/*
 * This file is part of NetSurf, http://netsurf.sourceforge.net/
 * Licensed under the GNU General Public License,
 *                http://www.opensource.org/licenses/gpl-license
 * Copyright 2004 John M Bell <jmb202@ecs.soton.ac.uk>
 */

#include <stdbool.h>
#include <string.h>

#include "libxml/HTMLtree.h"

#include "oslib/osfile.h"

#include "netsurf/utils/config.h"
#include "netsurf/content/content.h"
#include "netsurf/riscos/save_text.h"
#include "netsurf/utils/log.h"
#include "netsurf/utils/utils.h"

#ifdef WITH_TEXT_EXPORT

static void extract_text(xmlDoc *doc);
static void extract_text_from_tree(xmlNode *n);

static char *buffer = 0;
static int output_size = 0;

void save_as_text(struct content *c, char *path) {

        htmlParserCtxtPtr toSave;

	if (c->type != CONTENT_HTML) {
		return;
	}

	/* allocate a buffer the same size as the source
	 * the output is guaranteed to be less than this
	 */
	buffer = xcalloc(c->source_size, sizeof(char));

	toSave = htmlCreateMemoryParserCtxt(c->source_data, c->source_size);
	htmlParseDocument(toSave);

	extract_text(toSave->myDoc);

	if (output_size > 0) {
        	xosfile_save_stamped(path, 0xfff, (byte*)buffer,
	                             (byte*)buffer+output_size);
	}

	xmlFreeDoc(toSave->myDoc);
	htmlFreeParserCtxt(toSave);
	xfree(buffer);
}

void extract_text(xmlDoc *doc)
{
        xmlNode *html;

        /* find the html element */
        for (html = doc->children;
             html!=0 && html->type != XML_ELEMENT_NODE;
             html = html->next)
                ;
        if (html == 0 || strcmp((const char*)html->name, "html") != 0) {
                return;
        }

        extract_text_from_tree(html);
}

void extract_text_from_tree(xmlNode *n)
{
        xmlNode *this;
        char *text;
        int len = 0;
        int need_nl = 0;

        if (n->type == XML_ELEMENT_NODE) {
                if (strcmp(n->name, "dl") == 0 ||
                    strcmp(n->name, "h1") == 0 ||
                    strcmp(n->name, "h2") == 0 ||
                    strcmp(n->name, "h3") == 0 ||
                    strcmp(n->name, "ol") == 0 ||
                    strcmp(n->name, "title") == 0 ||
                    strcmp(n->name, "ul") == 0) {
                        need_nl = 2;
                }
                else if (strcmp(n->name, "applet") == 0 ||
                         strcmp(n->name, "br") == 0 ||
                         strcmp(n->name, "div") == 0 ||
                         strcmp(n->name, "dt") == 0 ||
                         strcmp(n->name, "h4") == 0 ||
                         strcmp(n->name, "h5") == 0 ||
                         strcmp(n->name, "h6") == 0 ||
                         strcmp(n->name, "li") == 0 ||
                         strcmp(n->name, "object") == 0 ||
                         strcmp(n->name, "p") == 0 ||
                         strcmp(n->name, "tr") == 0) {
                        need_nl = 1;
                }
                /* do nothing, we just recurse through these nodes */
        }
        else if (n->type == XML_TEXT_NODE) {
                text = squash_tolat1(n->content);
                len = strlen(text);
                strcat(buffer, text);
                output_size += len;
                xfree(text);
                return;
        }
        else {
                return;
        }

        /* now recurse */
	for (this = n->children; this != 0; this = this->next) {
                extract_text_from_tree(this);
	}

	if (need_nl) {
	        for (len = 0; len != need_nl; len++) {
	                strcat(buffer, "\n");
                        output_size += 1;
                }
	}
}

#endif