Ancestral recombination graphs (ARGs)は、組換えを起こす系統間の遺伝的関係を完全に表現したものであり、集団遺伝学において中心的な重要性を担っている。近年のシミュレーションおよび推論手法の飛躍的進歩により、ARGへの関心が高まっている。しかし、ARGのグラフィカルな構造を最大限に活用する方法は、研究者にとって依然として未解決の問題である。本稿では、インタラクティブなD3.js可視化ライブラリを用いてプログラム的にARGを描画するためのPythonパッケージ、tskit_arg_visualizerを紹介する。この可視化ツールは、ARGの概念の教育と、経験的データセットから推論されたARGの探索の両方において有用であることを示す。tskit_arg_visualizerの最新の安定版は、Python Package Index(https://pypi.org/project/tskit-arg-visualizer、現在v0.1.1)から入手できる。ドキュメントとパッケージの開発バージョンは、GitHub (https://github.com/kitchensjn/tskit_arg_visualizer) にある。
Tutorial
https://github.com/kitchensjn/tskit_arg_visualizer/blob/main/docs/tutorial.md
インストール
Github
mamba create -n tskit_arg_visualizer python=3.13 -y
conda activate tskit_arg_visualizer
pip install tskit_arg_visualizer
pip install jupyterlab
#Github
git clone https://github.com/kitchensjn/tskit_arg_visualizer.git
cd tskit_arg_visualizer
python -m pip install --upgrade pip
pip install -e .
pip install jupyterlab
実行方法
tskit_arg_visualizerは、tskitのTreeSequenceやJSONからARGを読み込み、ブラウザまたはJupyter Notebook上に組換えを起こす系統間の遺伝的関係を描画するツール。使い方を簡単に見ていきます。通常、jupyter labで進めます。
テストラン
1、jupyter立ち上げ、python環境で進める
jupyter lab
2、テストARG(Ancestral recombination graphs)を作成
import msprime
import tskit_arg_visualizer
ts = msprime.sim_ancestry(
samples=5,
recombination_rate=1e-8,
sequence_length=3000,
population_size=10000,
record_full_arg=True,
random_seed=1234
)
ts = msprime.sim_mutations(
ts,
rate=1e-7,
random_seed=4321
)
print(ts)

3、D3ARG形式に変換
d3arg = tskit_arg_visualizer.D3ARG.from_ts(
ts=ts,
progress=True
)

4、描画
d3arg.draw(
edge_type="line",
show_mutations=True,
rotate_tip_labels=True,
width=800,
height=600
)
ARGのグラフが出力される。ARGは、組換えを含むゲノムの祖先関係を、1本の系統樹ではなくグラフとして表したもの。

通常の系統樹なら1つのツリーとして読むが、ARGでは組換えがあるため、ゲノムの場所によって系統関係が変わる。
下側のノードは現在のサンプル、上側のノードは過去の祖先を表す。左軸は世代単位の時間であり、上に行くほど古い祖先になる。黒線は祖先から子孫への継承関係を示し、黄色の横線は変異を表す。下部のバーはゲノム座標で、今回の例では0から3000までの領域を表示している。下部のバーの区切りは、組換えなどによってlocal treeが切り替わるbreakpointを示している。
ゲノムバーにマウスを重ねると、その区間に対応するlocal treeが強調表示され、ゲノム位置ごとに系統関係が変化することを確認できる。

実データで使う場合は、VCFから直接利用できない。ARGを見たい場合は、まずARG推定ツールでTreeSequenceに変換する。推奨はtskit TreeSequenceを出力できるtsinferとtsdateの組み合わせで、これらのツールでVCFからTreeSequenceを作成、さらにdated TreeSequenceを得てtskit_arg_visualizerで可視化する。
論文より、
- tskit_arg_visualizer パッケージは tskit と、より広範な他のパッケージをベースに構築されている。tskit_arg_visualizer は独自の形式を採用しており、tskit とは完全に独立して使用できるが、両者には多くの概念的な類似点があり、両者はシームレスに統合されるように設計されている。
引用
tskit_arg_visualizer: interactive plotting of ancestral recombination graphs Open Access
James Kitchens , Yan Wong
Bioinformatics Advances, Volume 5, Issue 1, 2025
> import tskit_arg_visualizer
> help(tskit_arg_visualizer)
help
Help on package tskit_arg_visualizer:
NAME
tskit_arg_visualizer
PACKAGE CONTENTS
CLASSES
builtins.object
D3ARG
DrawInfo
IncludedObjects
class D3ARG(builtins.object)
| D3ARG(nodes, edges, mutations, breakpoints, num_samples, sample_order, default_node_style, time_units)
|
| Stores the ARG in a D3.js friendly format ready for plotting
|
| See 'Alternative Constructors' for common ways of creating this object
|
| Attributes
| ----------
| nodes : list
| List of node dicts that contain info about the nodes
| edges : list
| List of edge dicts that contain info about the edges
| breakpoints : list
| List of breakpoint dicts that contain info about the breakpoints
| num_samples : int
| The number of samples in the ARG (with (ts_flags & 1) == 1)
| sample_order : list
| Ordered list of sample IDs
|
| Alternative Constructors
| ------------------------
| from_ts(ts)
| Creates a D3ARG from a tskit.TreeSequence
|
| from_json(json)
| Creates a D3ARG from a saved custom JSON
|
| Methods
| -------
| draw(
| width=500,
| height=500,
| tree_highlighting=True,
| y_axis_labels=True,
| y_axis_scale="rank",
| line_type="ortho",
| subset_nodes=[],
| include_node_labels=True
| )
| Draws the ARG using D3.js
|
| set_node_labels(labels={})
| Customizes node labels in visualization
|
| reset_node_labels()
| Sets the node labels back to default values
|
| Methods defined here:
|
| __init__(self, nodes, edges, mutations, breakpoints, num_samples, sample_order, default_node_style, time_units)
| Initializes a D3ARG object
|
