1*b1cdbd2cSJim Jagielski/**************************************************************
2*b1cdbd2cSJim Jagielski *
3*b1cdbd2cSJim Jagielski * Licensed to the Apache Software Foundation (ASF) under one
4*b1cdbd2cSJim Jagielski * or more contributor license agreements.  See the NOTICE file
5*b1cdbd2cSJim Jagielski * distributed with this work for additional information
6*b1cdbd2cSJim Jagielski * regarding copyright ownership.  The ASF licenses this file
7*b1cdbd2cSJim Jagielski * to you under the Apache License, Version 2.0 (the
8*b1cdbd2cSJim Jagielski * "License"); you may not use this file except in compliance
9*b1cdbd2cSJim Jagielski * with the License.  You may obtain a copy of the License at
10*b1cdbd2cSJim Jagielski *
11*b1cdbd2cSJim Jagielski *   http://www.apache.org/licenses/LICENSE-2.0
12*b1cdbd2cSJim Jagielski *
13*b1cdbd2cSJim Jagielski * Unless required by applicable law or agreed to in writing,
14*b1cdbd2cSJim Jagielski * software distributed under the License is distributed on an
15*b1cdbd2cSJim Jagielski * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16*b1cdbd2cSJim Jagielski * KIND, either express or implied.  See the License for the
17*b1cdbd2cSJim Jagielski * specific language governing permissions and limitations
18*b1cdbd2cSJim Jagielski * under the License.
19*b1cdbd2cSJim Jagielski *
20*b1cdbd2cSJim Jagielski *************************************************************/
21*b1cdbd2cSJim Jagielski
22*b1cdbd2cSJim Jagielski
23*b1cdbd2cSJim Jagielski
24*b1cdbd2cSJim Jagielski#import "OOoContentDataParser.h"
25*b1cdbd2cSJim Jagielski
26*b1cdbd2cSJim Jagielski@implementation OOoContentDataParser
27*b1cdbd2cSJim Jagielski
28*b1cdbd2cSJim Jagielski- (id)init
29*b1cdbd2cSJim Jagielski{
30*b1cdbd2cSJim Jagielski    if ((self = [super init]) != nil) {
31*b1cdbd2cSJim Jagielski        shouldReadCharacters = NO;
32*b1cdbd2cSJim Jagielski        textContent = nil;
33*b1cdbd2cSJim Jagielski        runningTextContent = nil;
34*b1cdbd2cSJim Jagielski
35*b1cdbd2cSJim Jagielski        return self;
36*b1cdbd2cSJim Jagielski    }
37*b1cdbd2cSJim Jagielski
38*b1cdbd2cSJim Jagielski    return nil;
39*b1cdbd2cSJim Jagielski}
40*b1cdbd2cSJim Jagielski
41*b1cdbd2cSJim Jagielski- (void)parseXML:(NSData*)data intoDictionary:(NSMutableDictionary*)dict
42*b1cdbd2cSJim Jagielski{
43*b1cdbd2cSJim Jagielski    mdiValues = dict;
44*b1cdbd2cSJim Jagielski
45*b1cdbd2cSJim Jagielski    //NSLog(@"data: %@ %d", data, [data length]);
46*b1cdbd2cSJim Jagielski
47*b1cdbd2cSJim Jagielski    //init parser settings
48*b1cdbd2cSJim Jagielski    shouldReadCharacters = NO;
49*b1cdbd2cSJim Jagielski
50*b1cdbd2cSJim Jagielski    NSXMLParser *parser = [[NSXMLParser alloc] initWithData:data];
51*b1cdbd2cSJim Jagielski
52*b1cdbd2cSJim Jagielski    [parser setDelegate:self];
53*b1cdbd2cSJim Jagielski    [parser setShouldResolveExternalEntities:NO];
54*b1cdbd2cSJim Jagielski    [parser parse];
55*b1cdbd2cSJim Jagielski
56*b1cdbd2cSJim Jagielski    [parser release];
57*b1cdbd2cSJim Jagielski
58*b1cdbd2cSJim Jagielski    //NSLog(@"finished");
59*b1cdbd2cSJim Jagielski}
60*b1cdbd2cSJim Jagielski
61*b1cdbd2cSJim Jagielski- (void)parser:(NSXMLParser *)parser didStartElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qualifiedName attributes:(NSDictionary *)attributeDict
62*b1cdbd2cSJim Jagielski{
63*b1cdbd2cSJim Jagielski    // all text content is stored inside <text:p> elements
64*b1cdbd2cSJim Jagielski    if ([elementName isEqualToString:@"text:p"] == YES) {
