在Linux服务器上安装lxml

最近新接的活，第一个任务是处理一堆xml格式的专利文件，把里面的有效信息提取出来

由于公司的相关规定不容许把文件down到本地处理，只能在对方提供的远程服务器上写代码

因为xml里面的元素是XXX:YYYY这种带前缀的格式，用xml.etree的ElementTree死活解析不出来，最后从OverStack上找到了解释

ElementTree is not too smart about namespaces. You need to give the .find(), findall() and iterfind() methods an explicit namespace dictionary. This is not documented very well:

namespaces = {'owl': 'http://www.w3.org/2002/07/owl#'} # add more as needed root.findall('owl:Class', namespaces)

Prefixes are only looked up in the namespaces parameter you pass in. This means you can use any namespace prefix you like; the API splits off the owl: part, looks up the corresponding namespace URL in the namespaces dictionary, then changes the search to look for the XPath expression {http://www.w3.org/2002/07/owl}Class instead.

If you can switch to the lxml library things are better; that library supports the same ElementTree API, but collects namespaces for you in a .nsmap attribute on elements.

因而果断去装lxml

服务器上已经装了python2.7，然而除此以外的东西全都没有，所以又手动装了pip

执行pip install lxml 报了一大堆错，注意到：

Could not find function xmlCheckVersion in library libxml2. Is libxml2 installed?

查到lxml还须要先安装依赖库libxml2和libxslt的开发版

yum install libxml2-devel

yum install libxslt-devel

安装好依赖库后pip仍是报错

此次又认真读了一遍错误信息发现前面竟然还有一行：

unable to execute gcc: No such file or directory

好吧，连gcc都没有安装，难怪没法编译

yum install gcc

安装好之后pip install，竟然，仍是报错！

src/lxml/lxml.etree.c:84:20: 致命错误：Python.h：没有那个文件或目录

再查资料，原来还须要安装一个python-devel，这是Python的头文件和静态库包

yum install python-devel

安装完以后再次pip install lxml

终于见到了喜闻乐见的

Successfully installed lxml

至此，linux下的lxml安装完成