Find and count all the things nearby a place

findnear.py

import math
import pymongo
from pymongo import Connection

earth_radius_in_miles = 3959.0
radians_to_degrees = 180.0 / math.pi
radius = (0.0189393939 / earth_radius_in_miles) * radians_to_degrees # 100 foot radius

bbox_earth_radius = 3959.0 # miles
bbox_dist_from_center = 0.1 # miles

def sum_by_distinct_type(distinct_type, distinct_key, lat, lon):
    nearby_requests = db.windygrid.find({
        "where.location": {
        "$within": {"$center": [[lat, lon], radius ]}}, 
        "type": aggregate_type, 
        distinct_key: distinct_type})
    return {distinct_type: nearby_requests.count()}
    
def sum_types_by_locations_in_query(search_type_query, search_mask, mask_name_key, distinct_key, aggregate_type, db, table):
    cursor = db.windygrid.find(search_type_query, search_mask)
    result = {}
    location_counts = []
    print "running against query results of %d records" % cursor.count()
    for doc in cursor:
        if doc['where']['location'] == [-1, 1]:
            print "invalid coordinates (-1, 1) for %s" % doc['_id']
            continue
        lat = doc['where']['location'][0]
        lon = doc['where']['location'][1]
        # within circle
        #nearby_requests = db.windygrid.find({
        #  "where.location": {
        #    "$within": {"$center": [[lat, lon], radius ]}}, 
        #    "type": aggregate_type})
        #
        # within bounding box
        ll_x = lon - math.degrees(bbox_dist_from_center / bbox_earth_radius / math.cos(math.radians(lat)))
        ll_y = lat - math.degrees(bbox_dist_from_center / bbox_earth_radius)
        ur_x = lon + math.degrees(bbox_dist_from_center / bbox_earth_radius / math.cos(math.radians(lat)))
        ur_y = lat + math.degrees(bbox_dist_from_center / bbox_earth_radius)
        print "bbox = lowerleft:[%f, %f], upperright[%f, %f]" % (ll_x, ll_y, ur_x, ur_y)
        box = [[ll_y, ll_x], [ur_y, ur_x]]
        nearby_requests = db.windygrid.find({"where.location": {"$within": {"$box" : box}}, "type": aggregate_type})
        count = nearby_requests.count()
        distinct_types = db.windygrid.find({"type": aggregate_type}) \
                           .distinct(distinct_key)
        counts = [sum_by_distinct_type(t, distinct_key, lat, lon) \
                  for t in distinct_types]
        type_counts = []
        for c in counts:
           keys = c.keys()
           for key in keys:
               if c[key] > 0:
                   type_counts.append({key: c[key]})

        print "processed %s; it had %d records near it" % \
            (doc[mask_name_key], count)
        location_counts.append({"what": doc[mask_name_key], 
                                "location": doc['where']['location'], 
                                "total_count": count, 
                                "type_counts": type_counts})

    result["location_counts"] = location_counts
    return result

connection = Connection()
db = connection["test"]
search_type_query = {"type" : "business_licenses"}
search_mask = {"_id" : 0, "what.legal_name" : 1, "where.location" : 1}
mask_name_key = "what"
distinct_key = "what.type_code"
aggregate_type = "311"
table = "windygrid"

result = sum_types_by_locations_in_query(search_type_query, search_mask, mask_name_key, distinct_key, aggregate_type, db, table)
#could save to a db here
print result