Add free books to your eReader: Formatting tips
======
In my recent article, [A handy way to add free books to your eReader][1], I explained how to convert the plaintext indexes at [Project Gutenberg][2] to HTML and then EPUBs. But as one commenter noted, there is a problem in older indexes, where individual books are not always separated by an extra newline character.
I saw quite vividly the extent of the problem when I was working on the index for 2007, where you see things like this:
```
Audio: The General Epistle of James 22931
Audio: The Epistle to the Hebrews 22930
Audio: The Epistle of Philemon 22929
Sacrifice, by Stephen French Whitman 22928
The Atlantic Monthly, Volume 18, No. 105, July 1866, by Various 22927
The Continental Monthly, Vol. 6, No 3, September 1864, by Various 22926
The Story of Young Abraham Lincoln, by Wayne Whipple 22925
Pathfinder, by Alan Douglas 22924
[Subtitle: or, The Missing Tenderfoot]
Pieni helmivyo, by Various 22923
[Subtitle: Suomen runoja koulunuorisolle]
[Editor: J. Waananen] [Language: Finnish]
The Posy Ring, by Various 22922
```
My first reaction was, "Well, how bad can it be to just add newlines where needed?" The answer: "Really bad." After days of working this way and stopping only when the cramps in my hand became too annoying, I decided to revisit the problem. I thought I might need to do multiple Find-Replace passes, maybe keyed on things like `[Language: Finnish] `or maybe just the `]` bracket, but this seemed almost as laborious as the manual method.
Then I noticed a particular feature: For most instances where a newline was needed, a newline character was immediately followed by the capital letter of the next title. For lines where there was still more information about the book, the newline was followed by spaces. So I tried this: In the Find text box in [KWrite][3] (remember, we’re using regex), I put:
```
(\n[A-Z])
```
and in Replace, I put:
```
\n\1
```
For every match inside the parentheses, I added a preceding newline, retaining whatever the capital letter was. This worked extremely well. The few instances where it failed involved book titles beginning with a number or with quotes. I fixed these manually, but I could have put this:
```
(\n[0-9])
```
In Find and run Replace All again. Later, I also tried it with the quotes—this requires a backslash, like this:
```
(\n\”) and (\n\’)
```
One side effect is that a number of the listings were separated by three newline characters. Not an issue for XHTML, but easily fixed by putting in Find:
```
\n\n\n
```
and in Replace:
```
\n\n
```
To review the process with the new features:
1. Remove the preamble and other text you don’t want
2. Add extra newlines with the method shown above
3. Convert three consecutive newlines to two (optional)
4. Add the appropriate HTML tags at the beginning and end
5. Create the links based on finding `(\d\d\d\d\d)`, replacing with `\1`
6. Add paragraph tags by finding `\n\n` and replacing with `\n\n`
7. Add a `
` just before the ` ` tag at the end
8. Fix the headers, preceding each with `