BUGFIXES: handling filenames, reading float values, and event annotation

[dominikwelke]

fixed some bugs in the reader:

• handle . characters in file paths
• handle floating point numbers in header fields
• fix: handle fFrequency=0 case before adding fFrequency to datainfo dict
• fix: wrong lines selected in annotation slice

[dominikwelke]

hi @drammock
can we bump the version again?
will need this to progress mne-tools/mne-python#13176

[drammock]

I can review tomorrow (Thursday). No objection to bumping version after that

[dominikwelke]

thanks! no pressure :)

[drammock]

Since there's nothing else after this elif block, I think it's a bit cleaner to do the try-except method (even if it's a tad slower). This isn't a performance bottleneck (the loop only runs 16 times). Also the .replace(...).isnumeric() method won't handle negative numbers (probably that won't matter given the names of the fields above, they all appear to be strictly positive quantities, but better safe than sorry?)

Suggested change

	elif text.replace('.','',1).isnumeric() : # BUG (float not recognized) FIXED
	a[i] = float(text) # assign if it was a number
	try:
	a[i] = float(text)
	except ValueError:
	pass

[drammock]

Just so I'm clear, the bug here was adjusting fFrequency after it was entered into the datainfo dict, right? If so, let's avoid changes to this line, as there wasn't anything broken about it before, and we should try to minimize the diff with upstream so it's easier to pull in any upstream changes down the road.

Suggested change

	if fFrequency == 0. or fSampleTime != 0.: # BUG (order) FIXED
	if fFrequency == 0 or fSampleTime != 0:

[drammock]

[nitpick] I'm unsure if the comment will still be helpful in 6 months. On the assumption that we only make changes to the reader if there was a demonstrable bug (i.e., no new features), the diff will tell us where the bugfixes were, and "does not read correct lines" seems (to me at least) inferrable from the change of indices in the slice. Feel free to disregard though, if you disagree.

Suggested change

	annotations = contents[tixstart+1:tixstop].splitlines() # BUG (does not read correct lines) FIXED
	annotations = contents[tixstart+1:tixstop].splitlines()

[dominikwelke]

thanks for the suggestions @drammock - all make sense, all accepted

[drammock]

I've updated the PR's description to describe the 4 bugs addressed here. That should hopefully make it easier in future if we need to revisit / rebase / whatever. Thanks @dominikwelke