/usr/share/pyshared/pymc/graph.py is in python-pymc 2.2+ds-1.
This file is owned by root:root, with mode 0o644.
The actual contents of the file can be viewed below.
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 | import pymc as pm
import os
from copy import copy
__all__ = ['graph', 'moral_graph']
try:
import pydot
pydot_imported = True
except:
pydot_imported = False
def moral_graph(model, format='raw', prog='dot', path=None, name=None):
"""
moral_graph(model,format='raw', prog='dot', path=None)
Draws the moral graph for this model and writes it to path with filename name.
Returns the pydot 'dot' object for further user manipulation.
GraphViz and PyDot must be installed to use this function.
:Parameters:
model : PyMC Model instance
format : string
'ps', 'ps2', 'hpgl', 'pcl', 'mif', 'pic', 'gd', 'gd2', 'gif', 'jpg',
'jpeg', 'png', 'wbmp', 'ismap', 'imap', 'cmap', 'cmapx', 'vrml', 'vtx', 'mp',
'fig', 'svg', 'svgz', 'dia', 'dot', 'canon', 'plain', 'plain-ext', 'xdot'
prog : string
'dot', 'neato'
path : string
If model.__name__ is defined and path is None, the output file is
./'name'.'format'.
:Note:
format='raw' outputs a GraphViz dot file.
"""
if not pydot_imported:
raise ImportError('PyDot must be installed to use the moral_graph function.\n PyDot is available from http://dkbza.org/pydot.html')
model.moral_dot_object = pydot.Dot()
# Data are filled ellipses
for datum in model.observed_stochastics:
model.moral_dot_object.add_node(pydot.Node(name=datum.__name__, style='filled'))
# Stochastics are open ellipses
for s in model.stochastics:
model.moral_dot_object.add_node(pydot.Node(name=s.__name__))
gone_already = set()
for s in model.stochastics | model.observed_stochastics:
gone_already.add(s)
for other_s in s.moral_neighbors:
if not other_s in gone_already:
model.moral_dot_object.add_edge(pydot.Edge(src=other_s.__name__, dst=s.__name__, arrowhead='none'))
# Draw the graph
ext=format
if format=='raw':
ext='dot'
if name is None:
name = model.__name__
name = name + '.' + ext
if not path is None:
model.moral_dot_object.write(path=os.path.join(path,name), format=format, prog=prog)
else:
model.moral_dot_object.write(path='./'+name, format=format, prog=prog)
return model.moral_dot_object
def graph(model, format='raw', prog='dot', path=None, name=None, consts=False, legend=False,
collapse_deterministics = False, collapse_potentials = False, label_edges=True):
"""
graph( model,
format='raw',
prog='dot',
path=None,
name=None,
consts=False,
legend=True,
collapse_deterministics = False,
collapse_potentials = False)
Draws the graph for this model and writes it to path with filename name.
Returns the pydot 'dot' object for further user manipulation.
GraphViz and PyDot must be installed to use this function.
:Parameters:
model : PyMC Model instance
format : string
'ps', 'ps2', 'hpgl', 'pcl', 'mif', 'pic', 'gd', 'gd2', 'gif', 'jpg',
'jpeg', 'png', 'wbmp', 'ismap', 'imap', 'cmap', 'cmapx', 'vrml', 'vtx', 'mp',
'fig', 'svg', 'svgz', 'dia', 'dot', 'canon', 'plain', 'plain-ext', 'xdot'
prog : string
'dot', 'neato'
path : string
If model.__name__ is defined and path is None, the output file is
./'name'.'format'.
consts : boolean
If True, constant parents are included in the graph.
legend : boolean
If True, a graph legend is created.
collapse_deterministics : boolean
If True, all deterministic dependencies are collapsed.
collapse_potentials : boolean
If True, all potentials are converted to undirected edges.
