Skip to content

Latest commit

 

History

History
186 lines (150 loc) · 4.99 KB

getting-started.mdx

File metadata and controls

186 lines (150 loc) · 4.99 KB
title
Getting Started

MindsDB can be integrated with the most popular databases, as well as with the DBT and MLflow workflows.

To try out MindsDB right away without bringing in your own data or models, follow our Quickstart guide.

1. Create a MindsDB Cloud Account or Install MindsDB Locally

Create your [free MindsDB Cloud account](https://cloud.mindsdb.com/signup). To get started with a Docker installation, follow the MindsDB installation instructions using [Docker](/setup/self-hosted/docker/). You can also install MindsDB using pip on [Windows](/setup/self-hosted/pip/windows/), [Mac](/setup/self-hosted/pip/macos/), and [Linux](/setup/self-hosted/pip/linux/)

2. Connect to MindsDB from a SQL Client

If you do not have a preferred SQL client yet, we recommend using the [MindsDB SQL Editor](https://cloud.mindsdb.com/editor) or [DBeaver Community Edition](https://dbeaver.io/download/). Follow [this guide](/setup/cloud/) to set up your MindsDB SQL Editor. And [here](/connect/dbeaver/), you'll find how to connect to MindsDB from DBeaver. By default, on MindsDB Cloud the SQL Editor is already connected. Skip to step 3 a. Create a new MySQL connection. 
        ![DBeaver Create Connection](/assets/dbeaver-create-connection.png)

    b. Configure it using the parameters below, as well as your username and password.

        ```
        Host: `cloud.mindsdb.com`
        Port: `3306`
        Database: `mindsdb`
        ```

        ![DBeaver Configure Connection](/assets/dbeaver-configure-cloud-connection.png)
</Tab>
<Tab title="Local to Dbeaver">
    a. Create a new MySQL connection.

        ![DBeaver Create Connection](/assets/dbeaver-create-connection.png)

    b. Configure it using the following parameters:

        ```
        Host: `localhost`
        Port: `47335`
        Database: `mindsdb`
        Username: `mindsdb`
        Password: <leave it empty>
        ```

        ![DBeaver Configure Connection](/assets/dbeaver-configure-docker-connection.png)
</Tab>

3. Connect your Data to MindsDB Using CREATE DATABASE

CREATE DATABASE example_data
WITH ENGINE = "postgres",
PARAMETERS = {
  "user": "demo_user",
  "password": "demo_password",
  "host": "3.220.66.106",
  "port": "5432",
  "database": "demo"
};

4. Preview the Available Data Using SELECT

SELECT *
FROM example_data.demo_data.home_rentals
LIMIT 10;

5. Create a Model Using CREATE MODEL

If you already have a model in MLFlow, you can connect to your model.

```sql CREATE MODEL mindsdb.home_rentals_predictor FROM example_data (SELECT * FROM demo_data.home_rentals) PREDICT rental_price; ``` ```sql CREATE MODEL mindsdb.home_rentals_predictor FROM example_data (select * from demo_data.home_rentals) PREDICT rental_price USING url.predict='http://host.docker.internal:1234/invocations', format='mlflow', dtype_dict={"number_of_rooms": "categorical", "number_of_bathrooms": "categorical", "sqft": "integer", "days_on_market": "integer", "initial_price": "integer", "location": "categorical", "neighborhood":"categorical" }; ```

6. Make Predictions Using SELECT

SELECT rental_price
FROM mindsdb.home_rentals_predictor
WHERE number_of_bathrooms = 2
AND sqft = 1000;

On execution, we get:

+--------------+
| rental_price |
+--------------+
| 1130         |
+--------------+

7. Integrate your Predictions into the DBT Workflow

To do so, you need to make the following changes:

```yml profiles.yml mindsdb: type: mysql host: mysql.mindsdb.com user: [email protected] password: mindsdbpassword port: 3306 dbname: mindsdb schema: example_data threads: 1 keepalives_idle: 0 # default 0, indicating the system default connect_timeout: 10 # default 10 seconds ``` 
version: 2

models:
    - name: predicted_rentals
    description: "Integrating MindsDB predictions and historical   data"
with predictions as (
    SELECT hrp.rental_price as predicted_price, hr.rental_price as actual_price
    FROM mindsdb.home_rentals_predictor hrp
    JOIN exampleData.demo_data.home_rentals hr
    WHERE hr.number_of_bathrooms=2 AND hr.sqft=1000;
)
select * from predictions;
models:
  home_rentals:
    +materialized: view