def test_metadata_double_name(self, mock_etl2, mock_open):
edges = triangleEdges.copy()
edges['a1'] = triangleNodes.a1.map(lambda x: x+10)
graphistry.bind(source='src', destination='dst', node='id').plot(edges, triangleNodes)
dataset = mock_etl2.call_args[0][0]
self.assertIn('a1', dataset['attributes']['nodes'])
self.assertIn('a1', dataset['attributes']['edges'])
def test_metadata_no_nat(self, mock_etl2, mock_open):
edges = triangleEdges.copy()
edges['testDate'] = triangleNodes.a1.map(lambda x: None if x == 1 else datetime.datetime.now())
edges['testDate2'] = triangleNodes.a1.map(lambda x: None)
edges['testDate'] = pandas.to_datetime(edges.testDate, errors='ignore')
edges['testDate2'] = pandas.to_datetime(edges.testDate2, errors='ignore')
graphistry.bind(source='src', destination='dst', node='id').plot(edges)
dataset = mock_etl2.call_args[0][0]
for attrib in ['testDate', 'testDate2']:
for entry in list(dataset['attributes']['edges'][attrib]['aggregations'].values()):
self.assertFalse(isinstance(entry, pandas.tslib.NaTType))
def test_metadata_mandatory_fields(self, mock_etl2, mock_open):
graphistry.bind(source='src', destination='dst').plot(triangleEdges)
dataset = mock_etl2.call_args[0][0]
self.assertListEqual(list(dataset['attributes']['nodes']), [nid])
self.assertListEqual(list(dataset['attributes']['edges']), [])
self.assertEqual(dataset['encodings']['nodes'], {
'pointTitle': {'attributes': [nid]},
'nodeId': {'attributes': [nid]}
})
self.assertEqual(dataset['encodings']['edges'], {
'source': {'attributes': ['src']},
'destination': {'attributes': ['dst']}
})
def test_empty_graph(self, mock_error, mock_etl2, mock_warn, mock_open):
mock_error.side_effect = ValueError('error')
plotter = graphistry.bind(source='src', destination='dst')
with self.assertRaises(ValueError):
plotter.plot(pandas.DataFrame([]))
self.assertFalse(mock_etl2.called)
self.assertTrue(mock_error.called)
def test_no_src_dst(self, mock_etl2, mock_warn, mock_open):
with self.assertRaises(ValueError):
graphistry.bind().plot(triangleEdges)
with self.assertRaises(ValueError):
graphistry.bind(source='src').plot(triangleEdges)
with self.assertRaises(ValueError):
graphistry.bind(destination='dst').plot(triangleEdges)
with self.assertRaises(ValueError):
graphistry.bind(source='doesnotexist', destination='dst').plot(triangleEdges)
def test_metadata_no_nan(self, mock_etl2, mock_open):
edges = triangleEdges.copy()
edges['testNone'] = triangleNodes.a1.map(lambda x: numpy.nan)
edges['testNone'] = pandas.to_numeric(edges.testNone, errors='ignore')
edges['testInt'] = triangleNodes.a1.map(lambda x: numpy.nan if x%2 == 1 else 0)
edges['testFloat'] = triangleNodes.a1.map(lambda x: numpy.nan if x%2 == 1 else 0.5)
edges['testString'] = triangleNodes.a1.map(lambda x: numpy.nan if x%2 == 1 else 'foo')
edges['testBool'] = triangleNodes.a1.map(lambda x: numpy.nan if x%2 == 1 else True)
graphistry.bind(source='src', destination='dst', node='id').plot(edges)
dataset = mock_etl2.call_args[0][0]
for attrib in ['testInt', 'testFloat', 'testString', 'testBool', 'testNone']:
for entry in list(dataset['attributes']['edges'][attrib]['aggregations'].values()):
if entry is None or isinstance(entry, str):
pass
else:
self.assertFalse(numpy.isnan(entry))
def draw_graph(graph, nodes=None):
'''
采用pygraphistry 来绘制网络图，节点颜色目前还不能超过12种
:param graph: networkx.Graph/DiGraph
:param nodes: DataFrame,如果需要按社区颜色绘制，请传入带有社区信息的节点表, ['Id','modulraity_class']
:return: None
'''
graphistry.register(key='contact pygraphistry for api key')
ploter = graphistry.bind(source='Source',
destination='Target').graph(graph)
if nodes is not None:
ploter = ploter.bind(node='Id',
point_color='modularity_class').nodes(nodes)
ploter.plot()
return None
def test_api3_cudf_to_arrow_memoization_forgets(self):
maybe_cudf = None
try:
import cudf
maybe_cudf = cudf
except ImportError:
1
if maybe_cudf is None:
return
plotter = graphistry.bind()
df = maybe_cudf.DataFrame({'x': [0]})
arr1 = plotter._table_to_arrow(df)
for i in range(1, 110):
plotter._table_to_arrow(maybe_cudf.DataFrame({'x': [i]}))
assert not (arr1 is plotter._table_to_arrow(df))
def test_networkx2igraph(self):
import networkx as nx
ng = nx.complete_graph(3)
[x, y] = [int(x) for x in nx.__version__.split('.')]
