Installing and Configuring HCL Translator

Clara V2 provides multilanguage enablement through any the following services:

 

Notes:

Multilanguage enablement in Clara is obtained through machine translation. Therefore, it could sometimes produce inaccurate results.

If you are using HCL Translator and you find that the translation is not accurate, you can submit your feedback through the HCL Software portal by providing the original sentence and the correct translation.

In this way, the HCL Translator development team can use your input to re-train the language model and improve the translation quality.

 

HCL Translator is an API based service. It is based on Machine Learning models that have been trained and fine-tuned to understand context-specific terms or phrases used in the Workload Automation domain, thus greatly improving the user experience. It supports bidirectional automatic translation from English to French, German, Italian, Portuguese, and Spanish.

 

After Clara installation, you can customize multilanguage enablement in Clara, by selecting one of the available services. For details, see Translations.

 

To use HCL Translator, you must first install it. HCL Translator is available on HCL Software Factory (SoFy), a catalog of Kubernetes enabled products that can be deployed to a cloud-native environment.

It is also available, together with Clara, on HCL License Portal as a package for on-premises installation with Docker.

To deploy HCL Translator, run the following procedure.

 

Installation

  1. Download the Translator package from the source repository and extract it with a tool available on your system.

  1. Open a bash shell and move to the extracted directory with the following command:

cd <extract-dir>

where <extract-dir> is the path where you have extracted the package.

  1. Load the Docker images by running the following command:

./translator.sh load

  1. Install the Translator in a directory of your choice by running the following command:

./translator.sh install --install-dir=<install-dir>

 where <install-dir> is the installation directory of your choice (default value is /opt/hcl/).

  1. Download the language packages of your interest from the source repository and move them to <install-dir> without extracting them.

  1. On the bash shell move to the installation directory with the following command:

cd <install-dir>

  1.  Install the language packages by running the following command:

./translator.sh install-language --language-file=<lang-package>

where <lang_package> is the name of the language package file (e.g. `it.tar.gz`).

Notes: you can delete the package file after it has been successfully installed.

  1. Repeat the previous step for each language that you have downloaded.

  1. Configure the Translator options by following the Configuration section.

  1. Run the Translator with the following command:

./translator.sh up

 

 

Configuration

To configure the Translator, open the <install-dir> and edit the .env file with a text editor of your choice:

This file contains the environment variables used by HCL Translator in the from `VARIABLE=VALUE`. You can edit the `VALUE` based on your needs.

Here's the meaning of each variable:

INSTALL_DIR is the directory where HCL Translator is installed. The default value is `./` that represents the current working directory. You should not change this if you followed the standard installation procedure.

IMG_REGISTRY is the registry of the Docker images. You should not need to change this variable.

IMG_TAG is the tag of the Docker images that represents the Translator version. You should not need to change this variable.

NETWORK_NAME is the Docker network the translator must connect to. The default is `clara-net`, so the Translator uses the network of Clara and is not accessible from the outside. Change this only if you install the Translator on a different machine than Clara's.

EXTERNAL_NETWORK, if true (default), the Translator uses a pre-existing Docker network. If false, it creates a new one. If you want to use `clara-net`, you must leave the default value. Set this to false only if you install the Translator on a different machine than Clara's.

 

docker-compose.yaml

This file contains the configuration of docker-compose. You should not edit this file unless you install the Translator on a different machine than Clara's. In this case, you should uncomment the following lines by removing the `#` at the beginning:

        

        # ports:

        #     - <PORT>:5000

 

In this way, the Translator will be exposed to the outside. Replace <PORT> with the desired port that you want to expose.

WARNING: exposing the Translator to the outside is a potential security issue, because it is not protected by any authentication. So, be sure that the exposed port is only accessible by the Clara machine.

 

 

translator.sh script

Usage:    ./translator.sh COMMAND [OPTIONS]

 

This script manages the translator deployment.

 

Commands:

  load                  Loads the container images (required before the first start)

  install               Installs translator in a directory (by default in /opt/hcl/translator)

  install-language      Installs a language in a directory (by default the current working directory)

  up                    Creates the containers and starts the application

  stop                  Stops the containers

  restart               Restarts the containers

  down                  Removes the containers (but not volumes)

  down-volumes          Removes the containers and volumes

 

Options:

  --images              The archive path of the translator images (defaults to ./translator-images.tar.gz)

  --install-file        The archive path of the translator base files to install (defaults to ./translator.tar.gz)

  --install-dir         The directory where translator must be installed (defaults to /opt/hcl/translator)

  --language-file       The archive path of the language file to install (defaults to ./it.tar.gz)

  --language-dir        The directory where the language must be installed (defaults to the current working directory)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Leave feedback