Converting HTML to Docbook SGML/XML Using html2db

Introduction

html2db is a small utility to convert HTML to Docbook SGML/XML. It uses TidyLib for parsing the HTML. For information on Docbook, visit Docbook official site

Demo

The following are some of the documents converted using the tool.

NCURSES Introduction in html converted to Docbook SGML
I converted some LDP Documents to docbook. They are here.
VB6 To Tcl Cheat-Sheet converted to docbook on author's request.

Note: I made some manual changes to the LDP documents

Binaries

These binaries are provided for your convenience. Use them at your own risk.

A linux binary compiled with gcc on mandrake 8.2 for i386 architecture.
Windows binary, coming soon.

Instructions for compiling

You need tidy sources for compiling this.

Download the latest tidy source from here (5 July 2005 version)
Unzip it
```
    tar zxvf tidy_src.tgz
```
Download this patch and apply it
```
    pacth -p0 < newpatch
```
Compile tidy
```
    cd build/gmake
    make
```
Download html2db source from here
Unzip it
```
    tar zxvf html2db.tar.gz
```
Edit the Makefile. Change the TIDYDIR to point to tidy source dir
Compile it
```
    make
```

Usage

    Usage: ./html2db [option] <html file>
    Options:
        -dbsgml         Docbook SGML output
        -dbxml          Docbook XML output
        -help           Print Help Message

This produces the output in docbook SGML or XML. The docbook public identifiers are hardcoded for now. In future, I will add options.

The output is far from perfect, but it does most of the dirty work. I will prepare a web page explaining the transformations. It will take some time :-) For the impatient, look in the source.

For more information on docbook, visit Docbook Official Site.

Testing

If you want to test the tool, send me a mail informing me about it and follow these instructions.

Download the binary

Run tidy

   ./html2db <html file> 2>&t.err
   This would put errors in t.err file

Then validate the sgml

   nsgmls -s <sgml file> 2>&s.err
   This would put errors in s.err file

Send me both the error files along with html source and anything interesting you found. If you are sending lot of files use some naming convention like 1.html t1.err s1.err and zip them.
DO NOT send jade output. It's not necessary.