"""
if not pydot_imported:
raise ImportError('PyDot must be installed to use the graph function.\n PyDot is available from http://dkbza.org/pydot.html')
pydot_nodes = {}
pydot_subgraphs = {}
obj_substitute_names = {}
shown_objects = set([])
model.dot_object = pydot.Dot()
def get_obj_names(obj, key):
if isinstance(obj, pm.Stochastic):
if obj in obj_substitute_names:
return obj_substitute_names[obj]
if obj.observed:
datum = obj
# Data are filled ellipses
pydot_nodes[datum] = pydot.Node(name=datum.__name__, style='filled')
model.dot_object.add_node(pydot_nodes[datum])
shown_objects.add(datum)
obj_substitute_names[datum] = [datum.__name__]
else:
s = obj
# Stochastics are open ellipses
pydot_nodes[s] = pydot.Node(name=s.__name__)
model.dot_object.add_node(pydot_nodes[s])
shown_objects.add(s)
obj_substitute_names[s] = [s.__name__]
elif isinstance(obj, pm.Deterministic):
if obj in obj_substitute_names:
return obj_substitute_names[obj]
d = obj
# Deterministics are downward-pointing triangles
if not collapse_deterministics:
pydot_nodes[d] = pydot.Node(name=d.__name__, shape='invtriangle')
model.dot_object.add_node(pydot_nodes[d])
shown_objects.add(d)
obj_substitute_names[d] = [d.__name__]
else:
obj_substitute_names[d] = []
elif isinstance(obj, pm.Potential):
if obj in obj_substitute_names:
return obj_substitute_names[obj]
potential = obj
# Potentials are squares
if not collapse_potentials:
pydot_nodes[potential] = pydot.Node(name=potential.__name__, shape='box')
model.dot_object.add_node(pydot_nodes[potential])
shown_objects.add(potential)
obj_substitute_names[potential] = [potential.__name__]
else:
obj_substitute_names[potential]=[]
elif consts:
if key in obj_substitute_names:
return
else:
obj_substitute_names[key] = [key]
model.dot_object.add_node(pydot.Node(name=key, style='filled'))
return
else:
return
return obj_substitute_names[obj]
connected = []
def maybe_connect_parent(src, dst, label):
if (src,dst,label) in connected:
return False
else:
connected.append((src,dst,label))
model.dot_object.add_edge(pydot.Edge(src=src, dst=dst, label=label))
return True
def connect_parents(node):
if collapse_deterministics:
parent_tups = [(s.__name__, s) for s in node.extended_parents]
if consts:
parent_tups += filter(lambda x: not isinstance(x[1], pm.Variable), node.parents.items())
parent_dict = dict(parent_tups)
else:
parent_dict = node.parents
for key in parent_dict:
key_val = parent_dict[key]
label = label_edges*key or ''
if hasattr(key_val,'__name__'):
const_node_name = key_val.__name__
elif len(key_val.__str__()) <= 10:
const_node_name = key_val.__str__()
else:
const_node_name = key_val.__class__.__name__
if isinstance(key_val, pm.Variable):
if any([maybe_connect_parent(name, node.__name__, label) for name in get_obj_names(key_val, None)]):
connect_parents(key_val)
elif isinstance(key_val, pm.ContainerBase):
for var in key_val.variables:
if any([maybe_connect_parent(name, node.__name__, label) for name in get_obj_names(var, None)]):
connect_parents(var)
elif consts:
get_obj_names(key_val, const_node_name)
maybe_connect_parent(const_node_name, node.__name__, label)
# Create edges from parent-child relationships between nodes.
if collapse_potentials:
shownodes = model.variables
else:
shownodes = model.nodes
for node in shownodes:
get_obj_names(node,None)
if node in shown_objects:
connect_parents(node)
if collapse_potentials:
for potential in model.potentials:
if collapse_deterministics:
potential_parents = set()
for p in potential.extended_parents:
potential_parents.update(get_obj_names(p,None))
else:
potential_parents=set()
for parent in potential.parents.values():
if isinstance(parent, pm.Variable):
potential_parents |= set(get_obj_names(parent, None))
elif isinstance(parent, ContainerBase):
for ult_parent in parent.variables:
potential_parents |= set(get_obj_names(parent, None))
remaining_parents = copy(potential_parents)
for p1 in potential_parents:
remaining_parents.discard(p1)
for p2 in remaining_parents:
new_edge = pydot.Edge(src = p2, dst = p1, label=potential.__name__, arrowhead='none')
model.dot_object.add_edge(new_edge)
# Add legend if requested
if legend:
legend = pydot.Cluster(graph_name = 'Legend', label = 'Legend')
legend.add_node(pydot.Node(name='data', style='filled'))
legend.add_node(pydot.Node(name='stochastics'))
legend.add_node(pydot.Node(name='deterministics', shape='invtriangle'))
legend.add_node(pydot.Node(name='potentials', shape='box'))
if consts:
legend.add_node(pydot.Node(name='constants', style='filled'))
model.dot_object.add_subgraph(legend)
# Draw the graph
ext=format
if format=='raw':
ext='dot'
if name is None:
name = model.__name__
name = name + '.' + ext
if not path == None:
model.dot_object.write(path=os.path.join(path,name), format=format, prog=prog)
else:
model.dot_object.write(path='./' + name, format=format, prog=prog)
return model.dot_object
# Alias as dag
dag = graph
|