#!/usr/bin/env python

import json
import os
import re
import sys

# Scan one Cadabra notebook file for an \algorithm{...}{...}
# or \property{...}{...} command and use this to output HTML
# code to produce a table-of-contents of all manual pages.

# FIXME: the regex expressions below should be exchanged for
# something that parses latex properly, in order to deal with
# nested curly bracket situations. At the moment this will fail
# to correctly detect things like \algorithm{foobar}{Something {in} brackets}.

def scan_file(prevcat, dir, filename, ext):
    with open(dir+filename+ext) as jsonfile:
        data = json.load(jsonfile)
        jsonfile.close()

        # is this a multi-algorithm file?
        count=0
        package_name=""
        for cell in data["cells"]:
            if not "cells" in cell:
                continue
            src = cell["cells"][0]["source"]
            src = src.replace('\n',' ')
            m = re.search('\\\\algorithm{(.*)}.*', src)
            if m:
                count+=1
            m = re.search('\\\\property{(.*)}.*', src)
            if m:
                count+=1
            m = re.search('\\\\package{([^}]*)}{([^}]*)}.*', src)
            if m:
                package_name=m.group(1)
                package_desc=m.group(2)
        is_multi = (package_name!="")

        if is_multi:
            print('<tbody class="folding">')
            print('<tr class="package-name"><td colspan=4>'+package_name+'</td></tr>')
            print('<tr class="package-desc"><td colspan=4>'+package_desc+'</td></tr>')
        
        for cell in data["cells"]:
            if not "cells" in cell:
                continue
            src = cell["cells"][0]["source"]
            src = src.replace('\n',' ')
            m = re.search('\\\\algorithm{([^}]*)}{([^}]*)}.*', src)
            if m:
                algo=m.group(1)
                desc=m.group(2)
                cat=algo[0].upper()
                if cat==prevcat:
                    cat=''
                else:
                    prevcat=cat
                if is_multi:
                    cat=''
                m2 = re.search('([^\(]*)\(.*', algo)
                if m2:
                    algo = m2.group(1)
                print('<tr><td>'+cat+'<td></td><td><a href="manual/'+filename+'.html">'+algo+'</a></td>')
                print('<td>'+desc+'</td></tr>')
            m = re.search('\\\\property{([^}]*)}{([^}]*)}', src)
            if m:
                algo=m.group(1)
                desc=m.group(2)
                cat=algo[0].upper()
                if cat==prevcat:
                    cat=''
                else:
                    prevcat=cat
                if is_multi:
                    cat=''
                print('<tr><td>'+cat+'<td></td><td><a href="manual/'+filename+'.html">'+algo+'</a></td>')
                print('<td>'+desc+'</td></tr>')

        if is_multi:
            print('<tr><td colspan=4>&nbsp;</td></tr></tbody>')

    return prevcat


# Scan all manual pages for algorithms.

#lst=set()
#for file in os.listdir(dir):
#    filename, ext = os.path.splitext(file)
#    if ext=='.cnb':
#        lst.add( filename )

algodir=sys.argv[1]+"/"
cat=''
for nb in sorted(sys.argv[2:]):
    cat=scan_file(cat, algodir, nb, '.cnb')