Translator options dialog working

So I took my own advice and created a class to represent a dialog option. The author of a Translator codes some of these in his module as global variables, e.g.

import xlate_utils as XU
MAX_LINE_QUERY = XU.Dialog_Item( kind='number', label='Max Line',
                                tooltip='Maximum line width, default is 72',
                                minimum=50, maximum=90, result=72 )
...
OPTIONS_DIALOG = [ MAX_LINE_QUERY, ... ]

The Dialog_Item class definition has only one method, __init__ and it has about 120 lines of code that is mostly validation of the parameters as being the proper types. Errors get logged but some kind of object is always created that the rest of the code can display, if only as a QLabel "Bad Definition Here".

When the Translator is requested, the first thing that happens is to look into the loaded translator module for OPTIONS_DIALOG as a list of Dialog_Item objects. If it's there, a dialog is prepared. Here's one of the test cases in action.

So that all went together nicely. The code to build the dialog, display it, and capture the values set by the user is all quite compact.

Well, that was the fun part. The next part, which is really central to all Translators, is code to parse a document in DP format and reduce it to elemental parts, which can then be fed to any Translator for conversion. "Here's a Chapter Head, convert it. Here's a sub-head, convert it." And so on. But it goes much deeper than that, especially since I want to support nesting of, for example, a Poetry section inside a Footnote or Illustration, and so on. It's non-trivial.

I did this in V1 in a kind of hand-crafted semi-intuitive way. But the only consumer of the output of that parser was my own code for ASCII reflow and HTML conversion. Now I have to look to a generic consumer, a coder who is not me. And I do not want to expose any of the PPQT internals, like the way a document is stored, to the Translator. It needs a very clean, arms-length API. And I would like the parser to have a better foundation in computer science instead of just a big hacky loop with a bunch of state variables, as in V1.

I've been pointed to a partial DP document parser that I need to read. And I have some ideas of how to go at it. But it's a big and critical chunk of work that I need to do next.