Need to change words surrounded by _ to italics

[organized_chaos]

I have a book I purchased from Amazon but there is one problem with it.

Every word that should be in italics is surrounded by _underscores_. I want to edit this book and change the underscores to italics but I'm not sure how. I removed the DRM from the book and converted it to HTMLZ format. My plan was to replace the _ with the html code for italics. The only thing is, one underscore needs to be <i> and the other needs to be </i>. I can replace all underscores with <i> or </i> using the Find and Replace option... but I can't do half and half like I need to.

How can I accomplish my goal without manually replacing everything one at a time?

[Billi]

I think you can't do without a little manual work, but maybe you can limit the number of cases you have to change by hand.
- The beginning underscore has in most cases a blank in front of it - unless it is the beginning of a sentence -, so you should be able to search for "blank+_".
- The closing underscore is in most cases followed by a blank, a comma, a point or a question/exclamation mark - make use of this.

[pdurrant]

Unless you really, really want this book, ask Amazon for a refund and move on to something formatted properly.

If you do really, really want the book, still complain bitterly to Amazon about the poor formatting, but the way to fix it is like this.

Use a text editor that understands regular expressions. Search for something like:

_([a-z ]+?)_

and replace with

<i>\1</i>

And then check for any underscores that haven't been replaced. The first regular expression says "seach for an underscore and then the smallest number of letter and spaces possible until another underscore, and remember the text between the underscores". The second says "replace what was found by <i> followed by the first bit of remembered text, followed by </i>"

Originally Posted by organized_chaos

Every word that should be in italics is surrounded by _underscores_. I want to edit this book and change the underscores to italics but I'm not sure how.
[...]
How can I accomplish my goal without manually replacing everything one at a time?

[organized_chaos]

Originally Posted by pdurrant

If you do really, really want the book, still complain bitterly to Amazon about the poor formatting, but the way to fix it is like this.

Complaint has been registered with Amazon.

Use a text editor that understands regular expressions. Search for something like:

_([a-z ]+?)_

and replace with

<i>\1</i>

And then check for any underscores that haven't been replaced. The first regular expression says "seach for an underscore and then the smallest number of letter and spaces possible until another underscore, and remember the text between the underscores". The second says "replace what was found by <i> followed by the first bit of remembered text, followed by </i>"

This works for single words, but how can I make it work with entire sentences? I'm sorry, I've never needed to edit documents like this before.

For example:

_That's South America down there_, he decided, after rejecting the notion that it might be Africa. They had pretty much the same shape, and it was so hard to remember what Earth's continents looked like when there were so many other worlds. _But that's South America. And so that's North America just above it. The place where I was born._

Should be this:

That's South America down there, he decided, after rejecting the notion that it might be Africa. They had pretty much the same shape, and it was so hard to remember what Earth's continents looked like when there were so many other worlds. But that's South America. And so that's North America just above it. The place where I was born.

[pdurrant]

Originally Posted by organized_chaos

This works for single words, but how can I make it work with entire sentences? I'm sorry, I've never needed to edit documents like this before.

_([a-z ]+?)_

will work for words and spaces. If there might be punctuation involved, you could also add in those characters, e.g.

_([a-z ,.'"-;:]+?)_

or you could go for the slightly more chancy approach of

_([^_]+)_

which searches for an underscore, as many characters as possible that aren't an underscore, and then another underscore.

[organized_chaos]

_([a-z ,.'"-;:]+?)_ worked perfectly for me. I didn't know I could just add additional things to that (like punctuation). If there were numbers in the book would I need to add something like 0-9?

_([a-z ,0-9.'"-;:]+?)_

[pdurrant]

Originally Posted by organized_chaos

_([a-z ,.'"-;:]+?)_ worked perfectly for me. I didn't know I could just add additional things to that (like punctuation). If there were numbers in the book would I need to add something like 0-9?

_([a-z ,0-9.'"-;:]+?)_

Exactly so. the [] give a range of characters that can match. The + after the [] says that there can be as many matches as possible, but that it must match at least one character. I suppose, strictly speaking, the ? is unnecessary, since the match can't include underscores.

[user_none]

Since you converted to HTMLZ I'm assuming you're using calibre. Convert it again with heuristic processing and the italicize common cases option enabled.

[BillSmithBooks]

I think this will work for you...

This is what I do when formatting my plain text books to simple HTML.

1) Search for ^p_ and (space)_ and change them to ^p<i> and (space)<i> -- this finds all the beginning italics and the italics that start at the beginning of a paragraph.

2) Then just do a search for any remaining _ and change to </i>

This should work just fine as long as the author didn't do any duplicate _ _ marks. Or, rather than just do an auto-change, you could search for each change one at a time to make sure it is the right type of change.

[Jellby]

Originally Posted by BillSmithBooks

1) Search for ^p_ and (space)_ and change them to ^p<i> and (space)<i> -- this finds all the beginning italics and the italics that start at the beginning of a paragraph.

and misses italics that start right after a dash, or an opening quote, or an apostrophe, or a parenthesis, or in the middle of the word (which occurs sometimes)...

[organized_chaos]

Just FYI, pdurrant's suggestion solved my problem perfectly.