In general, image pixel data is not regarded as a standard TraP data product: stated policy is that archiving the large amount of pixel data generated by the sort of surveys that TraP is designed to process is simply impractical.

However, particularly when processing smaller volumes of data, or for pipeline testing and commissioning, it is convenient to be able to store a copy of the pixels that have been processed. The TraP therefore supports storing image data to a MongoDB database. MongoDB is a “document-oriented database”, which provides a convenient, low-overhead way to store and retrieve large volumes of image data.

It is worth noting that the data stored to MongoDB is not a bit-for-bit copy of the input images. We store only the pixel data together with world coordinate system information embedded in a FITS container; other metadata is stripped.

The use of MongoDB within the TraP is completely optional: it is perfectly possibly to run a complete survey without installing it. However, if storing image data is required, you will need to install MongoDB along with the TraP and configure the persistence step appropriately.