To improve the processing algorithm

As described, the data extraction method occurs in two stages:

1. To run the old data extraction algorithm
2. To apply the new text recognition stage to read NUM2 metadata.

The algorithm should be refactored such that the metadata type can be detected initially, and then read according to its type. This refactoring and rebuilding of the code base could resolve certain long-standing issues and may improve yield. Furthermore, it can enable the algorithm to be more adaptable to be able to read Alouette-2, ISIS-1 and ISIS-2 ionograms.

No files or tests are named. Start by locating the old extraction algorithm and the NUM2 text-recognition stage, then trace how metadata is currently processed; done means metadata type is detected before extraction and the code can route each type to the appropriate reader, with existing behavior preserved.

Written by the indexing model from the issue text.