if x == 1:
nx.set_node_attributes(ng, 'vattrib', 0)
nx.set_edge_attributes(ng, 'eattrib', 1)
else:
nx.set_node_attributes(ng, 0, 'vattrib')
nx.set_edge_attributes(ng, 1, 'eattrib')
(e, n) = graphistry.bind(source='src',
destination='dst').networkx2pandas(ng)
edges = pd.DataFrame({
'dst': {
0: 1,
1: 2,
2: 2
},
'src': {
0: 0,
1: 0,
2: 1
},
'eattrib': {
0: 1,
1: 1,
2: 1
}
})
nodes = pd.DataFrame({
'__nodeid__': {
0: 0,
1: 1,
2: 2
},
'vattrib': {
0: 0,
1: 0,
2: 0
}
})
assertFrameEqual(e, edges)
assertFrameEqual(n, nodes)
def test_igraph2pandas(self):
ig = igraph.Graph.Tree(4, 2)
ig.vs['vattrib'] = 0
ig.es['eattrib'] = 1
(e, n) = graphistry.bind(source='src', destination='dst').igraph2pandas(ig)
edges = pandas.DataFrame({
'dst': {0: 1, 1: 2, 2: 3},
'src': {0: 0, 1: 0, 2: 1},
'eattrib': {0: 1, 1: 1, 2: 1}
})
nodes = pandas.DataFrame({
'__nodeid__': {0: 0, 1: 1, 2: 2, 3: 3},
'vattrib': {0: 0, 1: 0, 2: 0, 3: 0}
})
assertFrameEqual(e, edges)
assertFrameEqual(n, nodes)
def test_networkx2igraph(self):
ng = networkx.complete_graph(3)
networkx.set_node_attributes(ng, 'vattrib', 0)
networkx.set_edge_attributes(ng, 'eattrib', 1)
(e, n) = graphistry.bind(source='src', destination='dst').networkx2pandas(ng)
edges = pandas.DataFrame({
'dst': {0: 1, 1: 2, 2: 2},
'src': {0: 0, 1: 0, 2: 1},
'eattrib': {0: 1, 1: 1, 2: 1}
})
nodes = pandas.DataFrame({
'__nodeid__': {0: 0, 1: 1, 2: 2},
'vattrib': {0: 0, 1: 0, 2: 0}
})
assertFrameEqual(e, edges)
assertFrameEqual(n, nodes)
def test_igraph2pandas(self):
ig = igraph.Graph.Tree(4, 2)
ig.vs['vattrib'] = 0
ig.es['eattrib'] = 1
(e, n) = graphistry.bind(source='src', destination='dst').igraph2pandas(ig)
edges = pandas.DataFrame({
'dst': {0: 1, 1: 2, 2: 3},
'src': {0: 0, 1: 0, 2: 1},
'eattrib': {0: 1, 1: 1, 2: 1}
})
nodes = pandas.DataFrame({
'__nodeid__': {0: 0, 1: 1, 2: 2, 3: 3},
'vattrib': {0: 0, 1: 0, 2: 0, 3: 0}
})
assertFrameEqual(e, edges)
assertFrameEqual(n, nodes)
def test_api3_cudf_to_arrow_memoization(self):
maybe_cudf = None
try:
import cudf
maybe_cudf = cudf
except ImportError:
1
if maybe_cudf is None:
return
plotter = graphistry.bind()
df = maybe_cudf.DataFrame({'x': [1]})
arr1 = plotter._table_to_arrow(df)
arr2 = plotter._table_to_arrow(df)
assert isinstance(arr1, pa.Table)
assert arr1 is arr2
arr3 = plotter._table_to_arrow(maybe_cudf.DataFrame({'x': [1]}))
assert arr1 is arr3
def modularity(cluster_result,
edgedata=None,
graph=None,
directed=True,
edge_weight='Weight'):
'''
:param cluster_result: 聚类结果，参考gephi输出的表，[Id,modulraity_class]
:param edgedata: 边数据，与graph给定其中一个
:param graph: networkx中的Graph/DiGraph
:param directed: 是否为有向图
:param edge_weight:
None/str, 计算模块度是否使用边的权重，如果使用，给定边权重的name
例如edge_weight='Weight'
如果不使用，请给定为None
:return: Q值
ps：
1.edgedata 和 graph 至少要给定一个
2.与gephi中计算的模块度结果已经对比过了，结果一致
'''
if edgedata is None and graph is not None:
edgedata = NetworkUnity.networkx2pandas(graph)
gr = graphistry.bind(source='Source',
destination='Target',
node='Id',
edge_weight=edge_weight)
ig = gr.pandas2igraph(edgedata, directed=directed)
nodes = pd.DataFrame(list(ig.vs['Id']), columns=['Id'])
community_data = pd.merge(nodes,
cluster_result,
left_on='Id',
right_on='Id',
how='left')
if edge_weight is None:
Q = ig.modularity(list(community_data['modularity_class']),
weights=None)
