You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 1 | #!/usr/bin/env python |
Jeremy Songster | ae01bba | 2016-07-11 15:39:17 -0700 | [diff] [blame] | 2 | """ |
| 3 | Created 2016 by ON.Lab |
| 4 | |
| 5 | Please refer questions to either the onos test mailing list at <onos-test@onosproject.org>, |
| 6 | the System Testing Plans and Results wiki page at <https://wiki.onosproject.org/x/voMg>, |
| 7 | or the System Testing Guide page at <https://wiki.onosproject.org/x/WYQg> |
| 8 | """ |
| 9 | |
| 10 | |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 11 | import time |
| 12 | import random |
| 13 | |
| 14 | class Graph: |
| 15 | """ |
| 16 | Graph class provides implementations of graph algorithms. |
| 17 | The functions currently supported include: |
| 18 | - Comparing two graphs with specified attributes for vertices and edges |
| 19 | - Getting DFI (Depth First Index) and back edges during a DFS |
| 20 | - Chain decomposition of a graph |
| 21 | - Finding (non-)cut-edges and vertices |
| 22 | """ |
| 23 | |
| 24 | def __init__( self ): |
| 25 | # We use a dictionary to store all information about the graph |
| 26 | self.graphDict = {} |
| 27 | # Depth-first index of each vertex |
| 28 | self.DFI = {} |
| 29 | self.currentDFI = 0 |
| 30 | # Parent vertex (and edge to that vertex) of each vertex in depth-first search tree |
| 31 | self.parentVertexInDFS = {} |
| 32 | self.parentEdgeInDFS = {} |
| 33 | # Back edges of the graph generated during DFS |
| 34 | self.backEdges = {} |
| 35 | # All chains in chain decomposition algorithm |
| 36 | self.chains = [] |
| 37 | |
| 38 | def update( self, graphDict ): |
| 39 | """ |
| 40 | Update the graph data. The current graph dictionary will be replaced by the |
| 41 | new one. |
| 42 | graphDict is in a dictionary which maps each vertex to a list of attributes. |
| 43 | An example of graphDict: |
| 44 | { vertex1: { 'edges': ..., 'name': ..., 'protocol': ... }, |
| 45 | vertex2: { 'edges': ..., 'name': ..., 'protocol': ... } } |
| 46 | Each vertex should at least have an 'edges' attribute which describes the |
| 47 | adjacency information. The value of 'edges' attribute is also represented by |
| 48 | a dictionary, which maps each edge (identified by the neighbor vertex) to a |
| 49 | list of attributes. |
| 50 | An example of the edges dictionary: |
| 51 | 'edges': { vertex2: { 'port': ..., 'type': ... }, |
| 52 | vertex3: { 'port': ..., 'type': ... } } |
| 53 | """ |
| 54 | self.graphDict = graphDict |
| 55 | return main.TRUE |
| 56 | |
| 57 | def compareGraphs( self, graphDictA, graphDictB, vertexAttributes=['edges'], edgeAttributes=['port'] ): |
| 58 | """ |
| 59 | Compare two graphs. |
| 60 | By default only the adjacency relationship, i.e. 'port' attribute in |
| 61 | 'edges' attribute for each vertex, is compared, To get other attributes |
| 62 | included, attribute name needs to be specified in the args, e.g. |
| 63 | vertexAttributes=[ 'edges', 'protocol' ] or |
| 64 | edgeAttributes=[ 'port', 'type' ] |
| 65 | Return main.TRUE if two graphs are equal, otherwise main.FALSE |
| 66 | """ |
| 67 | try: |
| 68 | result = main.TRUE |
| 69 | for vertex in set( graphDictA ).difference( graphDictB ): |
| 70 | result = main.FALSE |
| 71 | main.log.warn( "Graph: graph B: vertex {} not found".format( vertex ) ) |
| 72 | for vertex in set( graphDictB ).difference( graphDictA ): |
| 73 | result = main.FALSE |
| 74 | main.log.warn( "Graph: graph A: vertex {} not found".format( vertex ) ) |
| 75 | for vertex in set( graphDictA ).intersection( graphDictB ): |
| 76 | for vertexAttribute in vertexAttributes: |
| 77 | attributeFound = True |
| 78 | if vertexAttribute not in graphDictA[ vertex ]: |
