker Compose.

docker compose up

• Now, you can access the app on http://localhost:8080
• For doing prediction or get food recommendation, please read the API Endpoints section.

Development 💻

If you want to develop the model serving, you can follow this step:

• Clone the repository:

git clone https://github.com/MamMates/ml-model-serving.git

• Create and activate virtual environment:

python -m venv venv
source venv/bin/activate

• Install dependencies:

pip install -r requirements.txt

• To develop the FastAPI services, you can modify the compose.yml file, change the image in app services to build:

- image: putuwaw/mammates-model-serving
+ build: .

• Run the app:

docker compose up --build

• Now, you can access the app on http://localhost:8080

• You can also run the test using pytest:

pytest

API Endpoints 📡

List of available endpoints:

GET

GET / - Get hello world.

Response

{
  "status": true,
  "code": 200,
  "message": "OK",
  "data": {
    "message": "Hello World"
  }
}

GET /recommendation - Get food recommendation (top 5).

Name	Params	Required	Type	Description
user_id	Query	required	integer	The id of user. Example 14

Response

{
  "status": true,
  "code": 200,
  "message": "OK",
  "data": {
    "food_id": [13, 14, 12, 2, 18]
  }
}

GET /model/{model_name} - Get status and metadata from model.

Name	Params	Required	Type	Description
model_name	Path	required	string	The name of model. Example food_clf

Response

{
  "status": true,
  "code": 200,
  "message": "OK",
  "data": {
    "status": {
      "model_version_status": [
        {
          "version": "1",
          "state": "AVAILABLE",
          "status": {
            "error_code": "OK",
            "error_message": ""
          }
        }
      ]
    },
    "metadata": {
      "model_spec": {
        "name": "food_clf",
        "signature_name": "",
        "version": "1"
      },
      "metadata": {
        "signature_def": {
          "signature_def": {
            "serving_default": {
              "inputs": {
                "input_2": {
                  "dtype": "DT_FLOAT",
                  "tensor_shape": {
                    "dim": [
                      {
                        "size": "-1",
                        "name": ""
                      },
                      {
                        "size": "150",
                        "name": ""
                      },
                      {
                        "size": "150",
                        "name": ""
                      },
                      {
                        "size": "3",
                        "name": ""
                      }
                    ],
                    "unknown_rank": false
                  },
                  "name": "serving_default_input_2:0"
                }
              },
              "outputs": {
                "dense": {
                  "dtype": "DT_FLOAT",
                  "tensor_shape": {
                    "dim": [
                      {
                        "size": "-1",
                        "name": ""
                      },
                      {
                        "size": "10",
                        "name": ""
                      }
                    ],
                    "unknown_rank": false
                  },
                  "name": "StatefulPartitionedCall:0"
                }
              },
              "method_name": "tensorflow/serving/predict",
              "defaults": {}
            },
            "__saved_model_init_op": {
              "inputs": {},
              "outputs": {
                "__saved_model_init_op": {
                  "dtype": "DT_INVALID",
                  "tensor_shape": {
                    "dim": [],
                    "unknown_rank": true
                  },
                  "name": "NoOp"
                }
              },
              "method_name": "",
              "defaults": {}
            }
          }
        }
      }
    }
  }
}

POST

POST /predict - Predict category, rating, and price of image.

Name	Params	Required	Type	Description
province	Query	required	string	The province of seller. Example Bali
environment	Query	optional	string	The environment of seller. Default campus.
name	Query	optional	string	The name of food. Default null. Example donat ubi mawar
image	Body	required	file	The image to predict

Response

{
  "status": true,
  "code": 200,
  "message": "OK",
  "data": {
    "category": 2,
    "rating": 3,
    "price": 10000
  }
}

License 📝

This project is licensed under the MIT License. See the LICENSE file for details.