Separate out saving the file, updating the database, and identifying the
student number.  The last one is implemented such that it can be called outside
of the scanning process (i.e. it only relies on the information that's in the database).

Jamy Mahabier authored 6 years ago and Justin van der Krieken committed 6 years ago

See 7953eca8970e424a2aeac00f5a4bc9f35311a6b6

as per https://github.com/NaturalHistoryMuseum/pylibdmtx/issues/24
it is good to have such test

There is now a test for the id grid, so this comment is not relevant
anymore.

Since the code now identifies the exam on barcodes (e.g. DataMatrices),
it is not necessary for the name to be unique anymore. To allow this,
data storage is now done using the exam ID instead of its name.

An advantage of this is that it can't break anymore if the exam name
contains a character which is not allowed by the filesystem.

The margin parameter is removed to make things simpler for the corner
marker detection. The bottom bar is added to be able to identify the
orientation of a page without depending on the barcode recognition.

Note: Git seems to think two of the changes constitute a rename, but I
actually deleted one file and created another.

against this

Fix bug where the DataMatrix code and student ID grid could end up below
the exam elements in the generated PDF. Also add a test against this
bug.