ClickHouse/docs/en/guides/apply-catboost-model.md
Olga Revyakina f4f5e5c09e Start
2021-04-18 02:28:47 +03:00

7.0 KiB
Raw Blame History

toc_priority toc_title
41 Applying CatBoost Models

Applying a Catboost Model in ClickHouse

CatBoost is a free and open-source gradient boosting library developed at Yandex for machine learning.

With this instruction, you will learn to apply pre-trained models in ClickHouse by running model inference from SQL.

To apply a CatBoost model in ClickHouse:

  1. Create a Table.
  2. Insert the Data to the Table.
  3. Integrate CatBoost into ClickHouse (Optional step).
  4. Run the Model Inference from SQL.

For more information about training CatBoost models, see Training and applying models.

Prerequisites

If you dont have the Docker yet, install it.

!!! note "Note" Docker is a software platform that allows you to create containers that isolate a CatBoost and ClickHouse installation from the rest of the system.

Before applying a CatBoost model:

1. Pull the Docker image from the registry:

$ docker pull yandex/tutorial-catboost-clickhouse

This Docker image contains everything you need to run CatBoost and ClickHouse: code, runtime, libraries, environment variables, and configuration files.

2. Make sure the Docker image has been successfully pulled:

$ docker image ls
REPOSITORY                            TAG                 IMAGE ID            CREATED             SIZE
yandex/tutorial-catboost-clickhouse   latest              622e4d17945b        22 hours ago        1.37GB

3. Start a Docker container based on this image:

$ docker run -it -p 8888:8888 yandex/tutorial-catboost-clickhouse

1. Create a Table

To create a ClickHouse table for the training sample:

1. Start ClickHouse console client in the interactive mode:

$ clickhouse client

!!! note "Note" The ClickHouse server is already running inside the Docker container.

2. Create the table using the command:

:) CREATE TABLE amazon_train
(
    date Date MATERIALIZED today(),
    ACTION UInt8,
    RESOURCE UInt32,
    MGR_ID UInt32,
    ROLE_ROLLUP_1 UInt32,
    ROLE_ROLLUP_2 UInt32,
    ROLE_DEPTNAME UInt32,
    ROLE_TITLE UInt32,
    ROLE_FAMILY_DESC UInt32,
    ROLE_FAMILY UInt32,
    ROLE_CODE UInt32
)
ENGINE = MergeTree ORDER BY date

3. Exit from ClickHouse console client:

:) exit

2. Insert the Data to the Table

To insert the data:

1. Run the following command:

$ clickhouse client --host 127.0.0.1 --query 'INSERT INTO amazon_train FORMAT CSVWithNames' < ~/amazon/train.csv

2. Start ClickHouse console client in the interactive mode:

$ clickhouse client

3. Make sure the data has been uploaded:

:) SELECT count() FROM amazon_train

SELECT count()
FROM amazon_train

+-count()-+
|   65538 |
+-------+

3. Integrate CatBoost into ClickHouse

!!! note "Note" Optional step. The Docker image contains everything you need to run CatBoost and ClickHouse.

To integrate CatBoost into ClickHouse:

1. Build the evaluation library.

The fastest way to evaluate a CatBoost model is compile libcatboostmodel.<so|dll|dylib> library. For more information about how to build the library, see CatBoost documentation.

2. Create a new directory anywhere and with any name, for example, data and put the created library in it. The Docker image already contains the library data/libcatboostmodel.so.

3. Create a new directory for config model anywhere and with any name, for example, models.

4. Create a model configuration file with any name, for example, models/amazon_model.xml.

5. Describe the model configuration:

<models>
    <model>
        <!-- Model type. Now catboost only. -->
        <type>catboost</type>
        <!-- Model name. -->
        <name>amazon</name>
        <!-- Path to trained model. -->
        <path>/home/catboost/tutorial/catboost_model.bin</path>
        <!-- Update interval. -->
        <lifetime>0</lifetime>
    </model>
</models>

6. Add the path to CatBoost and the model configuration to the ClickHouse configuration:

<!-- File etc/clickhouse-server/config.d/models_config.xml. -->
<catboost_dynamic_library_path>/home/catboost/data/libcatboostmodel.so</catboost_dynamic_library_path>
<models_config>/home/catboost/models/*_model.xml</models_config>

!!! note "Note" You can change path to the CatBoost model configuration later without restarting server.

4. Run the Model Inference from SQL

For test model run the ClickHouse client $ clickhouse client.

Lets make sure that the model is working:

:) SELECT
    modelEvaluate('amazon',
                RESOURCE,
                MGR_ID,
                ROLE_ROLLUP_1,
                ROLE_ROLLUP_2,
                ROLE_DEPTNAME,
                ROLE_TITLE,
                ROLE_FAMILY_DESC,
                ROLE_FAMILY,
                ROLE_CODE) > 0 AS prediction,
    ACTION AS target
FROM amazon_train
LIMIT 10

!!! note "Note" Function modelEvaluate returns tuple with per-class raw predictions for multiclass models.

Lets predict the probability:

:) SELECT
    modelEvaluate('amazon',
                RESOURCE,
                MGR_ID,
                ROLE_ROLLUP_1,
                ROLE_ROLLUP_2,
                ROLE_DEPTNAME,
                ROLE_TITLE,
                ROLE_FAMILY_DESC,
                ROLE_FAMILY,
                ROLE_CODE) AS prediction,
    1. / (1 + exp(-prediction)) AS probability,
    ACTION AS target
FROM amazon_train
LIMIT 10

!!! note "Note" More info about exp() function.

Lets calculate LogLoss on the sample:

:) SELECT -avg(tg * log(prob) + (1 - tg) * log(1 - prob)) AS logloss
FROM
(
    SELECT
        modelEvaluate('amazon',
                    RESOURCE,
                    MGR_ID,
                    ROLE_ROLLUP_1,
                    ROLE_ROLLUP_2,
                    ROLE_DEPTNAME,
                    ROLE_TITLE,
                    ROLE_FAMILY_DESC,
                    ROLE_FAMILY,
                    ROLE_CODE) AS prediction,
        1. / (1. + exp(-prediction)) AS prob,
        ACTION AS tg
    FROM amazon_train
)

!!! note "Note" More info about avg() and log() functions.

Original article