landsearch/index.py

#!/usr/bin/python3

import mysql.connector
import requests
from bs4 import BeautifulSoup
import urllib.parse
import re
from sys import exit as exit
import json
import datetime
import custom_email
from tabulate import tabulate
from configparser import ConfigParser
from os import path

### TO DO ###
#
#  email results
#  allow this script to be called and work by itself (if __name__ == __main__)
#  Print useful reports (land only, house and land, etc)
#  Check if db entries no longer appear online (mark expired)
#  When checking online from various sites, check if address already exists in db
#    - if so, warn user and do not add
#  Add date_added to initial entries
#  Check results against database for changes
#   - update and add/change date_modified 
#  Add argument to run update query when results.py is calles
#  Add database column to hold parcel number.  Make links to GIS servers 
#
#  IDENTIFY NEW PROPERTIES!!
#
#  Automate db opening and closing when calling dbinsert()
#
#############

class Property:
  """Description of a proerty"""
  
  def __init__ (self, site_name, type, MLS, address, city, st, zip, \
    county, price, acres, title='', sqft=0, bedrooms=0, baths=0, description='', link=''):

    self.site_name = site_name
    self.type = type
    self.MLS = MLS
    self.title = title
    self.sqft = sqft
    self.bedrooms = bedrooms
    self.baths = baths
    self.address = address
    self.city = city
    self.st = st
    self.zip = zip
    self.county = county
    self.price = price
    self.acres = acres
    self.description = description
    self.link = link

    

class Search:
  """Universal Search Criteria"""

  def checktype(self, attribute):
    if not attribute == 'None':
      return attribute
    else:
      return ''

#  def __init__(self, county: list, lower_price=0, upper_price=500000, \
#    lower_acres=5, upper_acres=15, type=['farm','land','home'], lower_sqft='', upper_sqft='', \
#    lower_bedrooms='', upper_bedrooms=''):
  def __init__(self, file = '../landsearch.conf'):
    self.file = file
    if not path.exists(self.file):
      raise FileNotFoundError("The config file cannot be opened", self.file)
    try:
      config = ConfigParser()
      config.read(self.file)
      search_params = config['Search']
    except FileNotFoundError as err:
      print(err, "Using default search parameters.")
    except Exception as err:
      print(err, "Using default search parameters.")

    county = search_params.get('county', ['Gwinnett', 'Hall', 'Jackson', 'Walton', 'Barrow'])
    if isinstance(county, str):
      county = county.split(", ")
    type = search_params.get('type', ['farm', 'house', 'land'])
    if isinstance(type, str):
      type = type.split(", ")

    self.types=['land', 'farm', 'home', 'house']
    self.county = county
    self.lower_price = self.checktype(search_params.get('lower_price', 0))
    self.upper_price = self.checktype(search_params.get('upper_price', 525000))
    self.lower_acres = self.checktype(search_params.get('lower_acres', 5))
    self.upper_acres = self.checktype(search_params.get('upper_acres', 15))
    self.type = type			##accept list!
    self.lower_sqft = self.checktype(search_params.get('lower_sqft', ''))
    self.upper_sqft = self.checktype(search_params.get('upper_sqft', ''))
    self.lower_bedrooms = self.checktype(search_params.get('lower_bedrooms', ''))
    self.upper_bedrooms = self.checktype(search_params.get('upper_bedrooms', ''))
#    self.lower_price = search_params.get('lower_price', 0)
#    self.upper_price = search_params.get('upper_price', 525000)
#    self.lower_acres = search_params.get('lower_acres', 5)
#    self.upper_acres = search_params.get('upper_acres', 15)
#    self.lower_sqft = search_params.get('lower_sqft', '')
#    self.upper_sqft = search_params.get('upper_sqft', '')
#    self.lower_bedrooms = search_params.get('lower_bedrooms', '')
#    self.upper_bedrooms = search_params.get('upper_bedrooms', '')

    for property_type in self.type:
      assert property_type in self.types, ("Unknown type '" + property_type + "'. Property Type must be of type: " + str(self.types))
    