| This is the generalized function for initializing a D3ARG object. It is most
| often called by another method, such as from_ts() or from_json(), though it
| can be used separately if the parameters are in the correct format.
|
| Parameters
| ----------
| nodes : pandas.DataFrame
| Contains info about the nodes
| edges : pandas.DataFrame
| Contains info about the edges
| breakpoints : pandas.DataFrame
| Contains info about the breakpoints
| num_samples : int
| Number of samples in the ARG. Useful for various calculations when plotting
| sample_order : list or np.array
| Order of the sample ID from left to right when plotting
| default_node_style : dict
| time_units : string
| Time units for the node times (example is "generations").
|
| __str__(self)
| Prints attributes of D3ARG object
|
| draw(self, width=500, height=500, tree_highlighting=True, y_axis_labels=True, y_axis_title=None, y_axis_scale='rank', edge_type='line', variable_edge_width=False, include_underlink=True, sample_order=None, title=None, show_mutations=False, ignore_mutation_times=True, label_mutations=False, condense_mutations=False, is_notebook=None, rotate_tip_labels=False, zoom=0, styles=None, preamble=None, save_filename=None)
| Draws the D3ARG using D3.js by sending a custom JSON object to visualizer.js
|
| Parameters
| ----------
| width : int
| Width of the force layout graph plot in pixels (default=500)
| height : int
| Height of the force layout graph plot in pixels (default=500)
| tree_highlighting : bool
| Include the interactive chromosome at the bottom of the figure to
| to let users highlight trees in the ARG (default=True)
| y_axis_labels : bool, list, or dict
| Whether to include the y-axis on the left of the figure. By default, tick marks will be automatically
| chosen. You can specify a list of tick marks to use instead. You can also set custom text for tick marks
| using a dictionary where key is the time and value is the text. (default=True)
| y_axis_title : string
| Title of the y-axis that will be plotted on the left hand side of the figure. (default=None, uses
| `D3ARG.time_units` attribute)
| y_axis_scale : string
| Scale used for the positioning nodes along the y-axis. Options:
| "rank" (default) - equal vertical spacing between nodes
| "time" - vertical spacing is proportional to the time
| "log_time" - proportional to the log of time
| edge_type : string
| Pathing type for edges between nodes. Options:
| "line" (default) - simple straight lines between the nodes
| "ortho" - custom pathing (see pathing.md for more details, should only be used with full ARGs)
| variable_edge_width : bool
| Scales the stroke width of edges in the visualization will be proportional to the fraction of
| sequence in which that edge is found. (default=False)
| include_underlink : bool
| Includes an "underlink" for each edge gives a gap during edge crosses. This is currently only
| implemented for `edge_type="ortho"`. (default=True)
| sample_order : list
| Sample nodes IDs in desired order. Must only include sample nodes IDs, but does not
| need to include all sample nodes IDs. (default=None, order is set by first tree in tree sequence)
| title : str
| Title to be put at the top of the figure. (default=None, ignored)
| show_mutations : bool
| Whether to add mutations to the graph. Only available when `edge_type="line"`. (default=False)
| ignore_mutation_times : bool
| Whether to plot mutations evenly on edge (True) or at there specified times (False). (default=True, ignored)
| label_mutations : bool
| Whether to add the full label (position_index:inherited:derived) for each mutation. (default=False)
| condense_mutations : bool
| Whether to merge all mutations along an edge into a single mutation symbol. (default=False)
| is_notebook : bool
| Should the visualizer assume a notebook environment or show the visualization
| in a standalone HTML page? By default, the function will attempt to autodetect
| whether it is being called in a notebook environment. This may not work in
| some untested environments, in which case you may wish to set this explicitly
| to True (to force a notebook display) or False (to force a standalone HTML page).