| 79 | main.log.warn( "Graph: graph A -> vertex {}: attribute {} not found".format( vertex, vertexAttribute ) ) |
| 80 | attributeFound = False |
| 81 | if vertexAttribute not in graphDictB[ vertex ]: |
| 82 | attributeFound = False |
| 83 | main.log.warn( "Graph: graph B -> vertex {}: attribute {} not found".format( vertex, vertexAttribute ) ) |
| 84 | if not attributeFound: |
| 85 | result = main.FALSE |
| 86 | continue |
| 87 | else: |
| 88 | # Compare two attributes |
| 89 | attributeValueA = graphDictA[ vertex ][ vertexAttribute ] |
| 90 | attributeValueB = graphDictB[ vertex ][ vertexAttribute ] |
| 91 | # FIXME: the comparison may not work for (sub)attribute values that are of list type |
| 92 | # For attributes except for 'edges', we just rely on '==' for comparison |
| 93 | if not vertexAttribute == 'edges': |
| 94 | if not attributeValueA == attributeValueB: |
| 95 | result = main.FALSE |
| 96 | main.log.warn( "Graph: vertex {}: {} does not match: {} and {}".format( vertex, |
| 97 | vertexAttribute, |
| 98 | attributeValueA, |
| 99 | attributeValueB ) ) |
| 100 | # The structure of 'edges' is similar to that of graphs, so we use the same method for comparison |
| 101 | else: |
| 102 | edgeDictA = attributeValueA |
| 103 | edgeDictB = attributeValueB |
| 104 | for neighbor in set( edgeDictA ).difference( edgeDictB ): |
| 105 | result = main.FALSE |
| 106 | main.log.warn( "Graph: graph B -> vertex {}: neighbor {} not found".format( vertex, neighbor ) ) |
| 107 | for neighbor in set( edgeDictB ).difference( edgeDictA ): |
| 108 | result = main.FALSE |
| 109 | main.log.warn( "Graph: graph A -> vertex {}: neighbor {} not found".format( vertex, neighbor ) ) |
| 110 | for neighbor in set( edgeDictA ).intersection( edgeDictB ): |
| 111 | for edgeAttribute in edgeAttributes: |
| 112 | attributeFound = True |
| 113 | if edgeAttribute not in edgeDictA[ neighbor ]: |
| 114 | attributeFound = False |
| 115 | main.log.warn( "Graph: graph A -> vertex {} -> neighbor {}: attribute {} not found".format( vertex, |
| 116 | neighbor, |
| 117 | edgeAttribute ) ) |
| 118 | if edgeAttribute not in edgeDictB[ neighbor ]: |
| 119 | attributeFound = False |
| 120 | main.log.warn( "Graph: graph B -> vertex {} -> neighbor {}: attribute {} not found".format( vertex, |
| 121 | neighbor, |
| 122 | edgeAttribute ) ) |
| 123 | if not attributeFound: |
| 124 | result = main.FALSE |
| 125 | continue |
| 126 | else: |
| 127 | # Compare two attributes |
| 128 | attributeValueA = edgeDictA[ neighbor ][ edgeAttribute ] |
| 129 | attributeValueB = edgeDictB[ neighbor ][ edgeAttribute ] |
| 130 | if not attributeValueA == attributeValueB: |
| 131 | result = main.FALSE |
| 132 | main.log.warn( "Graph: vertex {} -> neighbor {}: {} does not match: {} and {}".format( vertex, |
| 133 | neighbor, |
| 134 | edgeAttribute, |
| 135 | attributeValueA, |
| 136 | attributeValueB ) ) |
| 137 | if not result: |
You Wang | 9a589b9 | 2016-07-14 09:43:21 -0700 | [diff] [blame] | 138 | #main.log.debug( "Graph: graphDictA: {}".format( graphDictA ) ) |
| 139 | #main.log.debug( "Graph: graphDictB: {}".format( graphDictB ) ) |
| 140 | pass |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 141 | return result |
| 142 | except TypeError: |
| 143 | main.log.exception( "Graph: TypeError exception found" ) |
| 144 | return main.ERROR |
| 145 | except KeyError: |
| 146 | main.log.exception( "Graph: KeyError exception found" ) |
| 147 | return main.ERROR |
| 148 | except Exception: |
| 149 | main.log.exception( "Graph: Uncaught exception" ) |
| 150 | return main.ERROR |
| 151 | |
| 152 | def getNonCutEdges( self ): |
| 153 | """ |
| 154 | Get a list of non-cut-edges (non-bridges). |
| 155 | The definition of a cut-edge (bridge) is: the deletion of a cut-edge will |
