From 8b30107a3efa6d8aa9a65439239c45c6801b27b9 Mon Sep 17 00:00:00 2001
From: Zingo Andersen <zingo@zingo.org>
Date: Thu, 24 Dec 2015 20:39:28 +0100
Subject: [PATCH] Add geolocation cache for distances below 3km

By reusing lookup's the server need to be contacted a lot less
Aslo names get a bit clustered and match together a bit better.
---
 elodie/constants.py    |  1 +
 elodie/geolocation.py  | 23 +++++++++++++---
 elodie/localstorage.py | 60 ++++++++++++++++++++++++++++++++++++++++--
 3 files changed, 78 insertions(+), 6 deletions(-)

diff --git a/elodie/constants.py b/elodie/constants.py
index ebb7c49..9622635 100644
--- a/elodie/constants.py
+++ b/elodie/constants.py
@@ -3,5 +3,6 @@ from os import path
 debug = True
 application_directory = '{}/.elodie'.format(path.expanduser('~'))
 hash_db = '{}/hash.json'.format(application_directory)
+location_db = '{}/location.json'.format(application_directory)
 script_directory = path.dirname(path.dirname(path.abspath(__file__)))
 exiftool_config = '%s/configs/ExifTool_config' % script_directory
diff --git a/elodie/geolocation.py b/elodie/geolocation.py
index 34b6819..e83232d 100644
--- a/elodie/geolocation.py
+++ b/elodie/geolocation.py
@@ -9,6 +9,7 @@ import sys
 import urllib
 
 from elodie import constants
+from elodie.localstorage import Db
 
 class Fraction(fractions.Fraction):
     """Only create Fractions from floats.
@@ -84,17 +85,31 @@ def get_key():
     return config.get('MapQuest', 'key')
 
 def place_name(lat, lon):
+
+    # Try to get cached location first
+    db = Db()
+    # 3km distace radious for a match
+    cached_place_name = db.get_location_name(lat, lon,3000)
+    if(cached_place_name is not None):
+        return cached_place_name
+
+    lookup_place_name = None;
     geolocation_info = reverse_lookup(lat, lon)
     if(geolocation_info is not None):
         if('address' in geolocation_info):
             address = geolocation_info['address']
             if('city' in address):
-                return address['city']
+                lookup_place_name = address['city']
             elif('state' in address):
-                return address['state']
+                lookup_place_name = address['state']
             elif('country' in address):
-                return address['country']
-    return None
+                lookup_place_name = address['country']
+
+    if(lookup_place_name is not None):
+        db.add_location(lat, lon, lookup_place_name)
+        # TODO: Maybe this should only be done on exit and not for every write.
+        db.update_location_db()
+    return lookup_place_name
 
 
 def reverse_lookup(lat, lon):
diff --git a/elodie/localstorage.py b/elodie/localstorage.py
index 3ca8375..eb343b5 100644
--- a/elodie/localstorage.py
+++ b/elodie/localstorage.py
@@ -1,7 +1,7 @@
 import hashlib
 import json
 import os
-
+from math import radians, cos, sqrt
 from elodie import constants
 
 class Db(object):
@@ -24,7 +24,22 @@ class Db(object):
                 self.hash_db = json.load(f)
             except ValueError:
                 pass
-        
+
+        # If the location db doesn't exist we create it.
+        # Otherwise we only open for reading
+        if not os.path.isfile(constants.location_db):
+            with open(constants.location_db, 'a'):
+                os.utime(constants.location_db, None)
+
+        self.location_db = []
+
+        # We know from above that this file exists so we open it for reading only.
+        with open(constants.location_db, 'r') as f:
+            try:
+                self.location_db = json.load(f)
+            except ValueError:
+                pass
+
     def add_hash(self, key, value, write=False):
         self.hash_db[key] = value
         if(write == True):
@@ -55,3 +70,44 @@ class Db(object):
                 buf = f.read(blocksize)
             return hasher.hexdigest()
         return None
+
+    # Location database
+    # Currently quite simple just a list of long/lat pairs with a name
+    # If it gets many entryes a lookup might takt to long and a better
+    # structure might be needed. Some speed up ideas:
+    # - Sort it and inter-half method can be used
+    # - Use integer part of long or lat as key to get a lower search list
+    # - Cache a smal number of lookups, photos is likey to be taken i clusters
+    #   around a spot during import.
+
+    def add_location(self, latitude, longitude, place, write=False):
+        data = {}
+        data['lat'] = latitude
+        data['long'] = longitude
+        data['name'] = place
+        self.location_db.append(data)
+        if(write == True):
+            self.update_location_db()
+
+    def get_location_name(self, latitude, longitude,threshold_m):
+        for data in self.location_db:
+            # As threshold is quite smal use simple math
+            # From http://stackoverflow.com/questions/15736995/how-can-i-quickly-estimate-the-distance-between-two-latitude-longitude-points
+            # convert decimal degrees to radians
+
+            lon1, lat1, lon2, lat2 = map(radians, [longitude, latitude, data['long'], data['lat']])
+
+            R = 6371000  # radius of the earth in m
+            x = (lon2 - lon1) * cos( 0.5*(lat2+lat1) )
+            y = lat2 - lat1
+            d = R * sqrt( x*x + y*y )
+            # Use if closer then threshold_km reuse lookup
+            if(d<=threshold_m):
+                #print "Found in cached location dist: %d m" % d
+                return data['name'];
+        return None
+
+
+    def update_location_db(self):
+        with open(constants.location_db, 'w') as f:
+            json.dump(self.location_db, f)