| rotate_tip_labels : bool
| Rotates tip labels by 90 degrees. (default=False)
| zoom : int
| The level of detail that you want. Larger numbers equate to less detail/more collapsing
| styles : list
| A list of css strings, one per selector. The ID of the current drawing will be
| appended to each string, so that the styles are unique to the current drawing.
| For example, [".labels {font-family: Times}"] will change the font of all the
| labels. Note that some styles are set from values in the dataframes stored in the
| D3ARG object, and cannot be altered using the 'styles' parameter. (default=None)
| preamble : str
| Extra HTML to be added to the D3ARG visualization. It is the user's
| responsibility to ensure that this is valid HTML. The preamble can be used
| to add additional information, such as a legend, to the visualization. To
| overlay any content on top of the plot, you may wish to wrap the preamble in a
| tag which has the "position: absolute" and "z-index: 1" styles set, e.g.
| `preamble='<div style="position:absolute; z-index:1; right:2em">My text</div>'`
| (default="")
| save_filename : str
| Filename to use when selecting "Download as" in the visualization
| (default=None, treated as "tskit_arg_visualizer")
|
| Returns
| -------
| info : DrawInfo
| An object with information about the drawing, for example
| the width and height and the objects (such as which nodes) which
| were included in the drawing.
|
| draw_genome_bar(self, width=500, windows=None, show_mutations=False, is_notebook=None)
| Draws a genome bar for the D3ARG using D3.js
|
| Parameters
| ----------
| width : int
| Width of the force layout graph plot in pixels (default=500)
| windows : list of lists
| Each list is are the start and end positions of the windows. Multiple windows can be included.
| (Default is None, ignored)
| show_mutations : bool
| Whether to add ticks for mutations along the genome bar
| is_notebook : bool
| Should the visualizer assume a notebook environment or show the visualization
| in a standalone HTML page? By default, the function will attempt to autodetect
| whether it is being called in a notebook environment. This may not work in
| some untested environments, in which case you may wish to set this explicitly
| to True (to force a notebook display) or False (to force a standalone HTML page).
|
| draw_node(self, seed_nodes, width=500, height=500, depth=1, y_axis_labels=True, y_axis_title=None, y_axis_scale='rank', tree_highlighting=True, title=None, show_mutations=False, ignore_mutation_times=True, label_mutations=False, condense_mutations=False, is_notebook=None, rotate_tip_labels=False, styles=None, preamble=None, save_filename=None)
| Draws a subgraph of the D3ARG using D3.js by sending a custom JSON object to visualizer.js.
|
| Parameters
| ----------
| seed_nodes : int or list
| Node ID or list of node IDs that will be central to the subgraph
| width : int
| Width of the force layout graph plot in pixels (default=500)
| height : int
| Height of the force layout graph plot in pixels (default=500)
| depth : int or list(int, int)
| Number of nodes above (older than) and below (younger than) the central
| node to include in the subgraph (default=1). If this is a list, the
| number of nodes above is taken from the first element and
| the number of nodes below from the last element.
