InterOp Format

This page describes the binary format for an InterOp file. The formats are broken down into two parts: (1) the header, which is written first, a single time for the file and (2) n-Records, which are written next, many times composing the rest of the file.

An InterOp file may have several versions of the format. This is important because certain values may not be populated in the model. For example, for the CorrectedInt InterOp file, the signal to noise attribute is only populated for version 2 of the format. This parser does not support all versions of the InterOp files, namely versions that were supported by the GA platform.

In 2017, we introduced the concept of "by-cycle" InterOp files, where every InterOp has a separate file for every cycle of data for which the file is produced. Each file has its own header and set of n-Records, but is otherwise similar to before.

The documentation for the model notes when an attribute is only populated by a specific version of the format.

The following are binary formats used only for testing purposes and are not officially supported: