Accepted custom sentiment analysis data formats

If you are trying to import your data into custom sentiment analysis, it has to follow a specific format. If you don't have data to import, you can create your project and use Language Studio to label your documents.

Labels file format

Your Labels file should be in the json format below to be used in importing your labels into a project.

{
    "projectFileVersion": "2023-04-15-preview",
    "stringIndexType": "Utf16CodeUnit",
    "metadata": {
        "projectKind": "CustomTextSentiment",
        "storageInputContainerName": "custom-sentiment-2",
        "projectName": "sa-test",
        "multilingual": false,
        "description": "",
        "language": "en-us"
    },
    "assets": {
        "projectKind": "CustomTextSentiment",
        "documents": [
            {
                "location": "document_1.txt",
                "language": "en-us",
                "sentimentSpans": [
                    {
                        "category": "positive",
                        "offset": 0,
                        "length": 60
                    },
                    {
                        "category": "neutral",
                        "offset": 61,
                        "length": 31
                    }
                ],
                "dataset": "Train"
            },
            {
                "location": "document_2.txt",
                "language": "en-us",
                "sentimentSpans": [
                    {
                        "category": "positive",
                        "offset": 0,
                        "length": 50
                    },
                    {
                        "category": "positive",
                        "offset": 51,
                        "length": 49
                    },
                    {
                        "category": "positive",
                        "offset": 101,
                        "length": 26
                    }
                ],
                "dataset": "Train"
            }
        ]
    }
}

Key Placeholder Value Example
multilingual true A boolean value that enables you to have documents in multiple languages in your dataset and when your model is deployed you can query the model in any supported language (not necessarily included in your training documents). See language support to learn more about multilingual support. true
projectName {PROJECT-NAME} Project name myproject
storageInputContainerName {CONTAINER-NAME} Container name mycontainer
sentimentSpans Array containing all the sentiments and their locations in the document.
documents Array containing all the documents in your project and list of the entities labeled within each document. []
location {DOCUMENT-NAME} The location of the documents in the storage container. Since all the documents are in the root of the container this should be the document name. doc1.txt
dataset {DATASET} The test set to which this file will go to when split before training. Learn more about data splitting here . Possible values for this field are Train and Test. Train
offset The inclusive character position of the start of a sentiment in the text. 0
length The length of the bounding box in terms of UTF16 characters. Training only considers the data in this region. 500
category The sentiment associated with the span of text specified. positive
offset The start position for the entity text. 25
length The length of the entity in terms of UTF16 characters. 20
language {LANGUAGE-CODE} A string specifying the language code for the document used in your project. If your project is a multilingual project, choose the language code of the majority of the documents. See Language support for more information about supported language codes. en-us

Next steps