#!/usr/bin/python

# sd2h - ScriptDirToHtml
# Makes HTML listings and a tarball for a directory of scripts.
#
# The listings are sdindex.html and atoz.html, chronologically and
# alphabetically sorted respectively.  They depend on
# http://www.cv.nrao.edu/~rreid/software/extractcomment and
# http://www.cv.nrao.edu/~rreid/software/style.css
#
# If run in <dir>, the tarball is <dir>/<dir>.tar.gz.
# All output gets written over if it is already present!
#
# Changes:
#   2009/12/09: Translated from bash to python, making it 2x faster by only
#               extracting the comments once.
#
#               The pages now use CSS (get the stylesheet above!) and both they
#               and the tarball are more careful to exclude unwanted files.
#   
# It is GNU software, and based on DirToHtml.  Please send us any
# customizations you have made so they can be incorporated in
# future releases.
# Have fun, Tomas IV., brunclik@mail.com, http://tom4.hyperlink.cz (DirToHtml)
# Rob Reid, http://www.cv.nrao.edu/~rreid/ (added script comment extraction)

from glob import glob
import os
import re
import shutil
import stat
import sys
import time

version = sys.argv[0] + " v2.1e"
verblurb = """
    """ + version + """
    ScriptDirToHTML, (G) Rob Reid 2009, based on
    DirToHtml, (G) Tomas Brunclik 2001

    Use under terms of GNU public licence.
"""

help = "Use: " + os.path.basename(sys.argv[0]) + """ [-hv | ["]<files>["]]

Generates html indices and a tarball of the current directory:
  sdindex.html: Sorted by last modification time.
  atoz.html:    Alphabetically sorted.
  $cwd.tar.gz:  The tarball.

 -h, --help		Show this help screen
 -v, --version		Version info
"""

fmask = ''
nargs = len(sys.argv)
if nargs > 1:
    if sys.argv[1][0] == '-':
        for i in xrange(1, min(nargs, 3)):
            if sys.argv[i].find('v') > 0:
                print verblurb
            if sys.argv[i].find('h') > 0:
                print help
        sys.exit(0)
    else:
        fmask = ' '.join(sys.argv[1:])

owd = os.getcwd()
scriptdir = os.path.split(owd)[-1]
for f in ('sdindex.html', 'atoz.html', scriptdir + '.tar.gz'):
    if os.path.isfile(f):
        os.remove(f)

tarball = scriptdir + '.tar.gz'

def getscripts(fmask=''):
    """
    Returns the filenames to make entries for, sorted by time.
    """
    lister = os.popen('ls -1t ' + fmask)
    fmask = []
    skippables = ('atoz.html', 'sdindex.html', 'style.css', 'index.html',
                  tarball)
    for s in lister:
        s = s.strip()
        if s[-1] != '~' and s not in skippables:
            fmask.append(s)
    lister.close()
    return fmask

fmask = getscripts(fmask)

def make_tarball(dirname, files, dest,
                 specialfiles=('COPYING', 'INSTALL', 'LICENSE', 'README')):
    os.mkdir(dirname)
    owd = os.getcwd()
    os.chdir(dirname)
    files = set(files)
    files.update(specialfiles)
    for f in files:
        absf = os.path.join(owd, f)
        if os.path.isfile(absf):
            try:
                os.symlink(absf, f)
            except Exception, e:
                print "Error", e, "symlinking", f, "to", absf
                #print "files =", ', '.join(files)
                sys.exit(1)
        else:
            print absf, "not found for", dest
    os.chdir(owd)
    os.system("tar chfz " + dest + ' ' + dirname)
    shutil.rmtree(dirname)

make_tarball(scriptdir, fmask, tarball)

def kilobytes(bytes):
    """
    Given an integer number of bytes, returns it as a string in kilobytes.
    """
    return "%.1fkb" % (bytes / 1024.0)
    

