Plugin for tidying ePub files

[martyger]

Originally Posted by CalibUser

The plugin should work now - there was an error in the filename that did not match the XML file in the plugin

I was able to add the plug-in; however I received an "Error Parsing Result XML: Premature end of document" with the following info:

F
atal Python error: Py_Initialize: unable to load the file system codec
ImportError: No module named 'encodings'

I tried this on 5 different files and got the same error message.

[eschwartz]

That appears to be a problem with Python, not the plugin.

Do you have the PYTHONPATH environment variable set to a python2 installation?

[Doitsu]

Originally Posted by eschwartz

That appears to be a problem with Python, not the plugin.

One would think so, but there definitely is a problem with the plugin. I don't get martyger's error message, but the plugin appears to be stuck in an endless loop. The only thing that works is the resize image option.

[gipsy]

Works fine for me. You must point to a text file for word list or it stuck in an endless loop

[Doitsu]

Originally Posted by gipsy

Works fine for me. You must point to a text file for word list or it stuck in an endless loop

I created a WordDictionary.txt dictionary with two words in it in \sigil-ebook\sigil\user_dictionaries and it's still stuck in an endless loop.

Besides, a plugin shouldn't rely on the existence of a file it doesn't create itself.

[gipsy]

Doitsu create and a words.txt and set it from Use word list from the choices when you run the plugin. It's for the new feature that automatically changes words.

[Doitsu]

Originally Posted by gipsy

Doitsu create and a words.txt and set it from Use word list from the choices when you run the plugin. It's for the new feature that automatically changes words.

With this file in place the plugin actually works, however, I'm still convinced that a plugin shouldn't rely on the existence of optional files.

If it requires this file, it should create a dummy file when it's first run.

[CalibUser]

My plugin did include a check for whether or not the file exists instead of creating a dummy file. However, I found a bug that was linked to the json file that stores user preferences. I resolved this problem and I have posted an update in the first thread of this post.

[gipsy]

@CalibUser
I added the "ακούσε|άκουσε" replacement in word list.
But it changes and the "ακούσει" word that is correct with "άκουσε(return)ι".
How can we change the code to find the whole word? Or the whole word followed by punctuation?

Thanks

[CalibUser]

I have amended the plugin so it should remove the erroneous return when "ακούσε|άκουσε" is put in word list.

[gipsy]

Originally Posted by CalibUser

I have amended the plugin so it should remove the erroneous return when "ακούσε|άκουσε" is put in word list.

Calib how you do that? I found another word and i don't want to bother you with single word

Edit: CalibUser i notice that in some greek words it doesnt find the whole word. Like the one in my last post "ακούσει, ακούσεις" that are correct words it changes them to "άκουσε(space)ι, άκουσε(space)ις" that are mispelled.
I also notice that it doesnt fix all the words. I play with the encoding of the words file to fix this.
In what enconding we must have the words file; it fixes the words when i have it in windows-greek. In utf-8 i have problem with the Ά it changed to '

[gipsy]

I finally get the FixP code to work with!

def IsFixP(m):
	"""
	This function examines a hyphenated word to see whether the hyphen is required.
	It is called by a regular expression function (re.sub) in FixCommonErrors()
	It returns the original expression if the non-hyphenated word is not in the dictionary,
	otherwise it returns the word without the hyphen
	"""
	FixP="Π"+m.group(2)
	FixP2=m.group(1)+m.group(2)

	if spell(FixP2):
		return(m.group(0))
	elif spell(FixP):
		print("FixP removed from: ", FixP)
		return ('Π'+m.group(2))
	else:
		return(m.group(0))

and

#Fixes Π in words that are misspelled
		if dictExists == True:
			CorrectText("Π fixes",r"(1\ Ι|ΓΙ|Γΐ|ΙΙ|II|Ι\ Ι|ΓΤ|ΙΊ|Ιί)[ ]?(\w+)(?![^<>]*>)(?!.*<body[^>]*>)", IsFixP)

It bypass the spelling correct words and changes the misspelled.
I think that i have some error in the CorrectText regex.
In the FixP sample that i had in a previously post... It doesn't correct the "Ι Ιήρε" And i have it in the CorrectText as "Ι\ Ι" .

Any thoughts???


EDIT: Found that it doesn't fix if the second part of the word (m.group(2)) is lowercase

[CalibUser]

Originally Posted by gipsy

Edit: CalibUser i notice that in some greek words it doesnt find the whole word. Like the one in my last post "ακούσει, ακούσεις" that are correct words it changes them to "άκουσε(space)ι, άκουσε(space)ις" that are mispelled.
I also notice that it doesnt fix all the words. I play with the encoding of the words file to fix this.
In what enconding we must have the words file; it fixes the words when i have it in windows-greek. In utf-8 i have problem with the Ά it changed to '

@gipsy: I have investigated the problem and I think the attached file may provide a solution. I have not attached it to the first post in the thread as it is just a test file to see whether this solution works. Please would you see whether this version of the plugin resolves the problem of the extra space and whether it works for other Greek words that have caused problems. If it works then I will update the plugin in the first thread with this solution.

Sigil uses the UTF-8 format so please ensure that your Greek word file is in this format (although this does not seem to matter for English words).

[gipsy]

@CalibUser Works fine with the words that I had problems!!!


ps. I have the IncorrectWords.txt saved as UTF-8 at Windows Notepad.

[CalibUser]

Originally Posted by gipsy

@CalibUser Works fine with the words that I had problems!!!

Thanks for testing my update. I am adding another update to the plugin so that apostrophes in the word list will be replaced by a straight or single quote depending on the type of quote that is used in the ePub. When I have managed to get this feature working I will post an update in the first thread.