3、centos6.9 下 OCRmypdf源码安装
Centos版本：
[root@bc22c4e1 ~]# cat /etc/issue
CentOS release 6.9 (Final)
3.1 安装前提要求
1）Python > 3.5
[root@bc22c4e1 ~]# python -V
Python 3.5.0
1
2
2）pip > 0.9.1
[root@bc22c4e1 ~]# pip -V
pip 9.0.1 from /usr/local/python3/lib/python3.5/site-packages (python 3.5)
1
2
3) Python3导入sqlite3成功
4）基础配置
*CentOS/RHEL 6.x*
# yum install gcc python-devel python-setuptools
# easy_install pip
# pip install fabric
1
2
3
4
5）其他功能配置
pdftotext依赖如下：
yum install poppler-utils
1
3.2 具体安装步骤
步骤1：下载git源文件。
git clone -b master https://github.com/jbarlow83/OCRmyPDF.git
步骤2：设置环境
python3 -m venv ./
1
步骤3：源码编译
source venv/bin/activate
1
步骤4：执行安装
cd OCRmyPDF
pip3 install .
1
2
4、OCRmyPDF安装中遇到的错误及解决方案
问题1：gs版本低。
错误如下：
Running setup.py install for ocrmypdf … error
Complete output from command /usr/local/bin/python -u -c “import setuptools, tokenize;file=’/tmp/pip-lio4mtqk-build/setup.py’;f=getattr(tokenize, ‘open’, open)(file);code=f.read().replace(‘\r\n’, ‘\n’);f.close();exec(compile(code, file, ‘exec’))” install –record /tmp/pip-qnapqha6-record/install-record.txt –single-version-externally-managed –compile:
Checking for tesseract >= 3.04…
Found tesseract 3.04.00
Checking for gs >= 9.15..
解决方案：
curl -O http://downloads.ghostscript.com/public/ghostscript-9.14.tar.gz &&
tar -xzf ghostscript-9.14.tar.gz &&
cd ghostscript-9.14 &&
./configure &&
make install &&
make so &&
cp ghostscript-9.14/sobin/libgs.so.9.14 /usr/lib &&
ln -s /usr/lib/libgs.so.9.14 /usr/lib/libgs.so &&
mkdir -p /etc/ld.so.conf.d/ &&
echo “/usr/lib/libgs.so” > /etc/ld.so.conf.d/libgs.conf &&
ldconfig &&
echo “Installing ghostscript finish” &&
gs
参考地址：https://unix.stackexchange.com/questions/79025/install-ghostscript-v-9-05-or-newer-on-centos
问题2：unpaper没有正确安装的错误提示。
解决方案：
步骤1：下载6.1版本unpaper
# cd /var/bin && wget https://www.flameeyes.eu/files/unpaper-6.1.tar.xz && tar -xvf unpaper-6.1.tar.xz
1
步骤2: 编译、安装、运行unpaper6.1
# cd unpaper-6.1 && ./configure && make && make install
1
参考地址：https://github.com/Flameeyes/unpaper/issues/44
问题3：qpdf没有安装错误提示。
解决方案：
qpdf编译、安装、运行。
./configure
make
make install
1
2
3
4
参考：https://github.com/qpdf/qpdf
问题4：
configure: error: Package requirements (libavformat libavcodec libavutil) were not met:
No package 'libavformat' found
No package 'libavcodec' found
No package 'libavutil' found
1
2
3
4
或者：
No package 'libavformat' found No package 'libavcodec' found No package 'libavutil' found
1
解决方案：
步骤1：安装依赖。
yum install libvorbis yasm freetype zlib bzip2 faac lame speex libvpx libogg libtheora x264 XviD openjpeg15 opencore-amr
1
步骤2：下载&安装
wget https://www.libav.org/releases/libav-10.5.tar.gz
tar xvf libav-10.5.tar.gz
cd libav-10.5
/configure –extra-cflags=-I/opt/local/include –extra-ldflags=-L/opt/local/lib –enable-gpl –enable-version3 –enable-libvpx
make
make install
参考：https://superuser.com/questions/850808/how-to-install-libav-tools-in-centos-6
问题5：”No package ‘libffi’ found” error during pip install
解决方案：
yum search ffi | grep python
yum install python-cffi
yum install libffi-devel
pip install –upgrade cffi
参考：https://github.com/Kozea/cairocffi/issues/14
5、安装成功标志。
ages (from reportlab>=3.3.0->ocrmypdf==5.2.post0+g3a7c341.d20170710)
Requirement already satisfied: pycparser in /home/centos001/lib/python3.5/site-packages (from cffi>=1.9.1->ocrmypdf==5.2.post0+g3a7c341.d20170710)
Installing collected packages: ocrmypdf
Running setup.py install for ocrmypdf ... done
Successfully installed ocrmypdf-5.2.post0+g3a7c341.d20170710
1
2
3
4
5
参考
python3、pip3安装参考 ：http://www.jianshu.com/p/6199b5c26725
sqlite3安装参考：http://www.cnblogs.com/greentomlee/p/6561509.html
小结
事件出真知，有问题就努力一个个排查问题，直到全部解决。（耗时2天）
