nested_dict 1.0.9

defaultdict extension for dictionaries with multiple levels of nesting

Hosted at http://code.google.com/p/nested-dict/ (Documentation at http://nested-dict.googlecode.com/hg/doc/build/html/index.html)

nested dictionaries

nested_dict extends collections.defaultdict to allow dictionaries with multiple levels of nesting to be created on the fly:

from nested_dict import *

nd = nested_dict()

nd["a"]["b"]["c"] = 311
nd["d"]["e"] = 311

Each nested level is create magically when accessed, a process known as “auto-vivification” in perl.

nested dictionaries of sets / lists and other collections

nested_dict also extends collections.defaultdict to allow dictionaries of lists, sets or other collections with a specified level of nesting level.

1) The old fashioned way using ugly syntax

d = dict()

d.setdefault(""1st group", []).append(3)
d.setdefault(""2nd group", []).append(5)
d.setdefault(""2nd group", []).append(8)
d.setdefault(""1st group", []).append(4)
d.setdefault(""2nd group", []).append(5)

2) default_dict adds lists automatically when required

from collections import defaultdict

dd = defaultdict(list)

dd["1st group"].append(3)
dd["2nd group"].append(5)
dd["2nd group"].append(8)
dd["1st group"].append(4)
dd["2nd group"].append(5)

3) nested_dict adds lists automatically when required for nested dictionaries

from nested_dict import nested_dict

# specify two levels of nesting
nd = nested_dict(2, list)

nd["1st group"]["subset a"].append(3)
nd["2nd group"]["subset a"].append(5)
nd["2nd group"]["subset b"].append(8)
nd["1st group"]["subset a"].append(4)
nd["2nd group"]["subset b"].append(5)

print nd

gives:

{'1st group': {'subset a': [3, 4]},
 '2nd group': {'subset b': [8, 5],
               'subset a': [5]}}

More examples:

“Auto-vivifying” nested dict

nd= nested_dict()
nd["mouse"]["chr1"]["+"] = 311
nd["mouse"]["chromosomes"]="completed"
nd["mouse"]["chr2"] = "2nd longest"
nd["mouse"]["chr3"] = "3rd longest"

for k, v in nd.iteritems_flat():
     print "%-30s=-%20s" % (k,v)

Gives

('mouse', 'chr3')             =-         3rd longest
('mouse', 'chromosomes')      =-           completed
('mouse', 'chr2')             =-         2nd longest
('mouse', 'chr1', '+')        =-                 311

Specifying the autovivified object

If you wish the nested dictionary to hold a collection rather than a scalar, you have to write:

nd["mouse"]["chr2"] = list()
nd["mouse"]["chr2"].append(12)

or:

nd["mouse"]["chr2"] = set()
nd["mouse"]["chr2"].add(84)

Which doesn’t seem very “auto” at all!

Instead, specify the collection in the constructor of nested_dict:

# two levels of nesting
nd2 = nested_dict(2, list)
nd2["mouse"]["chr2"].append(12)

# three levels of nesting
nd3 = nested_dict(3, set)
nd3["mouse"]["chr2"]["categorised"].add(3)

# counts
nd4 = nested_dict(2, int)
nd4["mouse"]["chr2"]+=4
nd4["human"]["chr1"]+=3
nd4["human"]["chr3"]+=4