isinstance check too strict

[bjodah]

Hi, thank you for making this project available--it looks really interesting.
I'm just getting started and immediately ran into this (when using from_nx):

> /usr/local/lib/python3.6/dist-packages/pyvis/network.py(197)add_node()
    195         :type y: num (optional)
    196         """
--> 197         assert isinstance(n_id, str) or isinstance(n_id, int)
    198         if label:
    199             node_label = label

ipdb> p n_id
2
ipdb> p type(n_id)
<class 'numpy.int64'>

one possible fix would be to check against the abstract base class:

ipdb> from numbers import Integral
ipdb> isinstance(n_id, Integral)
True

[boludo00]

Hi @bjodah ,
would you mind sharing the code you wrote to get this error? I can't seem to reproduce it using numpy and the from_nx method.

This is my output (Python3.6)

from pyvis.network import Network
import networkx as nx
import numpy as np
nxg = nx.Graph()
nxg.add_nodes_from(np.arange(3))
g=Network()
g.from_nx(nxg)
g
<class 'pyvis.network.Network'> |N|=3 |E|=0
g.get_nodes()
[0, 1, 2]
type(g.get_nodes()[0])
numpy.int64

[bjodah]

Sure, I'm using pyvis 0.1.4.1 (what pip tells me, saw no pyvis.__version__).
Here's a reproducer:

import sys
import networkx as nx
from networkx.drawing.nx_agraph import graphviz_layout
import numpy as np
import pyvis
from pyvis import network as net

print('\n'.join(['%s: %s' % (m.__name__, m.__version__) for m in (nx, np)]))
print('python: {}'.format(sys.version_info))
def mk_graph(from_to):
    dg = nx.DiGraph()
    dg.add_nodes_from([0, 1, 2])
    for src, dst_lst in from_to.items():
        for dst in dst_lst:
            dg.add_edge(src, dst)
    return dg


def viz(gr):
    pos = graphviz_layout(gr)
    g = net.Network(notebook=True)
    g.from_nx(gr)
    for i, node in enumerate(g.nodes):
        node['x'] = pos[i][0]
        node['y'] = pos[i][1]
        node['physics'] = False
    for edge in g.edges:
        edge['physics'] = False
    return g

dg = mk_graph({0: np.array([1, 2], dtype=int)})

viz(dg)

which gives me the following output in jupyter notebook:

networkx: 2.1
numpy: 1.15.0
python: sys.version_info(major=3, minor=6, micro=5, releaselevel='final', serial=0)
---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-18-60d494b73f4a> in <module>()
     31 dg = mk_graph({0: np.array([1, 2], dtype=int)})
     32 
---> 33 viz(dg)

<ipython-input-18-60d494b73f4a> in viz(gr)
     20     pos = graphviz_layout(gr)
     21     g = net.Network(notebook=True)
---> 22     g.from_nx(gr)
     23     for i, node in enumerate(g.nodes):
     24         node['x'] = pos[i][0]

/usr/local/lib/python3.6/dist-packages/pyvis/network.py in from_nx(self, nx_graph)
    554             for e in edges:
    555                 self.add_node(e[0], e[0], title=str(e[0]))
--> 556                 self.add_node(e[1], e[1], title=str(e[1]))
    557                 self.add_edge(e[0], e[1])
    558         else:

/usr/local/lib/python3.6/dist-packages/pyvis/network.py in add_node(self, n_id, label, shape, **options)
    195         :type y: num (optional)
    196         """
--> 197         assert isinstance(n_id, str) or isinstance(n_id, int)
    198         if label:
    199             node_label = label

AssertionError: 

[bjodah]

I can fix the code by simply doing:

dg.add_edge(src, int(dst))

instead, so no biggie.

[boludo00]

So strange, that code didnt throw an error for me
Oh well, glad you got it working. I'll play around more with the types that that JS api accepts.