| y_axis_labels : bool, list, or dict
| Whether to include the y-axis on the left of the figure. By default, tick marks will be automatically
| chosen. You can specify a list of tick marks to use instead. You can also set custom text for tick marks
| using a dictionary where key is the time and value is the text. (default=True)
| y_axis_title : string
| Title of the y-axis that will be plotted on the left hand side of the figure. (default=None, uses
| `D3ARG.time_units` attribute)
| y_axis_scale : string
| Scale used for the positioning nodes along the y-axis. Options:
| "rank" (default) - equal vertical spacing between nodes
| "time" - vertical spacing is proportional to the time
| "log_time" - proportional to the log of time
| tree_highlighting : bool
| Include the interactive chromosome at the bottom of the figure to
| to let users highlight trees in the ARG (default=True)
| title : str
| Title to be put at the top of the figure. (default=None, ignored)
| show_mutations : bool
| Whether to add mutations to the graph. (default=False)
| ignore_mutation_times : bool
| Whether to plot mutations evenly on edge (True) or at there specified times (False). (default=True, ignored)
| label_mutations : bool
| Whether to add the full label (position_index:inherited:derived) for each mutation. (default=False)
| condense_mutations : bool
| Whether to merge all mutations along an edge into a single mutation symbol. (default=False)
| is_notebook : bool
| Should the visualizer assume a notebook environment or show the visualization
| in a standalone HTML page? By default, the function will attempt to autodetect
| whether it is being called in a notebook environment. This may not work in
| some untested environments, in which case you may wish to set this explicitly
| to True (to force a notebook display) or False (to force a standalone HTML page).
| rotate_tip_labels : bool
| Rotates tip labels by 90 degrees. (default=False)
| styles : list
| A list of css strings, one per selector. The ID of the current drawing will be
| appended to each string, so that the styles are unique to the current drawing.
| For example, [".labels {font-family: Times}"] will change the font of all the
| labels. Note that some styles are set from values in the dataframes stored in the
| D3ARG object, and cannot be altered using the 'styles' parameter. (default=None)
| preamble : str
| Extra HTML to be added to the D3ARG visualization. It is the user's
| responsibility to ensure that this is valid HTML. The preamble can be used
| to add additional information, such as a legend, to the visualization. To
| overlay any content on top of the plot, you may wish to wrap the preamble in a
| tag which has the "position: absolute" and "z-index: 1" styles set, e.g.
| `preamble='<div style="position:absolute; z-index:1; right:2em">My text</div>'`
| (default="")
| save_filename : str
| Filename to use when selecting "Download as" in the visualization
| (default=None, treated as "tskit_arg_visualizer")
|
| Returns
| -------
| info : DrawInfo
| An object with information about the drawing, for example
| the width and height and the objects (such as which nodes) which
| were included in the drawing.
|
| draw_nodes = draw_node(self, seed_nodes, width=500, height=500, depth=1, y_axis_labels=True, y_axis_title=None, y_axis_scale='rank', tree_highlighting=True, title=None, show_mutations=False, ignore_mutation_times=True, label_mutations=False, condense_mutations=False, is_notebook=None, rotate_tip_labels=False, styles=None, preamble=None, save_filename=None)
|
| reset_all_breakpoint_fills(self)
| Sets the fill of genome bar blocks to the specified color
|
| reset_all_edge_colors(self)
| Resets the edge strokes to the default (#053e4e)
|
| reset_all_mutation_styles(self)
| Resets mutation styles to default
|
| reset_all_node_labels(self)
| Resets node labels to default (based on msprime IDs)
|
| reset_all_node_styles(self)
| Resets node styles to default (same as when assigned using D3ARG.from_ts)
|
| WARNING: This might not match the initial styles if using D3ARG.from_json
|
| set_all_breakpoint_fills(self, color)
| Sets the fill of genome bar blocks to the specified color
|
| set_all_edge_colors(self, color)
| Sets the edge strokes to the specified color
|
| set_all_mutation_styles(self, *, size=None, fill=None, stroke=None)
| Set all at once
|
| set_all_node_styles(self, size=None, symbol=None, fill=None, stroke=None, stroke_width=None)
| Sets the styling of all of the nodes at once for a specific option.
|
| If optional parameter not provided, that styling option will be ignored and unchanged.
