使用 Quartz 2D 解析 pdf 时获取文本位置

Posted 2023-02-24

【中文标题】使用 Quartz 2D 解析 pdf 时获取文本位置

【英文标题】：Getting text position while parsing pdf with Quartz 2D

【发布时间】：2011-04-07 08:45:07

【问题描述】：

关于pdf解析的另一个问题... 刚刚阅读 PDF Reference version 1.7 “5.3.1 Text-Positioning Operators”，我有点困惑。

我写了一些代码来获取转换矩阵和初始文本位置。

        CGPDFOperatorTableSetCallback (table, "MP", &op_MP);//Define marked-content point
    CGPDFOperatorTableSetCallback (table, "DP", &op_DP);//Define marked-content point with property list
    CGPDFOperatorTableSetCallback (table, "BMC", &op_BMC);//Begin marked-content sequence
    CGPDFOperatorTableSetCallback (table, "BDC", &op_BDC);//Begin marked-content sequence with property list
    CGPDFOperatorTableSetCallback (table, "EMC", &op_EMC);//End marked-content sequence

    //Text State operators
    CGPDFOperatorTableSetCallback(table, "Tc", &op_Tc);
    CGPDFOperatorTableSetCallback(table, "Tw", &op_Tw);
    CGPDFOperatorTableSetCallback(table, "Tz", &op_Tz);
    CGPDFOperatorTableSetCallback(table, "TL", &op_TL);
    CGPDFOperatorTableSetCallback(table, "Tf", &op_Tf);
    CGPDFOperatorTableSetCallback(table, "Tr", &op_Tr);
    CGPDFOperatorTableSetCallback(table, "Ts", &op_Ts);

    //text showing operators
    CGPDFOperatorTableSetCallback(table, "TJ", &op_TJ);
    CGPDFOperatorTableSetCallback(table, "Tj", &op_Tj);
    CGPDFOperatorTableSetCallback(table, "'", &op_apostrof);
    CGPDFOperatorTableSetCallback(table, "\"", &op_double_apostrof);

    //text positioning operators        
    CGPDFOperatorTableSetCallback(table, "Td", &op_Td);
    CGPDFOperatorTableSetCallback(table, "TD", &op_TD);
    CGPDFOperatorTableSetCallback(table, "Tm", &op_Tm);
    CGPDFOperatorTableSetCallback(table, "T*", &op_T);

    //text object operators
    CGPDFOperatorTableSetCallback(table, "BT", &op_BT);//Begin text object
    CGPDFOperatorTableSetCallback(table, "ET", &op_ET);//End text object

这是应用程序午餐后的输出：

    2010-09-02 15:09:23.041 testSearch[8251:207] op_BT begin
    Integer value: 0
    2010-09-02 15:09:23.043 testSearch[8251:207] op_BT end
    2010-09-02 15:09:23.043 testSearch[8251:207] op_Tf begin
    Integer value: 1
    2010-09-02 15:09:23.044 testSearch[8251:207] op_Tf end
    2010-09-02 15:09:23.044 testSearch[8251:207] op_Tm begin
    Float value: 557.364197
    2010-09-02 15:09:23.045 testSearch[8251:207] op_Tm end
    2010-09-02 15:09:23.045 testSearch[8251:207] op_TJ begin
    2010-09-02 15:09:23.046 testSearch[8251:207] Array string value [0]: F
    2010-09-02 15:09:23.046 testSearch[8251:207] Array integer value [1]: 94985208
    2010-09-02 15:09:23.047 testSearch[8251:207] Array string value [2]: r
    2010-09-02 15:09:23.047 testSearch[8251:207] Array integer value [3]: 94985208
    2010-09-02 15:09:23.048 testSearch[8251:207] Array string value [4]: o
    2010-09-02 15:09:23.048 testSearch[8251:207] Array integer value [5]: 94985208
    2010-09-02 15:09:23.049 testSearch[8251:207] Array string value [6]: m s
    2010-09-02 15:09:23.049 testSearch[8251:207] Array integer value [7]: 94985208
    2010-09-02 15:09:23.049 testSearch[8251:207] Array string value [8]: a
    2010-09-02 15:09:23.050 testSearch[8251:207] Array integer value [9]: 94985208
    2010-09-02 15:09:23.050 testSearch[8251:207] Array string value [10]: m
    2010-09-02 15:09:23.051 testSearch[8251:207] Array integer value [11]: 94985208
    2010-09-02 15:09:23.051 testSearch[8251:207] Array string value [12]: p
    2010-09-02 15:09:23.052 testSearch[8251:207] Array integer value [13]: 94985208
    2010-09-02 15:09:23.053 testSearch[8251:207] Array string value [14]: l
    2010-09-02 15:09:23.054 testSearch[8251:207] Array integer value [15]: 94985208
    2010-09-02 15:09:23.055 testSearch[8251:207] Array string value [16]: e t
    2010-09-02 15:09:23.055 testSearch[8251:207] Array integer value [17]: 94985208
    2010-09-02 15:09:23.057 testSearch[8251:207] Array string value [18]: o r
    2010-09-02 15:09:23.057 testSearch[8251:207] Array integer value [19]: 94985208
    2010-09-02 15:09:23.058 testSearch[8251:207] Array string value [20]: e
    2010-09-02 15:09:23.058 testSearch[8251:207] Array integer value [21]: 94985208
    2010-09-02 15:09:23.059 testSearch[8251:207] Array string value [22]: s
    2010-09-02 15:09:23.059 testSearch[8251:207] Array integer value [23]: 94985208
    2010-09-02 15:09:23.060 testSearch[8251:207] Array string value [24]: u
    2010-09-02 15:09:23.061 testSearch[8251:207] Array integer value [25]: 94985208
    2010-09-02 15:09:23.061 testSearch[8251:207] Array string value [26]: l
    2010-09-02 15:09:23.062 testSearch[8251:207] Array integer value [27]: 94985208
    2010-09-02 15:09:23.062 testSearch[8251:207] Array string value [28]: t
    2010-09-02 15:09:23.063 testSearch[8251:207] op_TJ end

如果有人熟悉文本矩阵和文本定位运算符，最好能解释一下这些东西是如何工作的。

如何使用 Tm（变换矩阵和其他数据）计算文本位置（或字形？）？

【问题讨论】：

你知道如何使用这些运算符了吗？谢谢！

【参考方案1】：

@Koteg：嗨！你终于设法让它工作了吗？对于 Tm，我可以获得所有六个值，但现在我看不到如何将单词的位置放入一行中...... 我有一个想法：如果我们在 Tj，只需获取字母之间的空格（每次都跳这个相同），然后使用 Tm，获取单词的位置。 在 TJ 的情况下，这要复杂得多：获取水平平移的值以减去数组每个部分的 Tm 矩阵，但是在该数组中搜索单词将比 Tj 更复杂。

顺便说一句，对于其他人：

for(size_t n = 0; n < CGPDFArrayGetCount(array); n += 2)

    if(n >= CGPDFArrayGetCount(array))
        continue;

    CGPDFStringRef string;
    success = CGPDFArrayGetString(array, n, &string);
    if(success)
    
        NSString *data = (NSString *)CGPDFStringCopyTextString(string);
        NSLog(@"array data : %@", data);

        [searcher.currentData appendFormat:@"%@", data];
        [data release];
    

    CGPDFReal real;
    success = CGPDFArrayGetNumber(array, n+1, &real);
    if(success)
    
        NSLog(@"array real : %f", real);
    

谢谢