else:
Q = ig.modularity(list(community_data['modularity_class']),
weights=list(ig.es[edge_weight]))
return Q
def test_au_e_enc_full(self):
g = graphistry.bind(source='s',
destination='d',
edge_color='c',
edge_title='t',
edge_label='l',
edge_weight='w',
edge_opacity='o',
edge_icon='i',
edge_size='s',
edge_source_color='sc',
edge_destination_color='dc')
g = g.encode_edge_color('c', ["green"], as_categorical=True)
au = ArrowUploader()
assert au.g_to_edge_encodings(g) == {
'bindings': {
'source': 's',
'destination': 'd',
'edge_color': 'c',
'edge_title': 't',
'edge_label': 'l',
'edge_weight': 'w',
'edge_opacity': 'o',
'edge_icon': 'i',
'edge_size': 's',
'edge_source_color': 'sc',
'edge_destination_color': 'dc'
},
'complex': {
'default': {
'edgeColorEncoding': {
'graphType': 'edge',
'encodingType': 'color',
'attribute': 'c',
'variation': 'categorical',
'colors': ['green']
}
}
}
}
def test_networkx2igraph(self):
ng = networkx.complete_graph(3)
networkx.set_node_attributes(ng, 'vattrib', 0)
networkx.set_edge_attributes(ng, 'eattrib', 1)
(e, n) = graphistry.bind(source='src',
destination='dst').networkx2pandas(ng)
edges = pandas.DataFrame({
'dst': {
0: 1,
1: 2,
2: 2
},
'src': {
0: 0,
1: 0,
2: 1
},
'eattrib': {
0: 1,
1: 1,
2: 1
}
})
nodes = pandas.DataFrame({
'__nodeid__': {
0: 0,
1: 1,
2: 2
},
'vattrib': {
0: 0,
1: 0,
2: 0
}
})
assertFrameEqual(e, edges)
assertFrameEqual(n, nodes)
def test_au_n_enc_full(self):
g = graphistry.bind(node='n',
point_color='c',
point_size='s',
point_title='t',
point_label='l',
point_weight='w',
point_opacity='o',
point_icon='i',
point_x='x',
point_y='y')
g = g.encode_point_color('c', ["green"], as_categorical=True)
au = ArrowUploader()
assert au.g_to_node_encodings(g) == {
'bindings': {
'node': 'n',
'node_color': 'c',
'node_size': 's',
'node_title': 't',
'node_label': 'l',
'node_weight': 'w',
'node_opacity': 'o',
'node_icon': 'i',
'node_x': 'x',
'node_y': 'y',
},
'complex': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'c',
'variation': 'categorical',
'colors': ['green']
}
}
}
}
def test_ipython(self, mock_util, mock_post, mock_open):
widget = graphistry.bind(source='src',
destination='dst').plot(triangleEdges)
self.assertIsInstance(widget, IPython.core.display.HTML)
def test_point_color(self):
assert graphistry.bind().encode_point_color('z')._point_color == 'z'
assert graphistry.bind().encode_point_color('z', ["red", "blue"], as_continuous=True)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'continuous',
'colors': ['red', 'blue']
}
},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', ["red", "blue"], as_categorical=True)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'colors': ['red', 'blue']
}
},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'truck': 'red'})._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': { 'fixed': { 'truck': 'red' } } }
}
},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'truck': 'red'}, default_mapping='blue')._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': { 'truck': 'red' },
'other': 'blue'
}
}
}
},
'current': {}
}
}
def test_bind_edges_nodes(self, mock_etl2, mock_open):
plotter0 = graphistry.bind(source='src').bind(destination='dst')
plotter1 = plotter0.bind(node='id').bind(point_title='a2')