|
| Parameters
| ----------
| size : int
| Size in pixels of the node
| symbol : string
| D3 symbol (see https://d3js.org/d3-shape/symbol)
| fill : string
| Color of the node, "#XXXXXX" form
| stroke : string
| Color of the stroke around the node, "#XXXXXX" form
| stroke_width : int
| Pixel width for the stroke around the node
|
| set_breakpoint_fills(self, colors)
| Set the fill of each breakpoint block in the ARG
|
| Parameters
| ----------
| colors : dict
| ID of the edge and its new color
|
| set_edge_colors(self, colors)
| Set the color of each edge in the ARG
|
| Parameters
| ----------
| colors : dict
| ID of the edge and its new color
|
| set_mutation_styles(self, styles)
| Individually control the styling of each mutation.
|
| Parameters
| ----------
| styles : dict
| A dictionary whose keys are mutation ids, and whose values are separate
| dictionaries containing any of the styling keys: size, fill, stroke.
| Only mutations that need styles updated need to be provided, and all
| styling keys are optional.
|
| set_node_labels(self, labels)
| Sets custom node labels
|
| Updates node labels based on the D3ARG node "id" using the labels dictionary.
| Final labels will always be strings. Do not rely on the ordering of the
| labels dictionary.
|
| Parameters
| ----------
| labels : dict
| ID of the node and its new label
|
| set_node_styles(self, styles)
| Individually control the styling of each node.
|
| Parameters
| ----------
| styles : dict
| A dictionary whose keys are node ids, and whose values are separate dicts
| containing any of the styling keys: size, symbol, fill, stroke, stroke_width.
| Only nodes that need styles updated need to be provided, and all styling
| keys are optional.
|
| set_node_x_positions(self, pos)
| Sets the x-axis positions of nodes
|
| Positions must be between 0 and 1
|
| Parameters
| ----------
| pos : dict
| ID of the node and its x-axis position (scaled between 0 and 1)
|
| subset_graph(self, seed_nodes, depth)
| Subsets the graph to focus around a specific node
|
| Parameters
| ----------
| seed_nodes : int or list
| Node ID or list of node IDs that will be central to the subgraph
| depth : int or list(int, int)
| Number of nodes above (older than) and below (younger than) the central
| node to include in the subgraph (default=1). If this is a list, the
| number of nodes above is taken from the first element and
| the number of nodes below from the last element.
|
| Returns
| -------
| included_nodes : pd.DataFrame
| The nodes to be plotted, potentially subset of original graph
| included_edges : pd.DataFrame
| The edges to be plotted, potentially subset of original graph
| included_mutations : pd.DataFrame
| The mutations to be plotted, potentially subset of original graph
| included_breakpoints : pd.DataFrame
| The breakpoints to be plotted, potentially subset of original graph
|
| ----------------------------------------------------------------------
| Class methods defined here:
|
| from_json(json)
| Converts a saved custom JSON into the D3ARG object
|
| Parameters
| ----------
| json : list of dictionaries
| the custom output of that is downloaded from the visualizer.
| See plotting.md for more details on the format of the json structure
|
| Returns
| -------
| D3ARG : a corresponding D3ARG object ready to be plotted
|
| from_ts(ts, ignore_unattached_nodes=False, progress=None, default_node_style=None)
| Converts a tskit tree sequence into a D3ARG object
|
| Parameters
| ----------
| ts : tskit.TreeSequence
| Note: if tree sequence marks recombination nodes, it must use 2-RE format similar to that
| used by msprime.sim_ancestry(...,record_full_arg=True).
| ignore_unattached_nodes : bool
| Whether to include all nodes or ignore nodes that are completely
| unattached. Default is False.
| progress : bool
| Show progress bars during conversion
| default_node_style : dict
| Customizable node stylings that will be set as default. Options include size, symbol, fill, stroke, and stroke_width
|
| Returns
| -------
| D3ARG : a corresponding D3ARG object ready to be plotted
|
| ----------------------------------------------------------------------
| Data descriptors defined here:
|
| __dict__
| dictionary for instance variables
|
| __weakref__
| list of weak references to the object
class DrawInfo(builtins.object)
| DrawInfo(width: int, height: int, uid: str, is_notebook: bool, included: tskit_arg_visualizer.IncludedObjects) -> None
|
| Stores information returned by drawing an ARG.
|
| Methods defined here:
|
| __eq__(self, other)
| Return self==value.
|
| __init__(self, width: int, height: int, uid: str, is_notebook: bool, included: tskit_arg_visualizer.IncludedObjects) -> None
| Initialize self. See help(type(self)) for accurate signature.