def getentries(fmask):
    entries = {}
    for script in fmask:
    #    echo <H2><A HREF=\"${i}\">${i}</A></H2> >> $page
    #    echo <font size=+2><b><A HREF=\"${i}\">${i}</A></b></font> \&nbsp\;\&nbsp\;\&nbsp\;\&nbsp\; "Last modified " `ls -o $i | awk '{printf "%s %s %s, %s bytes\n", $5, $6, $7, $4}'` <br> >> $page
        entries[script] = '''
<div class="entry">
  <div class="leftHeader">
    <A HREF="''' + script + '">' + script + '''</A>
  </div>
'''
        fstat = os.stat(script)
        #blmod=$(ls -o --full-time $i | sed 's/  */ /g' | cut -d ' ' -f 4-9)
        entries[script] += '''  <div class="rightHeader">
    ''' + kilobytes(fstat[stat.ST_SIZE]) + ', last modified '
        entries[script] += time.ctime(fstat[stat.ST_MTIME]) + '''
  </div>
  <div class="clear"></div>
  <div class="entryContent">
'''
        sfile = open(script)
        bangcheck = (sfile.readline()[:2] == '#!')
        
        fext = script.split('.')[-1].lower()

        if fext[:3] == 'htm':
            sfile.seek(0)
            for line in sfile:
                lcline = line.lower()
                if '<title>' in lcline:
                    titlestart = lcline.index('<title>') + 7
                    titleend = lcline.index('</title>')
                    entries[script] += line[titlestart:titleend]
                    break
        elif bangcheck or (fext == "py"):
        # Allows proportional fonts, but wrecks things that depend on
        # monospace alignment.
        #	echo <P>  >> $page
        #	extractcomment < $i | sed 's/^$/<p>/;s/$/<br>/' >> $page
        #	echo <P>  >> $page

            # Preserves monospaced arrangement.
            entries[script] += os.popen('extractcomment -h < ' + script).read()
        elif ('.tar' in script) or (script[-4:] == '.tgz') or (script[-4:]
                                                               == '.tbz'):
            tarcmd = 'tar tf'
            if script[-2] == 'g':
                tarcmd += 'z'
            elif 'bz' in script[-3:]:
                tarcmd += 'j'
            listing = os.popen(tarcmd + ' ' + script).read().split("\n")
            entries[script] += "  <ul>\n"
            for thing in listing:
                if not (re.match(r'.*/.*/', thing) or thing == ''):
                    entries[script] += '    <li>' + thing + "\n"
            entries[script] += "  </ul>\n"
        else:
            magic = os.popen('file ' + script).read()
            entries[script] += magic[len(script) + 2:]
        sfile.close()
        entries[script] += '''  </div>
  </div>
</div>
'''
    return entries

entries = getentries(fmask)

def makepage(sortord, fmask, entries):
    if sortord == "time":
        page = open('sdindex.html', 'w')
    else:
        fmask.sort()
        page = open('atoz.html', 'w')

    if scriptdir.lower() != 'scripts':
        title = scriptdir + ' scripts'
    else:
        title = 'Scripts'
        
    # HTML header
    header = '''<html>
<head>
<link rel="stylesheet" href="style.css" type="text/css">
<title>''' + title + '''</title>
<META name="generator" content="''' + version + '''">
</head>'''

    body = '''
<body>
'''
    # <H2>List of Files</H2>
    body += '<H1>' + title + '''</H1>
<p>
These scripts are available under the <a href="COPYING">GNU General Public License</a>.
</p>
<p>
Note that there is no warranty!  Although I have put some effort into making
them generally useful, they were mainly written to help with a task at hand.
So they may not yet work for all cases, and you may have some fixing to do.
Please send me any fixes or enhancements.
</p>
<hr>
<p><center>
<a href="''' + tarball + '''">The whole shebang ('''

    body += kilobytes(os.stat(tarball)[stat.ST_SIZE]) + ''')</a>
</p>
<hr>
<p><center>
'''
    
    if sortord == "time":
        body += '<a href="atoz.html">Sort alphabetically.</a>'
    else:
        body += '<a href="sdindex.html">Sort by last modification time.</a>'

    body += '''
</center><p>
<hr>
<p>
<table width=100% bgcolor=white cellpadding=4 rules=0 border=0>
'''

    # Hyperlink the files
    for script in fmask:
        body += entries[script]

    # Close the HTML.
    body += '''</table>
<P>
<HR>
<center>
Page generated by <A href="http://www.cv.nrao.edu/~rreid/software/sd2h">s2dh</A>,
based on <A href="http://tom4.hyperlink.cz/software.php?mylang=en&style=main#d2h">DirToHtml</A>,
a GNU generator of html directory listings.
</center>
</body>
</html>
'''
    page.write(header + body)
    page.close()

makepage("time", fmask, entries)
makepage("alpha", fmask, entries)