## FOR TESTING, PRINT ALL ATTRIBUTES OF SEARCH ##
#    print(vars(self))

    
class ImproperSearchError(Exception):
  def __init__ (self, search, message="Improper Search.  Must use instance of Search class"):
    self.search = search
    self.message = message
    super().__init__(self.message)

class MLSDATA:
  """Fetches and stores MLS Data
     Currently only supports GeorgiaMLS.com (GMLS)"""
  counties=['Gwinnett', 'Barrow', 'Hall', 'Jackson', 'Walton']
  GoogleAPIKey = 'AIzaSyAXAnpBtjv760W8YIPqKZ0dFXpwAaZN7Es'
  live_google = True

  def __init__ (self, mlstype):
    self.help = "This is a class that will retrieve MLS data from various sources, store the info in a database, and run queries on the data."
    self.mlstype = mlstype.lower()  ## Determines what kind of data is to be retreieve (gmls, Zillow, etc)
    self.cursor = ''
    self.cnx = ''
    self.new_listings = []

  def stringbuilder(self, search: Search, county):
    """ Takes Search class and build appropriate URL query based on mlstype.  Currently only supports gmls."""

    if self.mlstype == 'gmls':
      base_addr = 'https://www.georgiamls.com/real-estate/search-action.cfm?'
      params = [('cnty', county), \
        ('lpl', search.lower_price), ('lph', search.upper_price), \
        ('acresL', search.lower_acres), ('acresH', search.upper_acres), \
        ('sqftl', search.lower_sqft), ('sqfth', search.upper_sqft), \
        ('orderBy', 'b'), \
        ('scat', '1'), \
        ('sdsp', 'g')]
      for type in search.type:
        if 'land' in type.lower():
          params.append(('typ', 'll'))
        if 'farm' in type.lower():
          params.append(('typ', 'af'))
        if 'home' in type.lower():
          params.append(('typ', 'sd'))
        if 'house' in type.lower():
          params.append(('typ', 'sd'))
      search_string = base_addr + urllib.parse.urlencode(params)
      print(search_string)
      return search_string

  def break_address(self, address):
    """Takes an address string in the form 'street address|city, state zip' and returns a list"""
    street = address[:address.find('|')]
    csz = address[address.find('|')+1:]
    city = csz[:csz.find(',')]
    st = csz[csz.find(',')+1:].split(' ')[1]
    zip = csz[csz.find(',')+1:].split(' ')[2]
    split_address = [street, city, st, zip]
    return split_address