65*b1cdbd2cSJim Jagielski        runningTextContent = [NSMutableString new];
66*b1cdbd2cSJim Jagielski        shouldReadCharacters = YES;
67*b1cdbd2cSJim Jagielski        //NSLog(@"start");
68*b1cdbd2cSJim Jagielski    } else {
69*b1cdbd2cSJim Jagielski        return;
70*b1cdbd2cSJim Jagielski    }
71*b1cdbd2cSJim Jagielski
72*b1cdbd2cSJim Jagielski    //NSLog(@"start element %@", elementName);
73*b1cdbd2cSJim Jagielski}
74*b1cdbd2cSJim Jagielski
75*b1cdbd2cSJim Jagielski- (void)parser:(NSXMLParser *)parser didEndElement:(NSString *)elementName namespaceURI:(NSString *)namespaceURI qualifiedName:(NSString *)qName
76*b1cdbd2cSJim Jagielski{
77*b1cdbd2cSJim Jagielski    if (shouldReadCharacters == TRUE) {
78*b1cdbd2cSJim Jagielski        if (textContent == nil) {
79*b1cdbd2cSJim Jagielski            textContent = [NSMutableString new];
80*b1cdbd2cSJim Jagielski        } else if ([runningTextContent isEqualToString:@""] == NO) {
81*b1cdbd2cSJim Jagielski            // separate by whitespace
82*b1cdbd2cSJim Jagielski            [textContent appendString:@" "];
83*b1cdbd2cSJim Jagielski        }
84*b1cdbd2cSJim Jagielski        //NSLog(@"end");
85*b1cdbd2cSJim Jagielski
86*b1cdbd2cSJim Jagielski        [textContent appendString:[NSString stringWithString:runningTextContent]];
87*b1cdbd2cSJim Jagielski        [runningTextContent release];
88*b1cdbd2cSJim Jagielski    }
89*b1cdbd2cSJim Jagielski    shouldReadCharacters = NO;
90*b1cdbd2cSJim Jagielski}
91*b1cdbd2cSJim Jagielski
92*b1cdbd2cSJim Jagielski- (void)parser:(NSXMLParser *)parser foundCharacters:(NSString *)string
93*b1cdbd2cSJim Jagielski{
94*b1cdbd2cSJim Jagielski    if (shouldReadCharacters == NO) {
95*b1cdbd2cSJim Jagielski        return;
96*b1cdbd2cSJim Jagielski    }
97*b1cdbd2cSJim Jagielski    //NSLog(string);
98*b1cdbd2cSJim Jagielski
99*b1cdbd2cSJim Jagielski    [runningTextContent appendString:string];
100*b1cdbd2cSJim Jagielski
101*b1cdbd2cSJim Jagielski    //NSLog(@"currentElement: %@", currentElement);
102*b1cdbd2cSJim Jagielski    //NSLog(@"read: %@", string);
103*b1cdbd2cSJim Jagielski
104*b1cdbd2cSJim Jagielski}
105*b1cdbd2cSJim Jagielski
106*b1cdbd2cSJim Jagielski- (void)parser:(NSXMLParser *)parser parseErrorOccurred:(NSError *)parseError
107*b1cdbd2cSJim Jagielski{
108*b1cdbd2cSJim Jagielski    //NSLog(@"parsing finished with error");
109*b1cdbd2cSJim Jagielski    NSLog([NSString stringWithFormat:@"An error occured parsing the document. (Error %i, Description: %@, Line: %i, Column: %i)", [parseError code],
110*b1cdbd2cSJim Jagielski        [[parser parserError] localizedDescription], [parser lineNumber],
111*b1cdbd2cSJim Jagielski        [parser columnNumber]]);
112*b1cdbd2cSJim Jagielski
113*b1cdbd2cSJim Jagielski    if (runningTextContent != nil) {
114*b1cdbd2cSJim Jagielski        [runningTextContent release];
115*b1cdbd2cSJim Jagielski    }
116*b1cdbd2cSJim Jagielski    if (textContent != nil) {
117*b1cdbd2cSJim Jagielski        [textContent release];
118*b1cdbd2cSJim Jagielski    }
119*b1cdbd2cSJim Jagielski}
120*b1cdbd2cSJim Jagielski
121*b1cdbd2cSJim Jagielski- (void)parserDidEndDocument:(NSXMLParser *)parser
122*b1cdbd2cSJim Jagielski{
123*b1cdbd2cSJim Jagielski    if (textContent != nil && [textContent length] > 0) {
124*b1cdbd2cSJim Jagielski        [mdiValues setObject:[NSString stringWithString:textContent] forKey:(NSString*)kMDItemTextContent];
125*b1cdbd2cSJim Jagielski        [textContent release];
126*b1cdbd2cSJim Jagielski    }
127*b1cdbd2cSJim Jagielski}
128*b1cdbd2cSJim Jagielski
129*b1cdbd2cSJim Jagielski@end
130