Skip to main content

A python module to read and write the Newick format

Project description


Build Status PyPI

python package to read and write the Newick format.

Reading Newick

Since Newick specifies a format for a set of trees, all functions to read Newick return a list of newick.Node objects.

  • Reading from a string:

    >>> from newick import loads
    >>> trees = loads('(A,B,(C,D)E)F;')
    >>> trees[0].name
    >>> [ for n in trees[0].descendants]
    ['A', 'B', 'E']
  • Reading from a file-like object:

    >>> import io
    >>> from newick import load
    >>> with'fname', encoding='utf8') as fp:
    ...     trees = load(fp)
  • Reading from a path:

    >>> from newick import read
    >>> trees = read('fname')
    >>> import pathlib
    >>> trees = read(pathlib.Path('fname'))

Supported Newick dialects

While the set of reserved characters in Newick (;(),:) is relatively small, it's still often seen as too restrictive, in particular when it comes to adding more data to tree nodes. Thus, Newick provides two mechanisms to overcome this restriction:

  • quoted labels to allow arbitrary text as node names,
  • comments enclosed in square brackets.

Quoted node labels

Node labels in Newick may be quoted (i.e. enclosed in single quotes ') to make it possible to add characters which are otherwise reserved. The newick package supports quoted labels.

>>> from newick import loads
>>> print(loads("('A:B','C''D')'E(F)'")[0].ascii_art())

When creating Newick trees programmatically, names can be quoted (if necessary) automatically:

>>> from newick import Node
>>> print(Node("A(F')", auto_quote=True).name)
>>> print(Node("A(F')", auto_quote=True).unquoted_name)

Note: newick provides no support to parse structured data from node labels (as it can be found in the trees distributed by the Genome Taxonomy Database).

Additional information in comments

The "Newick specification" states

Comments are enclosed in square brackets and may appear anywhere

This has spawned a host of ad-hoc mechanisms to insert additional data into Newick trees.

The newick package allows to deal with comments in two ways.

  • Ignoring comments:
    >>> newick.loads('[a comment](a,b)c;', strip_comments=True)[0].newick
  • Reading comments as node annotations: Several software packages use Newick comments to store node annotations, e.g. *BEAST, MrBayes or TreeAnnotator. Provided there are no comments in places where they cannot be interpreted as node annotations, newick supports reading and writing these annotations:
    >>> newick.loads('(a[annotation],b)c;')[0].descendants[0].name
    >>> newick.loads('(a[annotation],b)c;')[0].descendants[0].comment
    >>> newick.loads('(a[annotation],b)c;')[0].newick
    Annotations may come before and/or after the : which separates node label and length:
  • >>> newick.loads('(a[annotation]:2,b)c;')[0].descendants[0].length
    >>> newick.loads('(a:[annotation]2,b)c;')[0].descendants[0].length
    >>> newick.loads('(a[annotation1]:[annotation2]2,b)c;')[0].descendants[0].comments
    ['annotation1', 'annotation2']

Note that square brackets inside quoted labels will not be interpreted as comments or annotations:

>>> newick.loads("('a[label]',b)c;")[0].descendants[0].name
>>> newick.loads("('a[label]',b)c;")[0].newick

Some support for reading key-value data from node comments is available as well. If the comment format follows the NHX spec or the &<key>=<value>,...-format used e.g. by the MrBayes or BEAST software, additional data can be accessed from the dict

>>> newick.loads('(A,B)C[&&NHX:k1=v1:k2=v2];')[0].properties
{'k1': 'v1', 'k2': 'v2'}

Note that we still don't support typed node properties. I.e. values in are always strings. Since typed properties tend to be specific to the application writing the newick, this level of support would require more knowledge of the creation context of the tree than can safely be inferred from the Newick string alone.

>>> newick.loads('(A,B)C[&range={1,5},support="100"];')[0].properties
{'range': '{1,5}', 'support': '"100"'}

Writing Newick

In parallel to the read operations there are three functions to serialize a single Node object or a list of Node objects to Newick format:

  • dumps(trees) -> str
  • dump(trees, fp)
  • write(trees, 'fname')

A tree may be assembled using the factory methods of the Node class:

  • Node.__init__
  • Node.create
  • Node.add_descendant

Manipulating trees

  • Displaying tree topology in the terminal:
    >>> import newick
    >>> tree = newick.loads('(b,(c,(d,(e,(f,g))h)i)a)')[0]
    >>> print(tree.ascii_art())
  • Pruning trees: The example below prunes the tree such that b, c and i are the only remaining leafs.
    >>> tree.prune_by_names(['b', 'c', 'i'], inverse=True)
    >>> print(tree.ascii_art())
  • Running a callable on a filtered set of nodes:
    >>> tree.visit(lambda n: setattr(n, 'name',, lambda n: in ['a', 'b'])
    >>> print(tree.ascii_art())
  • Removing (topologically) redundant internal nodes:
    >>> tree.prune_by_names(['B', 'c'], inverse=True)
    >>> print(tree.ascii_art())
        └─A ──c
    >>> tree.remove_redundant_nodes(keep_leaf_name=True)
    >>> print(tree.ascii_art())

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

newick-1.7.0.tar.gz (21.4 kB view hashes)

Uploaded Source

Built Distribution

newick-1.7.0-py2.py3-none-any.whl (15.2 kB view hashes)

Uploaded Python 2 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