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

# In[1]:


import pandas as pd, json, numpy as np
import matplotlib.pyplot as plt
get_ipython().run_line_magic('matplotlib', 'inline')


# In[2]:


url='http://en.wikipedia.org/wiki/List_of_airports_in_Romania'
df=pd.read_html(url)
df=df[0].loc[:17].T.set_index(0).T.loc[2:].set_index('IATA')


# In[3]:


df


# In[4]:


from pygeocoder import Geocoder
apik='AIzaSyDybC2OroTE_XDJTuxjKruxFpby5VDhEGk'


# In[5]:


locations={}
for i in df.index:
    results = Geocoder(apik).geocode(i+' airport romania')
    locations[i]=results[0].coordinates
    print i


# In[6]:


file("locations_ro.json",'w').write(json.dumps(locations))


# In[7]:


locations=json.loads(file('locations_ro.json','r').read())


# In[8]:


import requests


# In[9]:


airportialinks={}
for i in locations:
    print i,
    url='https://cse.google.com/cse?cx=partner-pub-6479063288582225%3A8064105798&cof=FORID%3A10&ie=UTF-8&q='+str(i)+'+airport+romania'
    m=requests.get(url).content
    z=pd.read_html(m)[5][0][0]
    z=z[z.find('http'):]
    airportialinks[i]=z
    print z


# In[10]:


#reformat
for z in airportialinks:
    airportialinks[z]=airportialinks[z].split('arrivals')[0].split('departures')[0].replace(' ','').replace('...','-international-')
    if airportialinks[z][-1]!='/':airportialinks[z]+='/' 
    #manual fixes
    if z=='TSR':airportialinks[z]='https://www.airportia.com/romania/timişoara-traian-vuia-airport/'
    print airportialinks[z]


# In[11]:


sch={}


# record schedules for 2 weeks, then augment count with weekly flight numbers.
# seasonal and seasonal charter will count as once per week for 3 months, so 12/52 per week. TGM separate, since its history is in the past.

# In[12]:


for i in locations:
    print i
    if i not in sch:sch[i]={}
    if i!='TGM':
        #march 11-24 = 2 weeks
        for d in range (11,25):
            if d not in sch[i]:
                try:
                    url=airportialinks[i]
                    full=url+'departures/201703'+str(d)
                    m=requests.get(full).content
                    sch[i][full]=pd.read_html(m)[0]
                    #print full
                except: pass #print 'no tables',i,d
    else:
        #november 17-30 = 2 weeks
        for d in range (17,31):
            if d not in sch[i]:
                try:
                    url=airportialinks[i]
                    full=url+'departures/201611'+str(d)
                    m=requests.get(full).content
                    sch[i][full]=pd.read_html(m)[0]
                    #print full
                except: pass #print 'no tables',i,d


# In[13]:


mdf=pd.DataFrame()


# In[14]:


for i in sch:
    for d in sch[i]:
        df=sch[i][d].drop(sch[i][d].columns[3:],axis=1).drop(sch[i][d].columns[0],axis=1)
        df['From']=i
        df['Date']=d
        mdf=pd.concat([mdf,df])


# In[15]:


mdf=mdf.replace('Hahn','Frankfurt')
mdf=mdf.replace('Hahn HHN','Frankfurt HHN')


# In[16]:


mdf['City']=[i[:i.rfind(' ')] for i in mdf['To']]
mdf['Airport']=[i[i.rfind(' ')+1:] for i in mdf['To']]


# In[17]:


file("mdf_ro_dest.json",'w').write(json.dumps(mdf.reset_index().to_json()))


# In[25]:


len(mdf)


# In[18]:


airlines=set(mdf['Airline'])


# In[19]:


cities=set(mdf['City'])


# In[20]:


file("cities_ro_dest.json",'w').write(json.dumps(list(cities)))
file("airlines_ro_dest.json",'w').write(json.dumps(list(airlines)))


# In[26]:


citycoords={}


# In[27]:


for i in cities:
    if i not in citycoords:
        if i==u'Birmingham': z='Birmingham, UK'
        elif i==u'Valencia': z='Valencia, Spain'
        elif i==u'Naples': z='Naples, Italy'
        elif i==u'St. Petersburg': z='St. Petersburg, Russia'
        elif i==u'Bristol': z='Bristol, UK'
        else: z=i
        citycoords[i]=Geocoder(apik).geocode(z)
        print i


# In[28]:


citysave={}
for i in citycoords:
    citysave[i]={"coords":citycoords[i][0].coordinates,
                 "country":citycoords[i][0].country}


# In[29]:


file("citysave_ro_dest.json",'w').write(json.dumps(citysave))