Uploads

When you upload your data to us, we are first storing it in it's raw form in our uploads storage. Every upload is kept, even if the data is being replaced by a newer upload. They're also stored exactly as you uploaded them, so you can always go back to the original data.

We currently offer the following ways to upload data:

• HTTP API
• Manual File Upload

Each raw upload can be any arbitrary JSON object or in the case of file uploads, json, csv or xls(x), in which case we split up the lines and convert it to json before it reaches the preprocessor. We recommend just uploading whatever format you already have, and transform it into something fitting for text generation in the preprocessor.

You can view your uploads in the uploads section of the pipeline, where we show you the history of all uploads to a pipeline.

Each upload is automatically preprocessed and then stored to the object storage.

Studio vs Cockpit

Differently to Cockpit, an upload will not be immediately matched to a specific object. We save all uploads, regardless if they result in a new object or update an existing one.

More to Come

You will soon have the possibility to replay uploads and send them through preprocessing again.