Python tree data library

I would like to determine the number of direct children (for non-leaf nodes). I imagine that unique identifiers will be needed then as references for relevant nodes. But the tree iteration is working with the attribute “name” (which might not be unique) so far. Thus I imagine also that the usage of paths would become helpful (so that an auxiliary node attribute can be avoided as long as pointer alternatives would not be applicable).

Will a configurable formatting become relevant for the display and further data processing of the calculated counts?

First of all, amazing work in here. I'm planing to use anytree to a huge kind of backtracking implementation. I will need to persist the tree to a file at some point, and of course, get the tree back from a file.

Are you planning to do it? Is it possible? Should i give it a try and then send the pull request?

Thank you

In the current implementation, there are some try-except blocks to check if an object has some attribute. To reduce the generated overhead by the exception handling logic, I propose to start using the getattr builtin function with a default value instead.

This change doesn't provide a big performance improvement, but it's a constant-factor cost removal with no public API implications.

To check the difference, I build a 1 million node random tree with a height of 25. The logic to build the anynode.AnyNode instances from the tree encoded as raw dictionaries decreased from ~11 seconds to ~8 seconds on my computer (not a big change, but better than nothing 🙂).

Works: 2.4.3 Breaks: 2.6.0

>>> from munch import Munch
>>> from anytree import NodeMixin
>>> class MyClass(Munch, NodeMixin):
...     pass
... 
Traceback (most recent call last):
  File "<stdin>", line 1, in <module>
TypeError: multiple bases have instance lay-out conflict

Where munch is https://github.com/Infinidat/munch

Googling around a bit this seems to be about conflicts in C implementations of classes on __foo__ attributes. So I did this:

>>> [k for k in dir(Munch) if k in dir(NodeMixin)]
['__class__', '__delattr__', '__dir__', '__doc__', '__eq__', '__format__', '__ge__', '__getattribute__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__sizeof__', '__str__', '__subclasshook__']

but it didn't make me feel much smarter.

Does this mean anything to you anytree peeps? I came here because obviously NodeMixin is meant to be mixed in so this feels like a bug rather than just a stupid thing I should not be attempting to do. Especially as it used to work until this version.

getting this error

FileNotFoundError: [Errno 2] No such file or directory: 'dot'

While I Run RenderTreeGraph(udo).to_picture("tree.png") Not building tree

Example:

from anytree import Node
top = Node("top", parent=None)
sub0 = Node("sub0", parent=top)
sub0sub0 = Node("sub0sub0", parent=sub0)
sub0sub1 = Node("sub0sub1", parent=sub0)
sub1 = Node("sub1", parent=top)

from anytree import Resolver
r = Resolver('name')

Getting the top node from sub0:

>>> r.get(sub0, '..')
Node('/top')

Using glob:

>>> r.glob(sub0, '..')
.../anytree/resolver.py in __glob(self, node, parts)
    165     def __glob(self, node, parts):
    166         nodes = []
--> 167         name = parts[0]
    168         remainder = parts[1:]
    169         # handle relative

IndexError: list index out of range

Hi, I would like to know if it's possible to remove children nodes from a Node with a lambda function criteria, something like:

outputnode = createnode(inputnode, filter_=lambda n: does_bbox_intersect(n.boundingbox, bounds))

This is just a small fix, but it allows to use the childiter function to not just sort, but also filter nodes when rendering. I need to "hide" some node types when rendering, and thus do this in childiter:

    def childiter(items):
        items = [i for i in items if i.type in lbuild.format.SHOW_NODES]
        return sorted(items, key=lambda item: (item.type, item.name))

However, this renders an additional node line without the node:

    ├── Module(modm:build)   Build System Generators
    │   ├── Module(modm:build:scons)   SCons Build Script Generator
    │   │   ╰── EnumerationOption(info.git) = Disabled in [Disabled, Info, Info+Status]
    ├── Module(modm:cmsis)   ARM CMSIS Support

After this fix:

    ├── Module(modm:build)   Build System Generators
    │   ╰── Module(modm:build:scons)   SCons Build Script Generator
    │       ╰── EnumerationOption(info.git) = Disabled in [Disabled, Info, Info+Status]
    ├── Module(modm:cmsis)   ARM CMSIS Support

cc @c0fec0de

Hallo, Just busy programming an Binary puzzle solver. What works: creating the nodes of the search tree for a solution works fine directly. But if i put the commands in a def ... The nodes are not to be found in command alinea of the notebook. It is probably my missing knowledge of ... How to get the nodes accessibke out of the function .. Help is apreciated

Is there any way to extract lowest common ancestor of nodes without iterating over the ancestor names and finding the common ancestor "manually"? For two nodes it's not too bad, but for more nodes I'm not actually sure how one would accomplish this.

Hi,

I am using anytree to generate a tree-view of pytest hooks based on a log file with indented hooks. https://github.com/pytest-dev/pytest/issues/3261

For that I needed to create a function to convert a list of indented strings to a tree. Apparently another users on stackoverflow have needed it: https://codereview.stackexchange.com/questions/176391/parsing-indented-text-and-inserting-its-data-into-a-tree https://stackoverflow.com/questions/17858404/creating-a-tree-deeply-nested-dict-from-an-indented-text-file-in-python https://stackoverflow.com/questions/32101253/converting-a-string-to-a-tree-structure-in-python?rq=1

I have created the following script for doing this and it is quite simple: https://github.com/Sup3rGeo/pytest/blob/master/doc/pytesthooks.py

I believe it makes sense to make this part of anytree as a "indented strings importer".

So if you agree I would gladly add a pull request for this. Thanks!

I am trying to load JSON data formatted as in the documentation and end up with

Traceback (most recent call last):
  File "d:\dev-pro\searchad\addm123.py", line 10, in <module>
    root = importer.import_(data)
  File "C:\Python310\lib\site-packages\anytree\importer\dictimporter.py", line 38, in import_
    return self.__import(data)
  File "C:\Python310\lib\site-packages\anytree\importer\dictimporter.py", line 47, in __import
    self.__import(child, parent=node)
  File "C:\Python310\lib\site-packages\anytree\importer\dictimporter.py", line 47, in __import
    self.__import(child, parent=node)
  File "C:\Python310\lib\site-packages\anytree\importer\dictimporter.py", line 47, in __import
    self.__import(child, parent=node)
  [Previous line repeated 5 more times]
  File "C:\Python310\lib\site-packages\anytree\importer\dictimporter.py", line 41, in __import
    assert isinstance(data, dict)
AssertionError

What does it mean, or more precisely - what kind of data can raise it?

~~I have empty children entries ("children": []), is this allowed? (just an idea of where the issue could come from)~~ removing "children" entries when empty does not change the issue.

per https://stackoverflow.com/questions/2358045/how-can-i-implement-a-tree-in-python

do you have functionality to count how many time particular branch was added

lets say marc = Node("Marc", parent=udo) lian = Node("Lian", parent=marc)

was added 12 times , then we want to see counts ?

I might misunderstand something, but the example for, e.g., leftsibling in the documentation of anytree/util/__init__.py states two things

1. It suggest only from anytree import Node
2. To directly call, e.g., leftsibling(dan)

For me, both did not work out. I had to

1. Import the method via from anytree.util import leftsibling
2. Do a, e.g., print(leftsibling(joe)) in order to get the desired output.

Taking the read-only example and implementing checks on post_attach instead of pre_attach, it does not undo the operation that is supposed to be atomic.

Sample code to replicate the issue below,

from anytree import NodeMixin


class SampleNode(NodeMixin):
    def __init__(self, foo, parent=None):
        super(SampleNode, self).__init__()
        self.foo = foo
        self.readonly = False
        self.parent = parent
    def _post_attach(self, parent):
        if self.root.readonly:
            raise ValueError()
    def _post_detach(self, parent):
        if self.root.readonly:
            raise ValueError()

a = SampleNode("a")
a0 = SampleNode("a0", parent=a)
a1 = SampleNode("a1", parent=a)
a1a = SampleNode("a1a", parent=a1)
a2 = SampleNode("a2", parent=a)

assert a1.parent == a

a.readonly = True
with pytest.raises(ValueError):
    a1.parent = a2

a1.parent == a2 # returns True, it should undo the operation

Hi there,

Hi there,

is there any easy way to create a tree from a python nested dictionary? For instance, I would like to turn: {'tax': {'a':{'c':{}, 'd':{}}, 'b': {'e':{}, 'f':{}}}}

tax --- a ------- c ------- d --- b ------- e ------- f

In general, I do not understand what format the json_importer takes as input. Could you please provide some examples? It there any online editor to first manually design the tree with nodes and arrows and then export it in the right format?

Thank you.

A few days ago I had an issue where anytree operations significantly slowed down over time. cProfile attributed the added time to this assertion. If I'm reading it correctly, it makes adding a child node take O(N) time, where N is the number of its siblings-to-be.

That N gets rather large in my case. Probably I should refactor my code to avoid that, but I felt I should report it anyway. Setting PYTHONOPTIMIZE=TRUE (to disable assertions) does work as a workaround.