Table descriptions

After partitioning and chunking, you can have Unstructured generate text-based summaries of detected tables.

This summarization is done by using models offered through these providers:

• GPT-4o, provided through OpenAI.
• Claude 3.5 Sonnet, provided through Anthropic.
• Claude 3.5 Sonnet, provided through Amazon Bedrock.

Here is an example of the output of a detected table using GPT-4o. Note specifically the text field that is added. Line breaks have been inserted here for readability. The output will not contain these line breaks.

{
    "type": "Table",
    "element_id": "5713c0e90194ac7f0f2c60dd614bd24d",
    "text": "The table consists of 6 rows and 7 columns. The columns represent 
        inhibitor concentration (g), bc (V/dec), ba (V/dec), Ecorr (V), icorr 
        (A/cm\u00b2), polarization resistance (\u03a9), and corrosion rate 
        (mm/year). As the inhibitor concentration increases, the corrosion 
        rate generally decreases, indicating the effectiveness of the 
        inhibitor. Notably, the polarization resistance increases with higher 
        inhibitor concentrations, peaking at 6 grams before slightly 
        decreasing. This suggests that the inhibitor is most effective at 
        6 grams, significantly reducing the corrosion rate and increasing 
        polarization resistance. The data provides valuable insights into the 
        optimal concentration of the inhibitor for corrosion prevention.",
    "metadata": {
        "text_as_html": "<table>...<full results omitted for brevity>...</table>",
        "filetype": "application/pdf",
        "languages": [
            "eng"
        ],
        "page_number": 1,
        "image_base64": "/9j...<full results omitted for brevity>...//Z",
        "image_mime_type": "image/jpeg",
        "filename": "7f239e1d4ef3556cc867a4bd321bbc41.pdf",
        "data_source": {}
    }
}

Here are two examples of the descriptions for detected tables. These descriptions are generated with GPT-4o by OpenAI:

The generated table’s summary will overwrite any previous contents in the text field. The table’s original content is available in the image_base64 field.

Any embeddings that are produced after these summaries are generated will be based on the new text field’s contents.

Generate table descriptions

To generate table descriptions, in an Enrichment node in a workflow, specify the following:

You can change a workflow’s table description settings only through Custom workflow settings.

Table summary descriptions are generated only when the Partitioner node in a workflow is set to use the High Res partitioning strategy and the workflow also contains a table description enrichment node.

Setting the Partitioner node to use Auto, VLM, or Fast in a workflow that also contains a table description enrichment node will not produce any table summary descriptions, and it could also cause the workflow to stop running or produce unexpected results.

Select Table, and then choose one of the following provider (and model) combinations to use:

• OpenAI (GPT-4o). Learn more.
• Anthropic (Claude 3.5 Sonnet). Learn more.
• Amazon Bedrock (Claude 3.5 Sonnet). Learn more.

Make sure after you choose the provider and model, that Table Description is also displayed. If Table Description and Table to HTML are both displayed, be sure to select Table Description.