TMultiGraph writing

[Pepesob]

Hello,

Recently, I was looking into implementing TMultiGraph writing; however, I got kind of stuck. When previously implementing TGraph writing, there was already an existing model class (for TGraphErrors and TGraphAsymmErrors too, although serialization for those still needs to be done). For TMultiGraph, I can't seem to find one.

I tried to read a TMultiGraph from an existing .root file. I ran this code to see available methods, fields, and members of f["myMultiGraph"]:

with uproot.open("tmultigraph_example.root") as f:
    p = f["myMultiGraph"]

    print(p)
    print(dir(p))
    print(p.members)

Output:

<TMultiGraph (version 2) at 0x017791151430>
['__class__', '__delattr__', '__dict__', '__dir__', '__doc__', '__enter__', '__eq__', '__exit__', '__format__', '__ge__', '__getattribute__', '__getstate__', '__gt__', '__hash__', '__init__', '__init_subclass__', '__le__', '__lt__', '__module__', '__ne__', '__new__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__setstate__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_bases', '_concrete', '_cursor', '_deeply_writable', '_file', '_format0', '_format_memberwise0', '_format_memberwise1', '_instance_version', '_is_memberwise', '_members', '_num_bytes', '_parent', '_serialize', '_to_writable', '_to_writable_postprocess', 'all_members', 'awkward_form', 'base', 'base_names_versions', 'bases', 'behaviors', 'check_numbytes', 'class_code', 'class_flags', 'class_rawstreamers', 'class_streamer', 'class_version', 'classname', 'close', 'closed', 'concrete', 
'cursor', 'empty', 'encoded_classname', 'file', 'has_member', 'hook_after_read_members', 'hook_before_postprocess', 'hook_before_read', 'hook_before_read_members', 'instance_version', 'is_instance', 'is_memberwise', 'member', 'member_names', 'members', 'num_bytes', 'parent', 'postprocess', 'read', 'read_member_n', 'read_members', 'read_numbytes_version', 'serialize', 'strided_interpretation', 'to_pyroot', 'to_writable', 'tojson', 'writable']
{'fGraphs': <TList of 2 items at 0x0173d15b9970>, 'fFunctions': None, 'fHistogram': None, 'fMaximum': -1111.0, 'fMinimum': -1111.0}

Then I tried to retrieve the TGraph axes and, interestingly enough, I got the expected output without errors:

with uproot.open("tmultigraph_example.root") as f:
    p = f["myMultiGraph"]

    print(p.member("fGraphs")[0].values())
    print()
    print(p.member("fGraphs")[1].values())

Output:

(array([0., 1., 2., 3., 4., 5., 6., 7., 8., 9.]), array([0., 1., 2., 3., 4., 5., 6., 7., 8., 9.]))

(array([0., 1., 2., 3., 4., 5., 6., 7., 8., 9.]), array([0., 1., 1.41421356, 1.73205081, 2., 2.23606798, 2.44948974, 2.64575131, 2.82842712, 3.]))

So it is possible to read a TMultiGraph from an existing .root file.

Then there is my question: what needs to be done to implement TMultiGraph writing? I would appreciate any guidance or tips on this topic, as well as any documentation regarding the .root data format and how serialization is handled.

For reference, I would like to be able to do the equivalent in Uproot:

import ROOT
from array import array

n_points = 10
x = [i for i in range(n_points)]
y1 = [i for i in range(n_points)]
y2 = [i**0.5 for i in range(n_points)]

graph1 = ROOT.TGraph(n_points, array('d', x), array('d', y1))
graph2 = ROOT.TGraph(n_points, array('d', x), array('d', y2))

mg = ROOT.TMultiGraph()
mg.Add(graph1)
mg.Add(graph2)

output_file = ROOT.TFile("tmultigraph_example.root", "RECREATE")
mg.Write("myMultiGraph")
output_file.Close()