plotter1.edges(triangleEdges).nodes(triangleNodes).plot()
self.assertTrue(mock_etl2.called)
def test_metadata_mandatory_fields(self, mock_etl2, mock_open):
edges = triangleEdges.copy()
edges['ustring'] = [u'abcdef', u'тйîbàüd', u'♜ ♞ ♝ ♛ ♚ ♝ ♞ ♜']
graphistry.bind(source='src', destination='dst').plot(edges)
self.assertTrue(mock_etl2.called)
def test_set_mode(self):
assert graphistry.bind().encode_point_color('z', categorical_mapping={'a': 'b'})._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': {'fixed': { 'a': 'b' } } }
}
},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'a': 'b'}, for_default=False, for_current=False)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'a': 'b'}, for_default=True, for_current=False)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': {'fixed': { 'a': 'b' } } }
}
},
'current': {}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'a': 'b'}, for_default=False, for_current=True)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': { },
'current': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': { 'fixed': { 'a': 'b' } } }
}
}
}
}
assert graphistry.bind().encode_point_color('z', categorical_mapping={'a': 'b'}, for_default=True, for_current=True)._complex_encodings \
== {
**TestPlotterEncodings.COMPLEX_EMPTY,
'node_encodings': {
'default': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': { 'fixed': { 'a': 'b' } } }
}
},
'current': {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': { 'categorical': { 'fixed': { 'a': 'b' } } }
}
}
}
}
def test_no_nodeid(self, mock_etl2, mock_warn, mock_open):
plotter = graphistry.bind(source='src', destination='dst')
with self.assertRaises(ValueError):
plotter.plot(triangleEdges, triangleNodes)
def test_bind_description(self):
plotter = graphistry.bind().description('d')
assert plotter._description == 'd'
def test_edge_icon(self):
assert graphistry.bind().encode_edge_icon('z')._edge_icon == 'z'
def test_au_n_enc_mt(self):
g = graphistry.bind()
au = ArrowUploader()
assert au.g_to_node_encodings(g) == {'bindings': {}}
def test_ipython(self, mock_util, mock_post, mock_open):
widget = graphistry.bind(source='src', destination='dst').plot(triangleEdges)
self.assertIsInstance(widget, IPython.core.display.HTML)
def test_pandas2igraph(self):
plotter = graphistry.bind(source='src', destination='dst', node='id')
ig = plotter.pandas2igraph(triangleEdges)
(e, n) = plotter.igraph2pandas(ig)
assertFrameEqual(e, triangleEdges[['src', 'dst']])
assertFrameEqual(n, triangleNodes[['id']])
def test_bind_edges_nodes(self, mock_etl2, mock_open):
plotter0 = graphistry.bind(source='src').bind(destination='dst')
plotter1 = plotter0.bind(node='id').bind(point_title='a2')
plotter1.edges(triangleEdges).nodes(triangleNodes).plot()
self.assertTrue(mock_etl2.called)
def test_bind_chain(self, mock_etl2, mock_open):
plotter0 = graphistry.bind(source='caca').bind(destination='dst',
source='src')
plotter0.plot(triangleEdges)
self.assertTrue(mock_etl2.called)
def test_point_size(self):
assert graphistry.bind().encode_point_size('z')._point_size == 'z'
def test_table_to_pandas_from_none(self):
plotter = graphistry.bind()
assert plotter._table_to_pandas(None) is None
def test_point_icon(self):
assert graphistry.bind().encode_point_icon('z')._point_icon == 'z'
