I need help!!!

[crutledge]

I'm at the verge of gibbering. Since all of the really smart folks hang out here, I know someone has the answer. The book I'm working on has 72 blocks similar the the following:

<div class="stanza">
  <p>“‘Tell me, my old friend, tell me why</p>
  <p>You sit and softly laugh by yourself.’</p>
  <p>‘It is because I am repeating to myself,</p>
  <p>Write! write</p>
  <p>Of the valiant strength,</p>
  <p>The calm, brave bearing</p>
  <p>Of the sons of the sea.’”</p>
</div>

I need to get rid of the <p>...</p> tags and replace with the following:

<div class="stanza">
  “‘Tell me, my old friend, tell me why<br />
  You sit and softly laugh by yourself.’<br />
  ‘It is because I am repeating to myself,<br />
  Write! write<br />
  Of the valiant strength,<br />
  The calm, brave bearing<br />
  Of the sons of the sea.’”<br />
</div>

I've been through every manual and web reference I can find and nothing I've tried has worked. Nothing comes close to finding a series of items between two markers. I know that anything can be done with regex's if you are only smart enough or can frame the right question to google.

[Jellby]

I would actually prefer the former markup, as it lets you control how each line wraps in narrow screens (typically, I'd have some largish negative indent; ideally, the wrapped part would be right-aligned)

I see you have spaces before the <p>, is that the case in every stanza and only inside stanzas?, then you could try something like:

search: "^ <p>(.*)</p>"
replace: " \1<br/>"

Otherwise, it gets complicated, I'd probably proceed stepwise: search for "<div class="stanza">, whatever, </p>\n<p>, whatever, </div>", where "whatever" stands for "any character, including newlines, multiple times, as few as possible", and repeat as needed.

[Fabe]

I would do it in two steps.
FIND </p>
REPLACE WITH <br />

And Then

FIND <p>
REPLACE WITH [blank]

This seems too easy, what an I missing? - Fabe

[crutledge]

Originally Posted by Jellby

I would actually prefer the former markup, as it lets you control how each line wraps in narrow screens (typically, I'd have some largish negative indent; ideally, the wrapped part would be right-aligned)

I see you have spaces before the <p>, is that the case in every stanza and only inside stanzas?, then you could try something like:

search: "^ <p>(.*)</p>"
replace: " \1<br/>"

Otherwise, it gets complicated, I'd probably proceed stepwise: search for "<div class="stanza">, whatever, </p>\n<p>, whatever, </div>", where "whatever" stands for "any character, including newlines, multiple times, as few as possible", and repeat as needed.

Please remember that there are many <p>...</p> that are not part of the stanza. What you describe would change every <p>...</p> in the file. The changes must be limited to those <p>...</p> between <div class="stanza"> and </div>

[Perkin]

Here's how I'd do it.
place cursor at END of file

Match case / Minimal / Regex / Direction Up
Search for

<div class="stanza">(.+)<p>(.+)</p>

Replace with

<div class="stanza">\1\2<br />

Then repeat 'replace' as needed for each line of current stanza, then after replacing the last line, do a 'find next'

[Fabe]

I'd go with my original Find/Replace settings, but go a line at a time with the mouse on REPLACE and a finger on the Enter key for Find Next.

Or I'd become a potato farmer. - Fabe

[Fabe]

????????

[KevinH]

Hi,

If you are okay running python you could try something like the following on each file in the epub that had stanza's in it (after unzipping it of course)

#!/usr/bin/env python                                                                                   
import sys
import os
def main(argv=sys.argv):
    if len(argv) != 3:
        print "syntax is:  python fixme.py INPUTFILE OUTPUTFILE"
        return 1
    infile = argv[1]
    outfile = argv[2]
    if not os.path.exists(infile):
        print "input file was not found"
        return 1

    data = file(infile,'rb').read()
    of = file(outfile,'wb')
    lines = data.split(os.linesep)
    instanza = False
    res = ''
    for line in lines:
        if line.find('<div class="stanza">') != -1:
            instanza = True
        if instanza :
            line = line.replace('<p>','')
            line = line.replace('</p>','<br />')
            if line.find('</div>') != -1:
                instanza = False
        line += os.linesep
        res += line
    of.write(res)
    of.close()

if __name__ == '__main__':
    sys.exit(main())

my test.html is:

<html>
<body>
<p>do not change me</p>
<div class="stanza">
  <p>“‘Tell me, my old friend, tell me why</p>
  <p>You sit and softly laugh by yourself.’</p>
  <p>‘It is because I am repeating to myself,</p>
  <p>Write! write</p>
  <p>Of the valiant strength,</p>
  <p>The calm, brave bearing</p>
  <p>Of the sons of the sea.’”</p>
</div>
<p>do not change me either</p>
</body>
</html>

And running:

python fixme.py test.html test_fixed.html

gives the following for test_fixed.html

<html>
<body>
<p>do not change me</p>
<div class="stanza">
  “‘Tell me, my old friend, tell me why<br />
  You sit and softly laugh by yourself.’<br />
  ‘It is because I am repeating to myself,<br />
  Write! write<br />
  Of the valiant strength,<br />
  The calm, brave bearing<br />
  Of the sons of the sea.’”<br />
</div>
<p>do not change me either</p>
</body>
</html>

This is of course only a simple test and the pasting of it here may cause problems if it messes up spacing and things, but a similar approach can be used for almost any mass change you want.

If you are desperate enough to want to give it a try, pm me with your e-mail and I will send you the python file.

If you are macosx, linux or unix based, you can use 'sed" to do this or awk or almost any simple scripting language like python (above) or perl or php, etc.

KevinH

[crutledge]

Thanks to all for the ideas which made me think harder and develop a solution. Hearing other ideas is a stimulant.

I finally used GREP. This allowed a logical sectioning of the file (which I just discovered) within the stanza tags. Once this was done, and the lines I was interested in were isolated from the rest of the code, a regex find and replace on the lines in the sections worked like a champ. Trying manually to do 75 stanzas X 5 to 10 lines per stanza would mind numbing.

I need to dig deeper into this GREP.

BTW, the file looks great>

[st_albert]

so, how did you use grep to solve this problem?

and what OS?

ETA: my textmod fu is perl >> sed > awk, but that's because I'm too frikkin lazy to learn much at my age. But occasionally I do get curious. now's your chance!

[Jellby]

Originally Posted by crutledge

Please remember that there are many <p>...</p> that are not part of the stanza. What you describe would change every <p>...</p> in the file. The changes must be limited to those <p>...</p> between <div class="stanza"> and </div>

Actually, the first solution only replaces <p>...</p> which are preceded by spaces.

The second was intended to find only <p>...</p> inside stanzas, but was flawed. In the "whatever" you'd have to include the restriction "does not contain '</div>'", since that restriction in often hard to express in regexp, I sometimes use this trick:

1 Replace "<div class="stanza">...</div>" with "¬...|" (¬ and | are unused characters in the rest of the file)
2 Find and replace "¬...<p>...</p>...|" where the "..." includes "[^|]", i.e., any character but |, this keeps the matching inside the ¬...| block.
3 Convert ¬ and | back into the <div class="stanza"> and </div>.

[crutledge]

Originally Posted by st_albert

so, how did you use grep to solve this problem?

and what OS?

ETA: my textmod fu is perl >> sed > awk, but that's because I'm too frikkin lazy to learn much at my age. But occasionally I do get curious. now's your chance!

I am not a GREP expert. I use PowerGREP from JGSOFT. I selected "FILE SECTIONING" and entered the following in the "SECTION SEARCH" pane.

<div class="stanza">.*?</div>

When I selected "PREVIEW" all of the lines within each stanza were displayed. Black Magic.

Below the "SECTION SEARCH" pane is a SEARCH AND REPLACE pane which allows modification of the data found.

So simple. Clever people these JSOFT folks. I've used their tools for years. I just have to be smart enough to use them. But they do have a great forum setup for help and information with a large following so I can ask my dumb questions.