Denkmalliste Erfurt PDF Scrape

convert.sh

http://www.erfurt.de/mam/ef/rathaus/buergerservice/doc/denkmalliste.pdf
pdftohtml -nodrm -xml -i -f 4 denkmalliste.pdf denkmalliste.xml

scrape.py

from lxml import etree
import pandas as pd

columns = ['strassenname', 'h_nr', 'h_nr_z', 'objektbezeichnung', 'eigenname_erlaeuterung', 'gem_krz', 'flur', 'flst_z', 'flst_n', 'stadtteil_krz', 'eKD_ENS', 'seite']

tree = etree.parse('denkmalliste.xml')
root = tree.getroot()

entries = []

for page in root.findall('page'):
    page_no = int(page.get('number'))
    e = {}
    e['seite'] = page_no
    old_left = 0
    for te in page.findall('text'):
        left = int(te.get('left'))
        text = te.text
        if text == None:
            continue
        if left == 1080:
            continue
        if left < old_left:
            #print e
            entries.append(e)
            
            # new entry
            old_left = 0
            e = {}
            e['seite'] = page_no
        
        #print left, text
        
        if left == 1013:
            e['eKD_ENS'] = text
        if left == 972:
            e['stadtteil_krz'] = text
        if left == 960 or left == 947:
            tmp = text.split()
            e['flst_n'] = tmp[0]
            if len(tmp) > 1:
                e['stadtteil_krz'] = tmp[1]
        elif left == 917 or left == 911:
            e['flst_z'] = text
        elif left == 884:
            e['flur'] = text
        elif left == 818:
            e['gem_krz'] = text
        elif left == 585:
            e['eigenname_erlaeuterung'] = text
        elif left == 400:
            e['objektbezeichnung'] = text
        elif left == 362 or left == 355:
            tmp = text.split()
            e['h_nr'] = tmp[0]
            if len(tmp) > 1:
                e['h_nr_z'] = tmp[1]
        elif left == 223:
            e['strassenname'] = text
        old_left = left

df = pd.DataFrame(data=entries, columns=columns)
df.to_csv('denkmalliste.csv', encoding='utf8', index=False)