def test_table_to_pandas_from_pandas(self):
plotter = graphistry.bind()
df = pd.DataFrame({'x': []})
assert isinstance(plotter._table_to_pandas(df), pd.DataFrame)
def test_edge_color(self):
assert graphistry.bind().encode_edge_color('z')._edge_color == 'z'
def test_table_to_arrow_from_none(self):
plotter = graphistry.bind()
assert plotter._table_to_arrow(None) is None
def test_composition(self):
# chaining + overriding
out = graphistry.bind()\
.encode_point_size('z', categorical_mapping={'m': 2})\
.encode_point_color('z', categorical_mapping={'a': 'b'}, for_current=True)\
.encode_point_color('z', categorical_mapping={'a': 'b2'})\
.encode_edge_color( 'z', categorical_mapping={'x': 'y'}, for_current=True)\
._complex_encodings
assert out['edge_encodings']['default'] == {
'edgeColorEncoding': {
'graphType': 'edge',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': {
'x': 'y'
}
}
}
}
}
assert out['edge_encodings']['current'] == {
'edgeColorEncoding': {
'graphType': 'edge',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': {
'x': 'y'
}
}
}
}
}
assert out['node_encodings']['default'] == {
'pointSizeEncoding': {
'graphType': 'point',
'encodingType': 'size',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': {
'm': 2
}
}
}
},
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': {
'a': 'b2'
}
}
}
}
}
assert out['node_encodings']['current'] == {
'pointColorEncoding': {
'graphType': 'point',
'encodingType': 'color',
'attribute': 'z',
'variation': 'categorical',
'mapping': {
'categorical': {
'fixed': {
'a': 'b'
}
}
}
}
}
def test_table_to_arrow_from_arrow(self):
plotter = graphistry.bind()
df = pd.DataFrame({'x': []})
arr = pa.Table.from_pandas(df)
assert isinstance(plotter._table_to_arrow(arr), pa.Table)
def test_pandas2igraph(self):
plotter = graphistry.bind(source='src', destination='dst', node='id')
ig = plotter.pandas2igraph(triangleEdges)
(e, n) = plotter.igraph2pandas(ig)
assertFrameEqual(e, triangleEdges[['src', 'dst']])
assertFrameEqual(n, triangleNodes[['id']])
def test_unknown_col_nodes(self, mock_etl2, mock_warn, mock_open):
plotter = graphistry.bind(source='src', destination='dst', node='id', point_title='doesnotexist')
plotter.plot(triangleEdges, triangleNodes)
self.assertTrue(mock_etl2.called)
self.assertTrue(mock_warn.called)
def test_bind_chain(self, mock_etl2, mock_open):
plotter0 = graphistry.bind(source='caca').bind(destination='dst', source='src')
plotter0.plot(triangleEdges)
self.assertTrue(mock_etl2.called)
def test_init_mt(self):
assert graphistry.bind(
)._complex_encodings == TestPlotterEncodings.COMPLEX_EMPTY
def test_init(self):
g = graphistry.bind()
assert g._style == None
def test_bind_edges(self, mock_etl2, mock_warn, mock_open):
plotter = graphistry.bind(source='src', destination='dst', edge_title='src')
plotter.plot(triangleEdges)
self.assertTrue(mock_etl2.called)
self.assertFalse(mock_warn.called)
def test_bind_name(self):
plotter = graphistry.bind().name('n')
assert plotter._name == 'n'
def test_no_ipython(self, mock_post, mock_open):
url = graphistry.bind(source='src', destination='dst').plot(triangleEdges)
self.assertIn('fakedatasetname', url)
self.assertIn('faketoken', url)
self.assertTrue(mock_open.called)
self.assertTrue(mock_post.called)
