Skip to main content

An interval tree data structure

Project description

Build Status License: MIT Coverage Status


itree is an interval tree data structure based on a self-balancing AVL binary search tree. Suitable for use with sequence features in bioinformatics.

Why itree?

  • itree is fast

itree implements an augmented search tree optmized for searching sets of intervals. The following benchmarks the performance of inserting, removing and searching for random intervals taken from the human chromosome 12 Gencode genes[1]:

  • itree is convenient

itree has a second-level interface for groups of objects (e.g. chromosomes):

>>> import itree, collections
>>> bed_records = [tuple(l.split()[:3]) for l in open('gencode.bed')] 
>>> i = collections.namedtuple('MyInterval', ['chrom','start','end'])
>>> t = itree.GroupedITree('chrom', [i(f[0], int(f[1]), int(f[2])) for f in bed_records])
>>>'chr15', 45167200, 45167300)) 
[MyInterval(chrom='chr15', start=45167213, end=45187956),
 MyInterval(chrom='chr15', start=45167250, end=45187952),
 MyInterval(chrom='chr15', start=45167213, end=45201175),
 MyInterval(chrom='chr15', start=45152663, end=45167526)]

Getting Started

  • Construction

Creating an interval tree object:

>>> import itree
>>> t = itree.ITree()
  • Insertion

Any item inserted into an interval tree must contain "start" and "end" attributes as integers.

>>> import collections
>>> i = collections.namedtuple('MyInterval', ['start','end'])
>>> t.insert(i(1,15))
>>> t.insert(i(3,20))
>>> t.insert(i(4,20))
>>> t.insert(i(5,15))
>>> t.insert(i(6,7))
  • Search

Search for all intervals overlapping a given interval

[MyInterval(start=3, end=20), MyInterval(start=4, end=20), MyInterval(start=1, end=15)]
  • Removal

Remove an interval exactly matching the given interval by its start and end attributes (but not necessarily the same object).

>>> t.pstring()

>>> t.remove(i(1,15))
>>> t.pstring()

The pstring method is mostly for debugging, but here we illustrate the rebalancing of the tree.

  • Grouping

A second-level itree object, GroupedITree, works as a proxy to itree objects which can be grouped by any hashable attribute or function:

>>> import itree, collections
>>> i = collections.namedtuple('Appointment', ['day','start','end'])
>>> appts = [i('Monday', 9, 13), i('Monday', 16, 17), i('Tuesday', 14, 15)]
>>> t = itree.GroupedITree(key='day', intervals=appts)
>>>'Monday', 11, 12))
[Appointment(day='Monday', start=9, end=13)]
>>>'Monday', 14, 15))

You may also use any arbitrary hashable value returned from a function as a key:

>>> i = collections.namedtuple('Appointment', ['day','month','start','end'])
>>> date_key = lambda appt: "{} {}".format(, appt.month)
>>> appts = [i(5, 'Jan', 9, 13), i(6, 'Jan', 16, 17), i(5, 'Feb', 14, 15)]
>>> t = itree.GroupedITree(key=date_key, intervals=appts)
>>>, 'Jan', 16, 17))   


[1]. generated with python3 scripts/ scripts/gencode.chr12.bed 500 10000 500 > benchmarking.txt.

Project details

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

itree-0.0.5.tar.gz (7.2 kB view hashes)

Uploaded Source

Built Distribution

itree-0.0.5-py3-none-any.whl (8.1 kB view hashes)

Uploaded Python 3

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page