import pandas as pd
import requests
from bs4 import BeautifulSoup

url='http://www.myanmars.net/enews/2010/myanmar20101113.html'
page=requests.get(url)
soup=BeautifulSoup(page.content)

def dictify(txt):
    attrs={}
    for item in txt.split(';'):
        itembits=item.split(':')
        attrs[itembits[0]]=itembits[1]
    return attrs

def approxify(x,y):
    if abs(x-y)<3:
        return True
    return False



txt='Pyithu Hluttaw representatives in the constituencies shown against them.'
leftCounters={}
for line in soup.find(text=txt).parent.parent.find_next_siblings():
    attrs=dictify(line['style'])
    currLeft=int(attrs['left'])
    if currLeft not in leftCounters:
        leftCounters[currLeft]=1
    else: leftCounters[currLeft]=leftCounters[currLeft]+1

for c in leftCounters:
    if leftCounters[c]>3: print(c,leftCounters[c])

colSep=600
lefts=[100,628]


#---
import re

prevLeft=0
prevTop=0
cols1={}
cols2={}
topCounters1=[]
topCounters2=[]
newState=False
currstate=''
stateReps={}
col=1
for line in soup.find(text=txt).parent.parent.find_next_siblings():
    attrs=dictify(line['style'])
    currLeft=int(attrs['left'])
    if leftCounters[currLeft]<4:
        continue

    currTop=int(attrs['top'])
    if currLeft<colSep:
        lines=cols1
        tops=topCounters1
    else:
        lines=cols2
        tops=topCounters2
        col=2
    #print(attrs, line.find('span').text)
    
    itemtxt=line.find('span').text
    for i in lefts:
        if approxify(i,currLeft):
            if not re.match('\d*\.',itemtxt):
                currstate=itemtxt
                stateReps[currstate]=(col,currTop)
                newState=True
    if newState:
        newState=False
        continue
                
    for i in tops:
        if approxify(currTop,i):
            currTop=i
    if currTop not in tops:
        tops.append(currTop)
    if currTop not in lines:
        lines[currTop]=[]
    lines[currTop].append(itemtxt)
    if currTop<prevTop:
        #new column
        pass
    prevLeft=currLeft
    prevTop=currTop

stateReps

cols1

cols2