Question

I am parsing content using Python and Beautiful Soup then writing it to a CSV file, and have run into a bugger of a problem getting a certain set of data. The data is ran through an implementation of TidyHTML that I have crafted and then other not needed data is stripped out.

The issue is that I need to retrieve all data between a set of <h3> tags.

Sample Data:

<h3><a href="Vol-1-pages-001.pdf">Pages 1-18</a></h3>
<ul><li>September 13 1880. First regular meeting of the faculty;
 September 14 1880. Discussion of curricular matters. Students are
 debarred from taking algebra until they have completed both mental
 and fractional arithmetic; October 4 1880.</li><li>All members present.</li></ul>
 <ul><li>Moved the faculty henceforth hold regular weekkly meetings in the
 President's room of the University building; 11 October 1880. All
 members present; 18 October 1880. Regular meeting 2. Moved that the
 President wait on the property holders on 12th street and request
 them to abate the nuisance on their property; 25 October 1880.
 Moved that the senior and junior classes for rhetoricals be...</li></ul>
 <h3><a href="Vol-1-pages-019.pdf">Pages 19-33</a></h3>`

I need to retrieve all of the content between the first closing </h3> tag and the next opening <h3> tag. This shouldn’t be hard, but my thick head isn’t making the necessary connections. I can grab all of the <ul> tags but that doesn’t work because there is not a one to one relationship between <h3> tags and <ul> tags.

The output I am looking to achieve is:

Pages 1-18|Vol-1-pages-001.pdf|content between and tags.

The first two parts have not been a problem but content between a set of tags is difficult for me.

My current code is as follows:

import glob, re, os, csv
from BeautifulSoup import BeautifulSoup
from tidylib import tidy_document
from collections import deque

html_path = 'Z:\\Applications\\MAMP\\htdocs\\uoassembly\\AssemblyRecordsVol1'
csv_path = 'Z:\\Applications\\MAMP\\htdocs\\uoassembly\\AssemblyRecordsVol1\\archiveVol1.csv'

html_cleanup = {'\r\r\n':'', '\n\n':'', '\n':'', '\r':'', '\r\r': '', '<img src="UOSymbol1.jpg"    alt="" />':''}

for infile in glob.glob( os.path.join(html_path, '*.html') ):
    print "current file is: " + infile

    html = open(infile).read()

    for i, j in html_cleanup.iteritems():
            html = html.replace(i, j)

    #parse cleaned up html with Beautiful Soup
    soup = BeautifulSoup(html)

    #print soup
    html_to_csv = csv.writer(open(csv_path, 'a'), delimiter='|',
                      quoting=csv.QUOTE_NONE, escapechar=' ')  
    #retrieve the string that has the page range and file name
    volume = deque()
    fileName = deque()
    summary = deque()
    i = 0
    for title in soup.findAll('a'):
            if title['href'].startswith('V'):
             #print title.string
             volume.append(title.string)
             i+=1
             #print soup('a')[i]['href']
             fileName.append(soup('a')[i]['href'])
             #print html_to_csv
             #html_to_csv.writerow([volume, fileName])

    #retrieve the summary of each archive and store
    #for body in soup.findAll('ul') or soup.findAll('ol'):
    #        summary.append(body)
    for body in soup.findAll('h3'):
            body.findNextSibling(text=True)
            summary.append(body)

    #print out each field into the csv file
    for c in range(i):
            pages = volume.popleft()
            path = fileName.popleft()
            notes = summary
            if not summary: 
                    notes = "help"
            if summary:
                    notes = summary.popleft()
            html_to_csv.writerow([pages, path, notes])

Answer 1

Extract content between </h3> and <h3> tags:

from itertools import takewhile

h3s = soup('h3') # find all <h3> elements
for h3, h3next in zip(h3s, h3s[1:]):
  # get elements in between
  between_it = takewhile(lambda el: el is not h3next, h3.nextSiblingGenerator())
  # extract text
  print(''.join(getattr(el, 'text', el) for el in between_it))

The code assumes that all <h3> elements are siblings. If it is not the case then you could use h3.nextGenerator() instead of h3.nextSiblingGenerator().