Integration: DeepL

Use DeepL translation services with Haystack

Authors
Dribia Data Research

Table of Contents

Overview

DeepL is a powerful translation services provider, offering high-quality translations in multiple languages. This integration allows you to use DeepL’s translation services with Haystack.

Installation

pip install deepl-haystack

Usage

Components

The DeepL Haystack integration introduces two components that can be used to obtain translations using the DeepL API.

  • The DeepLTextTranslator to translate plain text (Python strings).
  • The DeepLDocumentTranslator to translate Haystack Document objects.

API Key

To use the DeepL Haystack integration, you’ll need to provide a DeepL API key. You can get one by signing up at the DeepL API website.

Once obtained, make sure to export it as an environment variable named DEEPL_API_KEY in you working environment before running the examples below. Both the DeepLTextTranslator and the DeepLDocumentTranslator component constructors will expect this variable to be set.

An alternative way to provide the API key, although not recommended, would be to pass it through the api_key parameter of the components’ constructor, using the Haystack Secret utility.

Examples

Standalone

The following example shows how to translate a simple text:

from deepl_haystack import DeepLTextTranslator

translator = DeepLTextTranslator(source_lang="EN", target_lang="ES")

translated_text = translator.run("Hello, world!")
print(translated_text)
# {'translation': '¡Hola, mundo!', 'meta': {'source_lang': 'EN', 'target_lang': 'ES'}}

Here, instead, we show how to translate a list of Document objects:

from haystack.dataclasses import Document

from deepl_haystack import DeepLDocumentTranslator

translator = DeepLDocumentTranslator(source_lang="EN", target_lang="ES")

documents_to_translate = [
    Document(content="Hello, world!"),
    Document(content="Goodbye, Joe!", meta={"name": "Joe"}),
]

translated_documents = translator.run(documents_to_translate)
print(
    "\n".join(
        [f"{doc.content}, {doc.meta}" for doc in translated_documents["documents"]]
    )
)
# ¡Hola, mundo!, {'source_lang': 'EN', 'target_lang': 'ES'}
# ¡Adiós, Joe!, {'name': 'Joe', 'source_lang': 'EN', 'target_lang': 'ES'}

Pipeline

To use the DeepL components in a Haystack pipeline, you can use them as any other Haystack component.

from haystack import Pipeline
from haystack.components.converters import TextFileToDocument
from haystack.components.writers import DocumentWriter
from haystack.dataclasses.byte_stream import ByteStream
from haystack.document_stores.in_memory import InMemoryDocumentStore

from deepl_haystack import DeepLDocumentTranslator

document_store = InMemoryDocumentStore()

pipeline = Pipeline()
pipeline.add_component(instance=TextFileToDocument(), name="converter")
pipeline.add_component(
    instance=DeepLDocumentTranslator(target_lang="ES"),
    name="translator",
)
pipeline.add_component(
    instance=DocumentWriter(document_store=document_store), name="document_store"
)
pipeline.connect("converter", "translator")
pipeline.connect("translator", "document_store")
pipeline.run({"converter": {"sources": [ByteStream.from_string("Hello world!")]}})
print(document_store.filter_documents())
# [Document(id=..., content: '¡Hola, mundo!', meta: {'source_lang': 'EN', 'language': 'ES'})]

License

deepl-haystack is distributed under the terms of the MIT license.