Azure document Translation - how to translate a single file

Question

I've been looking at the Microsoft Azure Documentation for Document Translation:

https://learn.microsoft.com/en-us/azure/ai-services/translator/document-translation/how-to-guides/use-rest-api-programmatically?tabs=csharp#translate-documents

It shows how to translate every document in an Azure Blob Storage "source" container and write them to a "target" container.

What isn't clear to me is, what if there are multiple documents in the "source" container - how can I tell it to work on just one document?

Answer 1

To translate a specific document from the container, you can modify the input payload with - Specifying "storageType": "File" and provide source URL & SAS token for the specific blob/document (not for the container).

Below is sample code snippet for reference:

import requests

endpoint = "https://<NAME-OF-YOUR-RESOURCE>.cognitiveservices.azure.com/translator/text/batch/v1.1"
key =  '<YOUR-KEY>'
path = '/batches'
constructed_url = endpoint + path

payload= {
    "inputs": [
        {
            "storageType": "File",
            "source": {
                "sourceUrl": "https://my.blob.core.windows.net/source-en/source-english.docx?sv=2019-12-12&st=2021-01-26T18%3A30%3A20Z&se=2021-02-05T18%3A30%3A00Z&sr=c&sp=rl&sig=d7PZKyQsIeE6xb%2B1M4Yb56I%2FEEKoNIF65D%2Fs0IFsYcE%3D"
            },
            "targets": [
                {
                    "targetUrl": "https://my.blob.core.windows.net/target/try/Target-Spanish.docx?sv=2019-12-12&st=2021-01-26T18%3A31%3A11Z&se=2021-02-05T18%3A31%3A00Z&sr=c&sp=wl&sig=AgddSzXLXwHKpGHr7wALt2DGQJHCzNFF%2F3L94JHAWZM%3D",
                    "language": "es"
                },
                {
                    "targetUrl": "https://my.blob.core.windows.net/target/try/Target-German.docx?sv=2019-12-12&st=2021-01-26T18%3A31%3A11Z&se=2021-02-05T18%3A31%3A00Z&sr=c&sp=wl&sig=AgddSzXLXwHKpGHr7wALt2DGQJHCzNFF%2F3L94JHAWZM%3D",
                    "language": "de"
                }
            ]
        }
    ]
}
headers = {
  'Ocp-Apim-Subscription-Key': key,
  'Content-Type': 'application/json'
}

response = requests.post(constructed_url, headers=headers, json=payload)

print(f'response status code: {response.status_code}\nresponse status: {response.reason}\nresponse headers: {response.headers}')

For more details, please check this documentation.