source: gaycko/src/dom/dom.c @ 4716 Tweet

#include "dom.h"
#include <string.h>
#include "helpers/str.h"

gDOM *gaycko_convert_dom(htmlDocPtr doc) {

        gDOM *dom = malloc(sizeof(gDOM));
        dom->root = NULL;  
        htmlNodePtr root = xmlDocGetRootElement(doc);
        gElement **r = NULL;
        gElement *ret = NULL;

        if(root != NULL)
        {
                ret = explore(r, root, 0, NULL);
        }

        dom->root = ret;
        pretty_print(ret);

        return dom;
}


gElement *add_element(htmlNodePtr node) {

        gElement *e;
        e = malloc(sizeof(gElement));
        e->children_count  = 0;
        e->children        = NULL;
    e->attribute_count = 0;
        e->attributes      = NULL;
    
        e->name = (char*)strdup((const char*)node->name);

        if(!strncmp(e->name, "html", 4)) {
                e->type = ELEM_HTML;
        }else if(!strncmp(e->name, "head", 4)) {
                e->type = ELEM_HEAD;
        }else if(!strncmp(e->name, "title", 5)) {
                e->type = ELEM_TITLE;
        }else if(!strncmp(e->name, "meta", 4)) {
                e->type = ELEM_META;
        }else if(!strncmp(e->name, "body", 4)) {
                e->type = ELEM_BODY;
        }else if(!strncmp(e->name, "h1", 2)) {
                e->type = ELEM_H1;
        }else if(!strncmp(e->name, "br", 2) || !strncmp(e->name, "br ", 3)) {
                e->type = ELEM_BR;
        }else if(!strncmp(e->name, "p", 1)) {
                e->type = ELEM_P;
        }else if(!strncmp(e->name, "img", 3)) {
                e->type = ELEM_IMG;
        }else if(!strncmp(e->name, "text", 4)) {
                e->type = ELEM_TEXT;
                e->text = (char*)strdup((const char*)xmlNodeGetContent(node));
            strip_eol(e->text);
            strip_spaces(e->text);
        }else {
                e->type = ELEM_UNKNOW;
                printf("Unknow tag '%s'\n", e->name);   
        }
    
    if(node->properties) {
        for(xmlAttrPtr attr = node->properties; attr != NULL; attr = attr->next) {
            e->attributes = realloc(e->attributes, sizeof(gAttribute));
            e->attributes[e->attribute_count].name = strdup(attr->name);
            e->attributes[e->attribute_count].value = strdup(attr->children->content);
        }
    }
    
        return e;
}


gElement* explore(gElement **elem, htmlNodePtr element, unsigned int level, gElement *parent)
{
        unsigned int c = 0;
        for(htmlNodePtr node = element; node != NULL; node = node->next)
        {
                gElement *child = add_element(node);
                child->level = level;
        
                if(elem) elem  = realloc(elem, sizeof(gElement*)*(c+1));
                else elem = malloc(sizeof(gElement*));


                elem[c] = child;

                c++;

                if(node->type == XML_ELEMENT_NODE)
                {
                        if(node->children != NULL) {
                                explore(child->children,
                                        node->children,
                                        level+1,
                                        child);
                        } else {
                                free(child->children);
                                child->children = NULL;
                        }
                }         
                if(parent) {
                        parent->children = realloc(parent->children, sizeof(gElement*)*(parent->children_count+1));
                        parent->children[parent->children_count] = child;
                        parent->children_count++;
                }
        }
        return *elem;
}


#define LEVEL {unsigned int foo=0; for(foo=0; foo<elem->level; foo++) printf("    ");}

void pretty_print(gElement *elem) {
        if(!elem) {
                printf("elem is %p\n", elem);
                return;
        }

        if(elem->type == ELEM_TEXT) {
                LEVEL printf("%s\n", elem->text);
        } else {
                LEVEL printf("<%s>\n", elem->name);
        }
        
        unsigned int i;
        for(i=0; i < elem->children_count; i++) {
                pretty_print(elem->children[i]);
        }

        if(elem->type == ELEM_TEXT) {

        } else {
                LEVEL printf("</%s>\n", elem->name);
        }
}


#if 0
if(xmlStrcasecmp(node->name, (const xmlChar*)"A") == 0)
{
        for(xmlAttrPtr attr = node->properties; attr != NULL; attr = attr->next)
        {
                if(xmlStrcasecmp(attr->name, (const xmlChar*)"HREF") == 0)
                {
                        printf("Found link <%s>\n", node->children->content);
                }
        }
}
#endif