  def gmlsparser(self, URL, county, pages=''):
    """ Retrieve the website for georgiamls.com and returns a list of Property objects. 
        UNIQUE TO GEORGIAMLS.COM ONLY!!"""
    properties_list = []
    r = requests.get(URL)
    soup = BeautifulSoup(r.content, 'html5lib')
    if pages == '':
      try:
        pages = soup.find("div", {'class':'small listing-pagination-count'}).getText().strip().split(" ")[-1]
        current_page = soup.find("div", {'class':'small listing-pagination-count'}).getText().strip().split(" ")[-3]
      except AttributeError as err:
        print("No Results Found.")
        return
    else:
      print('pages already set to: ' + str(pages))
    for page in range(0, int(pages)):
      print('Processing Page: ' + str(page + 1) + ' of ' + str(pages))
      if not page == 0:
        next_URL = URL + '&start=' + str(((12*page)+1))
        soup = BeautifulSoup(requests.get(next_URL).content, 'html5lib')
      raw_listings = soup.findAll("div", {'class':'col-xs-12 col-sm-6 col-lg-4 text-center listing-gallery'})
      for listing in raw_listings:
        items = listing.findAll("p")					## 
        site_name = self.mlstype
        MLS = " ".join(items[3].getText().strip()[6:15].split())	## MLS NUMBER
        title = ''							## Listing Title (address if no title)
        price = items[0].string.strip()					## Price
        if self.mlstype == 'gmls':
          link = 'https://www.georgiamls.com' + listing.a['href']
        detail_request = requests.get(link)
        detail_soup = BeautifulSoup(detail_request.content, 'html5lib')
        details = detail_soup.findAll('tr')
        bedbath = details[1].findAll('td')[1].getText().strip().split('/')
        br = bedbath[0][:-3]
        ba = bedbath[1][:-3]
        baths = ba							## IF House is present
        bedrooms = br							## IF House is present
        address = ''
        for element in details:
          if 'sqft' in element.getText():
            sqft = element.findAll('td')[1].getText().strip()[:-5].replace(',','')
          if 'lot size' in element.getText().lower():
             acres = element.findAll('td')[1].getText().strip()[:-6]
          if 'Property Type' in element.getText():
            ptype = element.findAll('td')[1].getText().strip()
            if 'acreage' in ptype.lower():
              type = 'af'
            elif 'land lot' in ptype.lower():
              type = 'll'
            elif 'single family home' in ptype.lower():
              type = 'sf'
            else:
              type = 'unknown'
          if 'Address' in element.getText():
            if not address:  #Prevents finding the word 'address' elsewhere in the listings
              address = element.findAll('td')[1]
#7              print("TEST ADDRESS: ", element)
              street_address = list(address)[0].strip()
              csz = list(address)[2].strip()
              split_address = self.break_address(street_address + '|' + csz)
        description = detail_soup.find('div', {'id':'listing-remarks'}).getText().strip().replace('\t','')
        data = Property(site_name = self.mlstype,  \
		type = type, \
		MLS = MLS, \
		bedrooms = bedrooms, \
		baths = baths, \
		sqft = sqft, \
		address = split_address[0], \
		city = split_address[1].title(), \
		st = split_address[2].upper(), \
		zip = split_address[3], \
		county = county.title(), \
		price = price.replace('$','').replace(',',''), \
		acres = acres, \
		description = description, \
		link = link)
        properties_list.append(data)
        print('Scanned: ' + data.address)

    return properties_list

  def getmlsdata(self, search: Search, county):
    """This is the main entrypoint.  Takes arguments to pass to stringbuilder to create the URL.
       Selects appropriate parser based on self.mlstype from class intance.
       Needs any modifications from the standard search ($0 to $500,000, 5 to 15 acres, etc)
       See class search for more information.
       --> 9/1/20 - takes Search class as argument.  All properties are handled by the class <--"""
    if isinstance(search, Search):

##
# PROGRAM BREAKS HERE  - Used to loop for each county, not Search class contains list of counties.  Need to automate looping.
##

      if not county in self.counties:  ### FIX for lower()
        print("County " + county + " not regognized.  Exiting")
      else:
        print("Scanning for results in " + county + " using the " + self.mlstype.upper() + " database.")
        if self.mlstype == 'gmls':
          list = self.gmlsparser(self.stringbuilder(search, county), county)
        return list
    else:
     raise ImproperSearchError(search)

  def checkdb(self, criteria_dict):
    """Check dictionary of critera against database.
       Currently accepts keys: MLS, title, address (street number/name, not city/state/zip).
       Returns True if records exists."""
    if self.cursor:				## Check if DB is connected
      for criteria in criteria_dict:
		## Determine criteria passed, and execute queries for each
        if criteria == 'MLS':
          self.cursor.execute("SELECT COUNT(*) FROM properties WHERE MLS = %(MLS)s GROUP BY id", {criteria:criteria_dict[criteria]})
          if self.cursor.rowcount > 0: return self.cursor.rowcount # stop for loop if match already found.
        elif criteria == 'title':
          self.cursor.execute("SELECT COUNT(*) FROM properties WHERE title = %(title)s GROUP BY id", {criteria:criteria_dict[criteria]})
          if self.cursor.rowcount > 0: return self.cursor.rowcount # stop for loop if match already found.
        elif criteria == 'address':
          self.cursor.execute("SELECT COUNT(*) FROM properties WHERE address = %(address)s GROUP BY id", {criteria:criteria_dict[criteria]})
          if self.cursor.rowcount > 0: return self.cursor.rowcount # stop for loop if match already found.
        else:
          print("Cannot search on parameter: " + criteria)
      return self.cursor.rowcount
    else:
      print("Database is not connected or cursor not filled.  Use function 'connectdb()' to establish")

