Creating GUIDs in python using the first two chars to get a shard

guid_python.py

#!/usr/bin/env python

# Written by Warren Runk
# This file is free software in the public domain.


import base64
import uuid


"""
******* Deprecated. I went the route of explict shard in the node section.
See: https://gist.github.com/wrunk/b6d340297e7a9f3d97a0

Get a GUID - URL safe, Base32, all lowercase, looks like:

rwc5wtj73fdqjasiii5texzize
aruud7l6ijarfdk2a3kstwqmnu
p4zkxwv5hvfsbmp7o5b7cktb5i

We also provide the ability to get the bucket for the ID (0-1023)
"""


def _get_base_int(c):
    "Return the base int value of the base32 char (0-31)"
    c = ord(c)
    if c < 97:
        return c - 50
    return c - 91


def get_bucket_int(gid):
    """
    Return a base 10 integer representing the bucket of this guid. Will be between 0-1023 with decent
    distribution:

    Total # of bucket ids 1024
    [MAX] Buck id (0-1023) (642) # of ids generated in this bucket (1076)
    [MIN] Buck id (0-1023) (12) # of ids generated in this bucket (883)
    Median pos 511 977
    Median pos 512 977
    Median pos 513 977

    We take the first 2 character and count using their values. 
    Each char has 32 potential values it can be, so 32*32 = 1024. 
    Multiple the base of the first val by 32 and add to the second val.
    """
    return (_get_base_int(gid[0]) * 32) + _get_base_int(gid[1])


def get_new_guid():
    """
    Returns a length 26 globally unique ID.
    """
    return base64.b32encode(uuid.uuid4().bytes).lower().replace('=', '')


def print_bucket_distro_stats():
    from pprint import pprint
    meat_hap = {}
    for i in range(1000000):
        u = get_new_guid()
        buck = get_bucket_int(u)
        if buck not in meat_hap:
            meat_hap[buck] = 1
        meat_hap[buck] += 1

    pprint(meat_hap)
    print "Total # of bucket ids", len(meat_hap)
    mi = 20000
    mic = None
    ma = 0
    mac = None
    for k, v in meat_hap.items():
        if v < mi:
            mic = k
            mi = v
        if v > ma:
            mac = k
            ma = v
    print "[MAX] Buck id (0-1023) (%s) # of ids generated in this bucket (%i)" % (mac, ma)
    print "[MIN] Buck id (0-1023) (%s) # of ids generated in this bucket (%i)" % (mic, mi)
    print "Median pos 511", sorted(meat_hap.values())[511]
    print "Median pos 512", sorted(meat_hap.values())[512]
    print "Median pos 513", sorted(meat_hap.values())[513]


def print_sample():
    for i in range(20):
        u = get_new_guid()
        print "GUID: (%s) LEN (%i)" % (u, len(u))


if __name__ == '__main__':
    print_bucket_distro_stats()
    print_sample()