Pmc Xml Parsing

Question

[Alex]

I want to get article texts from PMC xml data with Python.

I used an event driven parsing but it doesn't work well with xml like this:

<p>There is a limimited ... Journal, 
<ext-link ext-link-type="uri" xlink:href="http://...index.asp">
http://...index.asp</ext-link>). Published ... </p>

I thinks the problem with tags-text-other_tag-text-tag sequences. And when I use event driven parsing I get something like this:

Code example:

n = 0
result = ""
for event, element in etree.iterparse(fh, events=['start', 'end']):
        if event == "start":
            n += 2
            result += "%s<%s>\n" % (" "*n, element.tag)
            result += "%s%s\n" % (" "*(n+2), element.text)
        if event == "end":
            result += "%s</%s>\n" % (" "*n, element.tag)
            n -= 2

output of this code:

      

        There is a limimited ... Journal, 
        <ext-link>
          <http://www.aapsj.org/theme_issues/virtual/index.asp>
        </ext-link>
        <xref>
          1
        </xref>
        <xref>
          2
        </xref>
        <xref>
          3
        </xref>
        <xref>
          4
        </xref>
      

I suspect that there is some simple solution for this case. How do you handle xml like this with Python?

Is there any tools/libraries for converting PMC XML to raw text?

[Pierre Lindenbaum]

show us the code please :-)

[Alex]

the code example added =)

[Pierre Lindenbaum]

I don't know your python parser but I would guess that it is only DATA oriented (it only expects text data between two tags).

[Alex]

Your googleable code (http://code.google.com/p/lindenb/source/browse/trunk/src/xsl/pubmed2exhibit.xsl) was a helpfull hint for right direction =) Now I know that any SAX parser in my case is bad idea, better to use XSLT

[Michael Schubert]

This might help.

Answer 1

[Luwening]

you would see this http://www.ncbi.nlm.nih.gov/pmc/pmcdoc/tagging-guidelines/article/style.html and the DTD http://dtd.nlm.nih.gov/publishing/ and Biopython may help you parse the xml file(also depend on DTD)

Answer 2

[Chris Maloney]

This question seems to be specifically about parsing XML with Python, and not about bioinformatics or about the NLM DTDs (JATS). I am not a Python user, but I'd suggest asking on the python mailing list, or Googling "parsing xml with python".