| 156 | increase the number of connected component of a graph. |
| 157 | The function is realized by impelementing Schmidt's algorithm based on |
| 158 | chain decomposition. |
| 159 | Returns a list of edges, e.g. |
| 160 | [ [ vertex1, vertex2 ], [ vertex2, vertex3 ] ] |
| 161 | """ |
| 162 | try: |
| 163 | if not self.depthFirstSearch(): |
| 164 | return None |
| 165 | if not self.findChains(): |
| 166 | return None |
| 167 | nonCutEdges = [] |
| 168 | for chain in self.chains: |
| 169 | for edge in chain: |
| 170 | nonCutEdges.append( edge ) |
You Wang | 9a589b9 | 2016-07-14 09:43:21 -0700 | [diff] [blame] | 171 | #main.log.debug( 'Non-cut-edges: {}'.format( nonCutEdges ) ) |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 172 | return nonCutEdges |
| 173 | except Exception: |
| 174 | main.log.exception( "Graph: Uncaught exception" ) |
| 175 | return None |
| 176 | |
| 177 | def getNonCutVertices( self ): |
| 178 | """ |
| 179 | Get a list of non-cut-vertices. |
| 180 | The definition of a cut-vertex is: the deletion of a cut-vertex will |
| 181 | increase the number of connected component of a graph. |
| 182 | The function is realized by impelementing Schmidt's algorithm based on |
| 183 | chain decomposition. |
| 184 | Returns a list of vertices, e.g. [ vertex1, vertex2, vertex3 ] |
| 185 | """ |
| 186 | try: |
| 187 | nonCutEdges = self.getNonCutEdges() |
| 188 | # find all cycle chains |
| 189 | cycleChains = [] |
| 190 | for chain in self.chains: |
| 191 | # if the source vertex of the first chain equals to the destination vertex of the last |
| 192 | # chain, the chain is a cycle chain |
| 193 | if chain[ 0 ][ 0 ] == chain[ -1 ][ 1 ]: |
| 194 | cycleChains.append( chain ) |
You Wang | 9a589b9 | 2016-07-14 09:43:21 -0700 | [diff] [blame] | 195 | #main.log.debug( 'Cycle chains: {}'.format( cycleChains ) ) |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 196 | # Get a set of vertices which are the first vertices of a cycle chain (excluding the first |
| 197 | # cycle chain), and these vertices are a subset of all cut-vertices |
| 198 | subsetOfCutVertices = [] |
| 199 | if len( cycleChains ) > 1: |
| 200 | for cycleChain in cycleChains[ 1: ]: |
| 201 | subsetOfCutVertices.append( cycleChain[ 0 ][ 0 ] ) |
You Wang | 9a589b9 | 2016-07-14 09:43:21 -0700 | [diff] [blame] | 202 | #main.log.debug( 'Subset of cut vertices: {}'.format( subsetOfCutVertices ) ) |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 203 | nonCutVertices = [] |
| 204 | assert nonCutEdges != None |
| 205 | for vertex in self.graphDict.keys(): |
| 206 | if vertex in subsetOfCutVertices: |
| 207 | continue |
| 208 | vertexIsNonCut = True |
| 209 | for neighbor in self.graphDict[ vertex ][ 'edges' ].keys(): |
| 210 | edge = [ vertex, neighbor ] |
| 211 | backwardEdge = [ neighbor, vertex ] |
| 212 | if not edge in nonCutEdges and not backwardEdge in nonCutEdges: |
| 213 | vertexIsNonCut = False |
| 214 | break |
| 215 | if vertexIsNonCut: |
| 216 | nonCutVertices.append( vertex ) |
You Wang | 9a589b9 | 2016-07-14 09:43:21 -0700 | [diff] [blame] | 217 | #main.log.debug( 'Non-cut-vertices: {}'.format( nonCutVertices ) ) |
You Wang | b1a16c3 | 2016-05-23 15:30:52 -0700 | [diff] [blame] | 218 | return nonCutVertices |
| 219 | except KeyError: |
| 220 | main.log.exception( "Graph: KeyError exception found" ) |
| 221 | return None |
| 222 | except AssertionError: |
| 223 | main.log.exception( "Graph: AssertionError exception found" ) |
| 224 | return None |
| 225 | except Exception: |
| 226 | main.log.exception( "Graph: Uncaught exception" ) |
| 227 | return None |
| 228 | |
| 229 | def depthFirstSearch( self ): |
| 230 | """ |
| 231 | This function runs a depth-first search and gets DFI of each vertex as well |
