HTML to LaTeX

This is a program that converts a collection of related HTML files into a single LaTeX file. Such a LaTeX file can be processed into PostScript file. I have done this for my pages related to the never written 7th book of Dune and the TransCoop pages. Both these pages contain a reference to a PostScript file, containing the contents of these root pages and all underlying pages.

Functionality

This tool consist of a single C program, with the name html2tex.c. The program is still under development, and thus still contains bugs. It does some checking of the HTML format, and detects some errors, but it does not verify everything and can still generate incorrect LaTeX output.

The program uses one input file that will be used as a frame work for the generated LaTeX file. The generated file will get the extension .tex. This frame work file has to contain valid LaTeX commands. In the file all lines starting with %html will be interpreted as special lines and replaced by the html2tex.c program.

The following command are recognized:

%html <fn>.html <level>
To include the file: <fn>.html as LaTeX at the given input line. The <level> should be an integer to specify the indentation depth of the headers.
%html -r <URL>
To indicate the URL directory at which the input file can be used. This URL is used to detect if any absolute URL's map to local files.
%html -m <rel-URL> <comp-URL>
To map relative URL's to complete URL's. Normally not needed.
%html -b
To indicate the place where the external URL's should be listed as LaTeX bibitems. If this command is not given (and also not the -b command line option), all external URL's are given as footnotes.

Besides the LaTeX file that is generated, the program will also generate a cross-reference file with the .ref extension, that contains alot of usefull information.

New: If the program is given an input file with the extension .html, it does not generate a LaTeX output file, but only analyse the file, and the files it references (if the -s option is given). A file with the extenstion .ref is generated.

The program has the following command line options:

-i : print info.
-w : print warning (and info).
-s : scan not include files. The program will scan all files that can be reached from the included files, and that are found in the directory (and its sub-directories) of the input file.
-r <URL> : root URL of document. This is needed to find out if a full URL represents a local file.
-b : make bibliograph. If this option is not given, references to external URL will appear in footnotes. The input file should contain a line with %html -b.

Extended examples

For a better understanding of how it works look at:

An example in the directory http://www.cs.utwente.nl/~faase/D7/:
The input file Dune which starts with main.html, and generates: Dune.tex and Dune.ref. The file Dune.ps contains the PostScript file after applying latex and dvips.
An example in the directory http://wwwtranscoop.cs.utwente.nl:8080/:
The input file transcoop which starts with transcoop.html, and generates: transcoop.tex and transcoop.ref. The file transcoop.ps contains the PostScript file after applying latex and dvips.

Known bugs

Output in .ref produces incorrect error messages for URL's.
The use of <H6> as a means of getting small bold font can produce strange results.
No support for IMG, PRE, BLINK, and more tags.
. . . more . . .

Revision history

May 2, 1995:

solved bug: program took first argument as output file name.
references in <ADDRESS> are omitted during output generation.

March 3, 1995:

solved bug in -s option. It now does a complete recursive search.
some extra parsing added. Still alot is missing. No compliance with any standard.
the program can now also except a single HTML file as input. It does not generate any LaTeX output.

How to obtain

If you want to have a try, here is the source of html2tex.c. I can compile it with Sun cc and gcc. No warranties! No version support! But feel free to email me.

Last update: May 2, 1995
Frans Faase

Edited from HTML Tools page: May 9, 1995
Michael Sofka