Mundraub.org shows plants in common land which can be freely harvested. This small script will harvest the data from Mundraub.org in a (hopefully) friendly way. It will output GeoJSON which can be easily transfered to GPX, SHP or whatnot. How to get the inital JSON is a small riddle for the user itself.

mundraeuber.py

#!/usr/bin/env python
# -*- coding: utf-8 -*-

import json
from geojson import Feature, Point, FeatureCollection
from lxml import etree, html
from urllib2 import urlopen
import re
from tqdm import *

# open file
with open('plant.json') as data_file:    
    data = json.load(data_file)

# declare some variables
i = 0
geo_json_list = []

# iterate over all elements of JSON
while i < len(data['features']):
    # loop for the tqdm progress bar
    for i in tqdm(range(len(data['features'])), desc = 'Getting data '):
        
        y = data['features'][i]['pos'][0]
        x = data['features'][i]['pos'][1]
        my_point = Point((float(x), float(y)))
        nid = data['features'][i]['properties']['nid']
                
        # get the description from the website
        url = "https://www.mundraub.org/node/" + nid
        soup = html.fromstring(urlopen(url).read().decode('utf-8'))
        croutons = soup.find_class("processed_text")
        
        for item in croutons:
            description = html.tostring(item,encoding='unicode', method='text')
            
        my_feature = Feature(geometry=my_point, properties={'nid': nid, 'description': description})
        geo_json_list.append(my_feature)    
            
        i += 1

# write file
with open('plant.geojson', 'rb+') as f:
    json.dump(FeatureCollection(geo_json_list), f)

Great, just what I was looking for. I had to change the filename in line 12 to match the one in line 43 (plant.geojson) in ordner not the get a file not found error.