| 232 | as generates the back edges |
| 233 | """ |
| 234 | try: |
| 235 | assert self.graphDict != None and len( self.graphDict ) != 0 |
| 236 | for vertex in self.graphDict.keys(): |
| 237 | self.DFI[ vertex ] = -1 |
| 238 | self.parentVertexInDFS[ vertex ] = '' |
| 239 | self.parentEdgeInDFS[ vertex ] = None |
| 240 | firstVertex = self.graphDict.keys()[ 0 ] |
| 241 | self.currentDFI = 0 |
| 242 | self.backEdges = {} |
| 243 | if not self.depthFirstSearchRecursive( firstVertex ): |
| 244 | return main.ERROR |
| 245 | return main.TRUE |
| 246 | except KeyError: |
| 247 | main.log.exception( "Graph: KeyError exception found" ) |
| 248 | return main.ERROR |
| 249 | except AssertionError: |
| 250 | main.log.exception( "Graph: AssertionError exception found" ) |
| 251 | return main.ERROR |
| 252 | except Exception: |
| 253 | main.log.exception( "Graph: Uncaught exception" ) |
| 254 | return main.ERROR |
| 255 | |
| 256 | def depthFirstSearchRecursive( self, vertex ): |
| 257 | """ |
| 258 | Recursive function for depth-first search |
| 259 | """ |
| 260 | try: |
| 261 | self.DFI[ vertex ] = self.currentDFI |
| 262 | self.currentDFI += 1 |
| 263 | for neighbor in self.graphDict[ vertex ][ 'edges' ].keys(): |
| 264 | edge = [ vertex, neighbor ] |
| 265 | backwardEdge = [ neighbor, vertex ] |
| 266 | if neighbor == self.parentVertexInDFS[ vertex ]: |
| 267 | continue |
| 268 | elif self.DFI[ neighbor ] == -1: |
| 269 | self.parentVertexInDFS[ neighbor ] = vertex |
| 270 | self.parentEdgeInDFS[ neighbor ] = backwardEdge |
| 271 | if not self.depthFirstSearchRecursive( neighbor ): |
| 272 | return main.ERROR |
| 273 | else: |
| 274 | key = self.DFI[ neighbor ] |
| 275 | if key in self.backEdges.keys(): |
| 276 | if not edge in self.backEdges[ key ] and\ |
| 277 | not backwardEdge in self.backEdges[ key ]: |
| 278 | self.backEdges[ key ].append( backwardEdge ) |
| 279 | else: |
| 280 | tempKey = self.DFI[ vertex ] |
| 281 | if tempKey in self.backEdges.keys(): |
| 282 | if not edge in self.backEdges[ tempKey ] and\ |
| 283 | not backwardEdge in self.backEdges[ tempKey ]: |
| 284 | self.backEdges[ key ] = [ backwardEdge ] |
| 285 | else: |
| 286 | self.backEdges[ key ] = [ backwardEdge ] |
| 287 | return main.TRUE |
| 288 | except KeyError: |
| 289 | main.log.exception( "Graph: KeyError exception found" ) |
| 290 | return main.ERROR |
| 291 | except Exception: |
| 292 | main.log.exception( "Graph: Uncaught exception" ) |
| 293 | return main.ERROR |
| 294 | |
| 295 | def findChains( self ): |
| 296 | """ |
| 297 | This function finds all the chains in chain-decomposition algorithm |
| 298 | """ |
| 299 | keyList = self.backEdges.keys() |
| 300 | keyList.sort() |
| 301 | vertexIsVisited = {} |
| 302 | self.chains = [] |
| 303 | for vertex in self.graphDict.keys(): |
| 304 | vertexIsVisited[ vertex ] = False |
| 305 | try: |
| 306 | for key in keyList: |
| 307 | backEdgeList = self.backEdges[ key ] |
| 308 | for edge in backEdgeList: |
| 309 | chain = [] |
| 310 | currentEdge = edge |
| 311 | sourceVertex = edge[ 0 ] |
| 312 | while True: |
| 313 | currentVertex = currentEdge[ 0 ] |
| 314 | nextVertex = currentEdge[ 1 ] |
| 315 | vertexIsVisited[ currentVertex ] = True |
| 316 | chain.append( currentEdge ) |
| 317 | if nextVertex == sourceVertex or vertexIsVisited[ nextVertex ] == True: |
| 318 | break |
| 319 | currentEdge = self.parentEdgeInDFS[ nextVertex ] |
| 320 | self.chains.append( chain ) |
| 321 | return main.TRUE |
| 322 | except KeyError: |
| 323 | main.log.exception( "Graph: KeyError exception found" ) |
| 324 | return main.ERROR |
| 325 | except Exception: |
| 326 | main.log.exception( "Graph: Uncaught exception" ) |
| 327 | return main.ERROR |