Refactor main validate function and calls to it

Following @stweil's comment here I noticed some weirdness with the validate function: depending on a number of conditions, the output tuple varies in length (2 or 3 elements) and contents. This makes things unnecessarily complex and hard to understand. We now have (numbers added here for reference):

        if self.isValid is not None:
            return (self.isValid, self.tests, self.characteristics)  #[1]
        elif self.returnOffset is None:
            return (self.tests, self.characteristics)  #[2]
        else:
            return (self.tests, self.characteristics, self.returnOffset) #[3]

Variation 1 is only used once in jpylyzer.py:

isValidJP2, tests, characteristics = bv.BoxValidator("JP2", fileData).validate()

Variation 3 is used only once:

resultTilePart, characteristicsTilePart, offsetNext = BoxValidator(
    marker, self.boxContents, offset).validate() #[3]

Then there's this weird one which looks like 3 but it isn't:
```
resultSOT, characteristicsSOT, tilePartLength = BoxValidator(
        'startOfTile', segContents).validate() 
```
Here a comment in the corresponding function says that "that unlike other marker validation functions this one returns a third result, which is the total tile-part length (psot)!"
All remaining calls are variation 1.

So this is a bit of a mess. A better approach might be to rewrite the main function as:

def validate(self):
    """Generic box validation function"""
    try:
        to_call = getattr(self, "validate_" + self.boxType)
    except AttributeError:
        shared.printWarning(
            "ignoring '" + self.boxType + "' (validator function not yet implemented)")

    else:
        to_call()

    return self

And then use self.tests, self.characteristics, self.isValid, self.offsetNext and self.tilePartLength as needed.

openpreserve / jpylyzer

Refactor main validate function and calls to it #114