[diitaz93]

Hi,
I just starting using pyvis and I still get this error. I am building a networkx graph from an adjacency matrix, so I can't use Bjorn's workaround. I am working in a jupyter notebook, with python 3.8.10, networkx 2.5.1 and pyvis 0.1.9. My code is as follows:

import networkx as nx
import scipy.sparse as sp
from pyvis.network import Network

M = sp.random(5,5,density=0.6)
G = nx.from_scipy_sparse_matrix(M)
nt = Network('500px', '500px')
nt.from_nx(G)
nt.show("nx.html")

which throws the same error as above:

---------------------------------------------------------------------------
AssertionError                            Traceback (most recent call last)
<ipython-input-10-fa438620fb54> in <module>
      6 G = nx.from_scipy_sparse_matrix(M)
      7 nt = Network('500px', '500px')
----> 8 nt.from_nx(G)
      9 nt.show("nx.html")

~/miniconda3/envs/netx/lib/python3.8/site-packages/pyvis/network.py in from_nx(self, nx_graph, node_size_transf, edge_weight_transf, default_node_size, default_edge_weight)
    611                 nodes[e[1]]['size'] = int(node_size_transf(nodes[e[1]]['size']))
    612                 self.add_node(e[0], **nodes[e[0]])
--> 613                 self.add_node(e[1], **nodes[e[1]])
    614 
    615                 if 'weight' not in e[2].keys():

~/miniconda3/envs/netx/lib/python3.8/site-packages/pyvis/network.py in add_node(self, n_id, label, shape, **options)
    205         :type y: num (optional)
    206         """
--> 207         assert isinstance(n_id, str) or isinstance(n_id, int)
    208         if label:
    209             node_label = label

AssertionError: 

[Yquetzal]

Hi, great project.
Same error here for a similar case.

[Economax]

Same here, trying to build a network graph from a scipy sparse matrix, converting to networkx and then pyvis.
Originating graph is from scikit-learn's nearest neighbor algorithm.

[SPA20]

Any answer to the above fault ?
I see the same error:

from pyvis.network import Network

nodes =[{"color": "red", "id": "2", "label": "N2", "shape": "o", "size": 50},
{"color": "red", "id": "3", "label": "N3", "shape": "o", "size": 50},
{"color": "red", "id": "1", "label": "N1", "shape": "o", "size": 50},
{"color": "red", "id": "4", "label": "N4", "shape": "o", "size": 50}]

edges=[{"arrows": "to", "font_size": 8, "from": "1", "label": "L12 : 500", "to": "2", "weight": 500},
{"arrows": "to", "font_size": 8, "from": "2", "label": "L231 : 1", "to": "3", "weight": 1},
{"arrows": "to", "font_size": 8, "from": "2", "label": "L232 : 1", "to": "3", "weight": 1},
{"arrows": "to", "font_size": 8, "from": "3", "label": "L34 : 500", "to": "4", "weight": 500},
{"arrows": "to", "font_size": 8, "from": "4", "label": "L41 : 500", "to": "1", "weight": 500}]

G = Network(width="800px", height='600px', heading='Test connectivity', layout=True, bgcolor='#aaaaaa', directed=True)
G.add_nodes(nodes)
G.add_edges(edges)
G.show('pyg.html')

Error:
C:\Apps\python.exe C:/.../ScratchMG2.py
Traceback (most recent call last):
File "C:\Apps\lib\site-packages\pyvis\network.py", line 261, in add_nodes
node = int(node)
TypeError: int() argument must be a string, a bytes-like object or a number, not 'dict'

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
File "C:/.../ScratchMG2.py", line 17, in
G.add_nodes(nodes)
File "C:\Apps\lib\site-packages\pyvis\network.py", line 265, in add_nodes
assert isinstance(node, str)
AssertionError

Process finished with exit code 1

why is the code trying to check whether the data "node" is a string... As per documentation, the request "add_nodes" should parse a LIST data structure.

[dianamindroc]

Having the same error and could not find a fix yet.

[Rocks-n-Code]

I resolved my similar error by dropping the indexes from my DataFrame with null values in the src and dst columns.

[ignaciogavier]

After some investigation on the source code pyvis.readthedocs.io/en/latest/_modules/pyvis/network.html I found that the issue is that the only accepted format to add a node is either string and integer (check the add_node method):

assert isinstance(n_id, str) or isinstance(n_id, int)

A way to solve this issue is converting all nodes to string before transferring from networkx.Graph to pyvis.Network.

[matriculus]

After some investigation on the source code pyvis.readthedocs.io/en/latest/_modules/pyvis/network.html I found that the issue is that the only accepted format to add a node is either string and integer (check the add_node method):

assert isinstance(n_id, str) or isinstance(n_id, int)

A way to solve this issue is converting all nodes to string before transferring from networkx.Graph to pyvis.Network.

It is a bad design to force network nodes to be only int or str. Networkx takes any object and pyvis is based on that and it doesn't. Seems artificial bottleneck.