  def getGoogle(self, property):
   """Supplies date from Google Distance Matrix API to populate
      distance_to_work
      time_to_work
      distance_to_school
      time_to_school
      Costs money, so it should only be called when inserting a new db record.
      Returns distance in METERS (1m = 0.000621371 mi) and time in SECONDS
      returns fully populated Propery object."""
   print("Fetching live Google Data.  $$")
# Build Request
   destination1 = 'Hebron Christian Acadamy' ## Working query for Hebron Christian Acadamy
   destination2 = 'JHRJ+FJ Atlanta, Georgia' ## Plus code for Hourly parking at Int'l Terminal, KATL
   params = {}
   params['units'] = 'imperial'
   params['origins'] = property.address + ', ' + property.city + ' ' + property.st
   params['destinations'] = 'Hebron Christian Acadamy|JHRJ+FJ Atlanta, Georgia'
   params['key'] = self.GoogleAPIKey
   baseURL = 'https://maps.googleapis.com/maps/api/distancematrix/json?'
   API_URL = baseURL + urllib.parse.urlencode(params)
#   print(API_URL)

#    Send Request and capture result as json
   try:
     google_result = requests.get(API_URL).json()
     if google_result['status'] == 'OK':
       property.distance_to_school = google_result['rows'][0]['elements'][0]['distance']['value']
       property.time_to_school = google_result['rows'][0]['elements'][0]['duration']['value']
       property.distance_to_work = google_result['rows'][0]['elements'][1]['distance']['value']
       property.time_to_work = google_result['rows'][0]['elements'][1]['duration']['value']
   except:
     print("ERROR: Failed to obtain Google API data")

#Load sample data for testing:
#   with open('complex.json') as f:
#      data = json.load(f)
#      google_result = data
### end testing json ###



  def insertrecord(self, property, work_address=None, school_address=None):
    """Inserts record into database.  Takes argument Property class object.
       FUTURE - add date_added field to insert operation."""
    if self.cursor:
      criteria_dict = property.__dict__
      criteria_dict['Date_Added'] = str(datetime.date.today())
      placeholder_columns = ", ".join(criteria_dict.keys())
      placeholder_values = ", ".join([":{0}".format(col) for col in criteria_dict.keys()])
      qry = "INSERT INTO properties ({placeholder_columns}) VALUES {placeholder_values}".format(placeholder_columns=placeholder_columns, placeholder_values=tuple(criteria_dict.values()))
      self.cursor.execute(qry)
      self.cnx.commit()
      print("Inserted " + criteria_dict['MLS'] + " | " + criteria_dict['address'] + " into database.")
    else:
      print("Database is not connected or cursor not filled.  Use function 'connectdb()' to establish")

  def connectdb(self, host='192.168.100.26', user='landsearchuser', password='1234', database='landsearch'):
    """Connects to database and returns a cursor object"""
    self.cnx = mysql.connector.connect(host=host, user=user, password=password, database=database, buffered=True)
    self.cursor = self.cnx.cursor()
    return self.cursor

  def closedb(self):
    """Cleanly close the db."""
    self.cursor.close()
    self.cnx.close()

  def dbinsert(self, properties: list):
    """Inserts records into database.  Takes list of Property class objects"""
    if not properties == None:
      if not isinstance(properties, list):
        raise TypeError('type list required')
      for property in properties:
        if not self.checkdb({'MLS': property.MLS, 'address': property.address}):
          if self.live_google: self.getGoogle(property)		## <- This will populate distance and time fields if set TRUE
          self.insertrecord(property)
          self.new_listings.append(property)
        else:
          print(property.MLS + ' | ' + property.address + ' is already in db.  Not inserted.')
##REMOVE FOR TESTING###
#          self.new_listings.append(property)
#######################
    else:
      print("Empty dataset.  No records to insert.")

