How to remove nodes by both name and condition for Networkx?

I have an edge list like below. I want to remove all nodes start with A or C, and degree <= 1.

(A2,A1)
(A1,B1)
(B1,C2)
(C2,B2)
(B2,C1)
(B1,A3)

E.g, for the list above, I want to get:

(B1,C2)
(C2,B2)

How can I do it?

You can do this to remove nodes that start with 'A' or 'C', and have a degree <= 1:

import networkx as nx

G = nx.DiGraph()

G.add_edges_from([
    ("A2", "A1"),
    ("A1", "B1"),
    ("B1", "C2"),
    ("C2", "B2"),
    ("B2", "C1"),
    ("B1", "A3"),
])


def node_filter(node):
    return (
        (node.startswith('A') or
        node.startswith('C')) and
        G.out_degree(node) <= 1
    )

G.remove_nodes_from(list(nx.subgraph_view(G, filter_node=node_filter).nodes))

However in your example, the node 'C2' will be deleted because it has an out-degree of 1 and begins with C. Therefore, the edge (C2, B2) and (B1, C2) will be deleted. You are only left with the nodes B1 and B2.

print(G.edges)
G.remove_nodes_from(list(nx.subgraph_view(G, filter_node=node_filter).nodes))
print(G.edges)
print(G.nodes)


[('A2', 'A1'), ('A1', 'B1'), ('B1', 'C2'), ('B1', 'A3'), ('C2', 'B2'), ('B2', 'C1')]
[]
['B1', 'B2']

Credit to @Tom McLean

import networkx as nx

G = nx.Graph()
G.add_edges_from([('A2', 'A1'), 
                  ('A1', 'B1'), 
                  ('B1', 'C2'), 
                  ('C2', 'B2'), 
                  ('B2', 'C1'), 
                  ('B1', 'A3')])

def node_filter(node):
      return(
          (node.startswith('A') or 
          node.startswith('C')) and 
          G.degree(node) <= 1)

for i in range(0,len(G)):
    G.remove_nodes_from(list(nx.subgraph_view(G, filter_node = node_filter).nodes))
    i = i+1

print(G.nodes)
print(G.edges)

Will result in

['B1', 'C2', 'B2']
[('B1', 'C2'), ('C2', 'B2')]