Day3 summary hu guangda
完成工作：
今天比较忙，没干什么。poppler基本整理清楚了，开始考虑接口的问题。
一个比较容易实现的想法是，显示采用GCPDF＊库，取词用poppler。我暂时写了一个接口如下，如果有任何不方便的地方或修改建议，请告诉我。
1
/*
2
* PdfAnalyze.h
3
* PaperReader
4
*
5
* Created by Guangda Hu on 11-4-20.
6
* Copyright 2011 Ruse. All rights reserved.
7
*
8
*/
9
10
#ifndef PDFANALYZE_H
11
#define
PDFANALYZE_H
12
13
#include
<
cstdlib
>
14
#include
"
TextOutputDev.h
"
15
#include
"
PDFDoc.h
"
16
#include
"
GooString.h
"
17
#include
"
CharTypes.h
"
18
#include
"
UnicodeMap.h
"
19
#include
"
Dict.h
"
20
21
class
PdfAnalyze {
22
public
:
23
24
/*
25
Construct Function.
26
filename: PDF file to open.
27
ownerPW: owner password.
28
userPW: user password.
29
*/
30
PdfAnalyze(
const
char
*
filename,
const
char
*
ownerPW
=
NULL,
const
char
*
userPW
=
NULL);
31
32
/*
33
Set encoding. No need to call this if use "UTF-8".
34
Available encodings are "Latin1", "ASCII7", "Symbol", "ZapfDingbats", "UTF-8" and "UCS-2".
35
*/
36
void
setEncoding(
const
char
*
encName);
37
38
/*
39
Set end of line. Candidates are "unix" (LF, default), "dos" (CR + LF), "mac" (CR).
40
*/
41
void
setEOL(
const
char
*
eolName);
42
43
/*
44
Return the number of pages.
45
*/
46
int
getNumPages();
47
48
/*
49
Query PDF information dict.
50
key: can be "Title", "Subject", "Keywords", "Author", "Creator", "Producer", "CreationData" or "LastModifiedData".
51
If the dict is not usable, or key is not in dict, NULL is returned.
52
For date information, the return string may start with "D:".
53
*/
54
const
char
*
pdfInfo(
const
char
*
key);
55
56
/*
57
Process a page. This is automatically called in the first query in that page.
58
You can process all pages at the beginning to provide a more smooth interface, if this processing takes long.
59
*/
60
void
touchPage(
int
page);
61
62
/*
63
Query a word and its bounding box by page number and position.
64
The return string is normalized to NKFC (ready for search).
65
See
http://en.wikipedia.org/wiki/Unicode_equivalence
#Normal_forms for full definition.
66
NULL is returned if no word there.
67
*/
68
const
char
*
wordQuery(
int
page,
double
x,
double
y,
double
*
xMin
=
NULL,
double
*
yMin
=
NULL,
69
double
*
xMax
=
NULL,
double
*
yMax
=
NULL);
70
71
/*
72
Return the TextWord object hitted by the positition.
73
Font information, color, raw content can be accessed from TextWord.
74
Do not modify TextWord content, or wordQuery may not work as you expected because I'm using cache.
75
*/
76
TextWord
*
getWord(
int
page,
double
x,
double
y);
77
78
/*
79
Get text from a rectangle.
80
The return is NOT normalized, so there may be ligatures.
81
*/
82
const
char
*
textQuery(
int
page,
double
xMin,
double
yMin,
double
xMax,
double
yMax);
83
84
/*
85
Search text in a page.
86
87
Start & End:
88
If <startAtTop> is true, starts looking at the top of the page;
89
else if <startAtLast> is true, starts looking immediately after the last find result;
90
else starts looking at <xMin>,<yMin>.
91
If <stopAtBottom> is true, stops looking at the bottom of the page;
92
else if <stopAtLast> is true, stops looking just before the last find result;
93
else stops looking at <xMax>,<yMax>.
94
95
Encoding:
96
See
http://en.wikipedia.org/wiki/UTF-8
for the conversion between UTF-32 and UTF-8.
97
Normally use UTF-8 for ASCII based string, but it is converted to UTF-32 implicitly and then call the second function.
98
So also provide a Unicode (UTF-32 no higher than 0x10ffff) version for efficiency.
99
100
Automatically NKFC-normalized.
101
*/
102
bool
findText(
int
page,
const
char
*
s,
int
len,
bool
startAtTop,
bool
stopAtBottom,
bool
startAtLast,
bool
stopAtLast,
103
bool
caseSensitive,
bool
backward,
double
*
xMin,
double
*
yMin,
double
*
xMax,
double
*
yMax);
//
UTF-8
104
bool
findText(
int
page,
const
Unicode
*
s,
int
len,
bool
startAtTop,
bool
stopAtBottom,
bool
startAtLast,
bool
stopAtLast,
105
bool
caseSensitive,
bool
backward,
double
*
xMin,
double
*
yMin,
double
*
xMax,
double
*
yMax);
//
Unicode
106
107
/*
108
Start from the first reference.
109
*/
110
void
resetRef();
111
112
/*
113
Return the next reference, without "[1]" or "[2]".
114
NULL if reaches the last reference.
115
*/
116
const
char
*
nextReference();
117
118
~
PdfAnalyze();
119
120
private
:
121
122
const
double
resolution;
123
124
UnicodeMap
*
uMap;
125
126
PDFDoc
*
doc;
127
128
Object info;
129
Dict
*
infoDict;
130
GooString lastDict;
131
132
int
numPages;
133
TextPage
**
pages;
134
135
const
int
cache_size;
136
int
*
cacheCnt;
137
TextWord
**
wordCache;
138
GooString
*
queryCache;
139
140
GooString
*
lastText;
141
142
int
refPage, refStartPage;
143
TextFlow
*
refFlow,
*
refStartFlow;
144
TextBlock
*
refBlock,
*
refStartBlock;
145
TextLine
*
refLine,
*
refStartLine;
146
GooString lastRef;
147
};
148
149
#endif
/* PDFANALYZE_H */
另外，我不太清楚objective c和c＋＋传递字符串的方法。改成.mm就可以混合编程，但不知道内存分配会不会有问题，尤其是这种指针在objective c里能不能用我完全不知道。（objective c怎么写hello world我都不会。）
明天计划：
写完PdfAnalyze.cpp，脱离对poppler纠结状态。