  def alerts(self):
    pass

  def email(self):
    body = ''
    data = []
    subj = "New Real Estate Listings for " + str(datetime.date.today())
    for listing in self.new_listings:
      row = []
      body += listing.MLS + " | " + listing.address + " | " + listing.acres + " | " + listing.price + " | " + listing.link + "\n"
      row.append(listing.MLS)
      row.append(listing.address)
      row.append('{:0,.2f}'.format(float(listing.acres)))
      row.append(listing.sqft)
      row.append('${:0,.0f}'.format(int(listing.price)))
      row.append(listing.time_to_school/60 if hasattr(listing, 'time_to_school') else 'NA')
      row.append(listing.link)
      data.append(row)
    body = """\
Daily Real Estate Search Report\n
The following properties have been found which may be of interest.\n
"""
    results = tabulate(data, headers=['MLS', 'Address', 'Acres', 'sqft', 'Price', 'Time to School', 'link'])
    body += results
    sendto = ['stagl.mike@gmail.com', 'M_Stagl@hotmail.com']
    mymail = custom_email.simplemail(subj, body, sendto)
    
    if len(self.new_listings) > 0:
      try:
        mymail.sendmail()
      except Exception as e:
        print("Error sending email.  " + e)
    else:
      print("No new listings. Email not sent")
# REMOVE AFTER TESTING #
      mymail.sendmail()
########################

###########  BEGIN CODE ###############33


if __name__ == '__main__':

  gmls = MLSDATA('GMLS')	# Create MLSDATA object

  mysearch = Search()  		# Create a custom search object
#  print(len(mysearch.county))
#  print(mysearch.county[0])
  myresults = []
 
## Create function in MLSDATA module:
#    - takes counties from configparser and calls getmlsdata for each county.
#    - Compiles results into single list and returns that list
#    - User code would look something like this:
#    _   mysearch = Search()
#    _   mydata = gmls.findalllistings(mysearch)  # This would control the looping of counties and return a list like normal
#    _   gmls.dbinsert(myresults)                # This would automate db opening and closing
  for county in mysearch.county:
    print("local search: ", county)
    mysearch = Search()		## Search used to take county as parameter, so this loop would work.  Now Search class contains list.  loop must occur in getmlsdata module
    mydata = gmls.getmlsdata(mysearch, county)

    for listing in mydata:
      myresults.append(listing)

#  print(len(myresults))
#  print(myresults[0].address)
  gmls.connectdb()
  gmls.dbinsert(myresults)
  gmls.closedb()
#
#  gmls.email()
#
#print()
#print(str(len(gmls.new_listings)) + " new properties found!")
#print()
#for listing in gmls.new_listings:
#  print(listing.MLS, listing.address)
#  gmls = MLSDATA('GMLS')
#
#  #new_properties = []
#
##  for county in ['Jackson']:							### FIX
#  for county in gmls.counties:							### FIX
#    mysearch = Search(county, type=['farm', 'house', 'land'], upper_price=525000)		### FIX
#    mydata = gmls.getmlsdata(mysearch)
#
#    gmls.connectdb()
#    gmls.dbinsert(mydata)
#    gmls.closedb()
#
#  gmls.email()
#
#print()
#print(str(len(gmls.new_listings)) + " new properties found!")
#print()
#for listing in gmls.new_listings:
#  print(listing.MLS, listing.address)
#
#