# coding: utf-8

# In[8]:


import pywikibot
from collections import namedtuple
import hashlib
import re
import time
from base64 import b64encode

import requests
import json


# In[50]:


pybot_username="Venkateswaran raman"
pybot_docname="ललितविस्तरः.pdf"
pybot_pages="1:453"


# In[51]:


class OCRException(Exception):
    pass

class retry(object):
    
    def __init__(self,retries,sleep,exception):
        self.retries = retries
        self.sleep = sleep 
        self.exception = exception
        
    def __call__(self, f):
        def inner(*args):
            for i in range(self.retries + 1):
                try:
                    text = f(*args)
                    return text
                except Exception as e:
                    if type(e) is  self.exception:
                        if i == (self.retries):
                            print("GIVING UP!!... Please retry after sometime ")
                            raise(e)
                        else:
                            print("OCR Extract failed.\n======== \nRETRY {}: \n======== \nseleeping for {} seconds before retry".format(i+1,self.sleep))
                            time.sleep(self.sleep)
                    else:
                        raise(e)
                        
        return inner    


# ### Function Definitions

# In[52]:


def to_sans(numb):
    """
    Input a valid integer. Function takes in an integer (in ASCII) and converts it to devanagari. returns a string. 
    """
    str_numb ="{}".format(numb)
    x =  [chr(ord(i) + 2358) for i in str_numb]
    return ''.join(x)

def to_eng(sans_num):
    """
    Input a valid devanagari number (as string). 
    Function converts it to english and returns the ascii equivalent (returns number). 
    """
    x =  [chr(ord(i) - 2358) for i in sans_num]
    return int(''.join(x))

def get_range(p):
    """
    arguments is a range in format [start:end:step]
    returns a generator object - 
    """
    regex = re.compile("([^:\s]*):?")
    matches = regex.findall(p)
    matches.pop()
    matches.extend((None,None,None))
    (start,end,step)  = matches[:3]

    #default missing values
    if not step: step=1
    if not start:    start = 0
    if not end: end = start
    print("start={}; end={}; step={};".format(start,end,step))

    return(int(start),int(end)+1,int(step))


# In[53]:


PageInfo = namedtuple('PageInfo',['url','pageno','actualpage','ocr_url'])
def generate_pages(docname,rng,displayno):
    (start,end,step) = get_range(rng)
    for i in range(start,end,step):
        page = "पृष्ठम्:{}/{}".format(docname, to_sans(i))
        ocr_url = get_url(docname,i,"jpg")
        yield(PageInfo(page,displayno,i,ocr_url))
        displayno = displayno + step
        
## helper functions for generating text from google ocr
def get_url_md5(docname):
    m = hashlib.md5()
    m.update(docname.encode("utf-8"))
    md5 = m.hexdigest()
    return "/{}/{}/".format(md5[0],md5[0:2])

def get_url(docname,pageno,image_format):
    urlprefix="https://upload.wikimedia.org/wikipedia/commons/thumb/"
    urlsuffix="page{}-1024px-".format(pageno)
    md5text = get_url_md5(docname)
    return"{}{}{}/{}.{}.{}".format(urlprefix,md5text,docname,urlsuffix,docname,image_format)

@retry(retries=5,sleep=25,exception=OCRException)
def url_ocr(image_uri):
    req_url = "https://tools.wmflabs.org/ws-google-ocr/api.php?image={}&lang=sa".format(image_uri)
    print(req_url)
    try:
        resp = requests.post(req_url,timeout=30)
        resp.raise_for_status()
    except requests.exceptions.HTTPError as errh:
        print ("Http Error:",errh)
        raise OCRException(errh)
        
    except requests.exceptions.ConnectionError as errc:
        print ("Error Connecting:",errc)
        raise OCRException(errc)        
     
    except requests.exceptions.Timeout as errt:
        print ("Timeout Error:",errt)
        raise OCRException(errt)        
        
    except requests.exceptions.RequestException as err:
        print ("OOps: Something Else",err)
        raise OCRException(err)        

    if resp is None:
        raise OCRException("Empty resp object returned..")        

    ocrjson = json.loads(resp.text)
    ocrtext=ocrjson.get("text")
    return ocrtext


# In[54]:


pages= generate_pages(pybot_docname,pybot_pages,1)


# In[ ]:
#for p in pages:
#    print(p)


# In[ ]:


site = pywikibot.Site("sa","wikisource")
for p in pages:
    print(p)
    page = pywikibot.Page(site,p.url)
    header="<noinclude><pagequality level=\"1\" user=\"{}\" /></noinclude>".format(pybot_username)
    footer="<noinclude></noinclude>"
    if page.text =='':
        ocrtext = url_ocr(p.ocr_url)
        pagetext = "{}{}{}".format(header,ocrtext,footer)
        page.text = pagetext
        page.status=u"OCR done"
        page.save(summary=u'OCR done.')
        #page.save(u'OCR done')
    else:
        print("page text already exists. Skipping ocr")
    #print(text)
    print("=======================================================================")