|
| __repr__(self)
| Return repr(self).
|
| ----------------------------------------------------------------------
| Data descriptors defined here:
|
| __dict__
| dictionary for instance variables
|
| __weakref__
| list of weak references to the object
|
| ----------------------------------------------------------------------
| Data and other attributes defined here:
|
| __annotations__ = {'height': <class 'int'>, 'included': <class 'tskit_...
|
| __dataclass_fields__ = {'height': Field(name='height',type=<class 'int...
|
| __dataclass_params__ = _DataclassParams(init=True,repr=True,eq=True,or...
|
| __hash__ = None
|
| __match_args__ = ('width', 'height', 'uid', 'is_notebook', 'included')
class IncludedObjects(builtins.object)
| IncludedObjects(nodes: list[int] | None = None) -> None
|
| Stores the IDs of the various objects included in a drawing.
|
| Methods defined here:
|
| __eq__(self, other)
| Return self==value.
|
| __init__(self, nodes: list[int] | None = None) -> None
| Initialize self. See help(type(self)) for accurate signature.
|
| __repr__(self)
| Return repr(self).
|
| ----------------------------------------------------------------------
| Data descriptors defined here:
|
| __dict__
| dictionary for instance variables
|
| __weakref__
| list of weak references to the object
|
| ----------------------------------------------------------------------
| Data and other attributes defined here:
|
| __annotations__ = {'nodes': list[int] | None}
|
| __dataclass_fields__ = {'nodes': Field(name='nodes',type=list[int] | N...
|
| __dataclass_params__ = _DataclassParams(init=True,repr=True,eq=True,or...
|
| __hash__ = None
|
| __match_args__ = ('nodes',)
|
| nodes = None
FUNCTIONS
calculate_evenly_distributed_positions(num_elements, start=0, end=1, round_to=0)
Returns a list of `num_elements` evenly distributed positions on a given `length`
Parameters
----------
num_elements : int
Number of positions to be returned
start : int
Start of position range
end : int
End of position range
round_to : int
The number of decimals to round to
Returns
-------
List of float positions
convert_time_to_position(t, min_time, max_time, scale, unique_times, h_spacing, height, y_shift=0)
Calculates a y-axis position corresponding to a time on various axis scales
Parameters
----------
t : int or float
Time to convert
min_time : int or float
Minimum time along axis
max_time
Maximum time along axis
scale : str
Axis scale to use (Options: "rank", "time", or "log_time).
unique_times : list
All of the node and mutation times used for determining rank of time.
h_spacing : float
Gap between ranks
height : int or float
Height of the visualizer
y_shift : int or float
Translation to add to the position. (default=0)
draw_D3(arg_json, styles=None, is_notebook=None)
extract_x_positions_from_json(arg_json)
Extracts the x position of nodes from json loaded from a saved ARG. This
can be passed into D3ARG.set_node_x_positions to retain the same x positions.
Parameters
----------
arg_json : json
The loaded json from the saved ARG
Returns
-------
x_pos_01 : dict
Nodes' x-axis positions scaled between 0 and 1
Examples
--------
>>> import json
>>> from pathlib import Path
>>> d3arg.set_node_x_positions( # assume a D3ARG object named d3arg
... pos=extract_x_positions_from_json(
... json.loads(Path("tskit_arg_visualizer.json").read_text())
... )
... )
>>> d3arg.draw() # draw, with fixed nodes for loaded x positions
map_value(n, start1, stop1, start2, stop2)
Map a value to a new range
From SO: https://stackoverflow.com/questions/44338698/p5-js-map-function-in-python
Parameters
----------
n : int or float
start1 : int or float
stop1 : int or float
start2 : int or float
stop2 : int or float
Returns
-------
mapped
running_in_notebook()
Checks whether the code is being executed within a Jupyter Notebook.
Adapted from https://stackoverflow.com/questions/15411967/how-can-i-check-if-code-is-executed-in-the-ipython-notebook
Returns
-------
bool
True if being executed within a Jupyter Notebook, False otherwise
DATA
default_mutation_styles = {'condensed': {'fill': 'pink', 'stroke': '#0...
VERSION
Using __version__ is deprecated. Please use `importlib.metadata.version('tskit_arg_visualizer')` instead
FILE
/Users/kazu/Downloads/clinker/tskit_arg_visualizer/tskit_arg_visualizer/__init__.py