Parameters *GetParameters(int argc, char *argv[])
{
Parameters *parameters;
parameters = (Parameters *) malloc(sizeof(Parameters));
if (parameters == NULL)
OutOfMemoryError("GetParameters:parameters");
// initialize default parameter settings
parameters->directed = TRUE;
parameters->labelList = AllocateLabelList();
return parameters;
}
int main(int argc, char **argv)
{
Graph *g1, *g2;
LabelList *labelList = NULL;
BOOLEAN directed = TRUE; // 'e' edges are considered directed
ULONG maxVertices;
VertexMap *mapping;
double matchCost;
if (argc != 3)
{
fprintf(stderr, "usage: %s <graph file> <graph file>\n", argv[0]);
exit(1);
}
labelList = AllocateLabelList();
g1 = ReadGraph(argv[1], labelList, directed);
g2 = ReadGraph(argv[2], labelList, directed);
if (g1->numVertices < g2->numVertices)
{
maxVertices = g2->numVertices;
mapping = (VertexMap *) malloc(sizeof(VertexMap) * maxVertices);
matchCost = InexactGraphMatch(g2, g1, labelList, MAX_DOUBLE, mapping);
}
else
{
maxVertices = g1->numVertices;
mapping = (VertexMap *) malloc(sizeof(VertexMap) * maxVertices);
matchCost = InexactGraphMatch(g1, g2, labelList, MAX_DOUBLE, mapping);
}
printf("Match Cost = %f\n", matchCost);
PrintMapping(mapping, maxVertices);
free(mapping);
return 0;
}
Parameters *GetParameters(int argc, char *argv[])
{
Parameters *parameters;
int i;
double doubleArg;
ULONG ulongArg;
BOOLEAN limitSet = FALSE;
FILE *outputFile;
ULONG argumentExists;
parameters = (Parameters *) malloc(sizeof(Parameters));
if (parameters == NULL)
OutOfMemoryError("parameters");
// initialize default parameter settings
parameters->directed = TRUE;
parameters->limit = 0;
parameters->numBestSubs = 3;
parameters->beamWidth = 4;
parameters->valueBased = FALSE;
parameters->prune = FALSE;
strcpy(parameters->outFileName, "none");
parameters->outputToFile = FALSE;
parameters->outputLevel = 2;
parameters->allowInstanceOverlap = FALSE;
parameters->threshold = 0.0;
parameters->evalMethod = EVAL_MDL;
parameters->iterations = 1;
strcpy(parameters->psInputFileName, "none");
parameters->predefinedSubs = FALSE;
parameters->minVertices = 1;
parameters->maxVertices = 0; // i.e., infinity
parameters->compress = FALSE;
parameters->mdl = FALSE;
parameters->mdlThreshold = 0.0;
parameters->mpsThreshold = 0.0;
parameters->prob = FALSE;
parameters->mps = FALSE;
parameters->maxAnomalousScore = MAX_DOUBLE;
parameters->minAnomalousScore = 0.0;
parameters->noAnomalyDetection = TRUE;
parameters->norm = 1;
parameters->similarity = 1.0;
if (argc < 2)
{
fprintf(stderr, "input graph file name must be supplied\n");
exit(1);
}
// process command-line options
i = 1;
while (i < (argc - 1))
{
if (strcmp(argv[i], "-beam") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg == 0)
{
fprintf(stderr, "%s: beam must be greater than zero\n", argv[0]);
exit(1);
}
parameters->beamWidth = ulongArg;
}
else if (strcmp(argv[i], "-compress") == 0)
{
parameters->compress = TRUE;
}
else if (strcmp(argv[i], "-eval") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if ((ulongArg < 1) || (ulongArg > 3))
{
fprintf(stderr, "%s: eval must be 1-3\n", argv[0]);
exit(1);
}
parameters->evalMethod = ulongArg;
}
else if (strcmp(argv[i], "-iterations") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
parameters->iterations = ulongArg;
}
else if (strcmp(argv[i], "-limit") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg == 0)
{
fprintf(stderr, "%s: limit must be greater than zero\n", argv[0]);
exit(1);
}
parameters->limit = ulongArg;
limitSet = TRUE;
}
else if (strcmp(argv[i], "-maxsize") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg == 0)
{
fprintf(stderr, "%s: maxsize must be greater than zero\n", argv[0]);
exit(1);
}
parameters->maxVertices = ulongArg;
}
else if (strcmp(argv[i], "-minsize") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg == 0)
{
fprintf(stderr, "%s: minsize must be greater than zero\n", argv[0]);
exit(1);
}
parameters->minVertices = ulongArg;
}
else if (strcmp(argv[i], "-nsubs") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg == 0)
{
fprintf(stderr, "%s: nsubs must be greater than zero\n", argv[0]);
exit(1);
}
parameters->numBestSubs = ulongArg;
}
else if (strcmp(argv[i], "-out") == 0)
{
i++;
strcpy(parameters->outFileName, argv[i]);
parameters->outputToFile = TRUE;
}
else if (strcmp(argv[i], "-output") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if ((ulongArg < 1) || (ulongArg > 5))
{
fprintf(stderr, "%s: output must be 1-5\n", argv[0]);
exit(1);
}
parameters->outputLevel = ulongArg;
}
else if (strcmp(argv[i], "-overlap") == 0)
{
parameters->allowInstanceOverlap = TRUE;
}
else if (strcmp(argv[i], "-prune") == 0)
{
parameters->prune = TRUE;
}
else if (strcmp(argv[i], "-ps") == 0)
{
i++;
strcpy(parameters->psInputFileName, argv[i]);
parameters->predefinedSubs = TRUE;
}
else if (strcmp(argv[i], "-threshold") == 0)
{
i++;
sscanf(argv[i], "%lf", &doubleArg);
if ((doubleArg < (double) 0.0) || (doubleArg > (double) 1.0))
{
fprintf(stderr, "%s: threshold must be 0.0-1.0\n", argv[0]);
exit(1);
}
parameters->threshold = doubleArg;
}
else if (strcmp(argv[i], "-undirected") == 0)
{
parameters->directed = FALSE;
}
else if (strcmp(argv[i], "-valuebased") == 0)
{
parameters->valueBased = TRUE;
}
else if (strcmp(argv[i], "-mdl") == 0)
{
i++;
argumentExists = sscanf(argv[i], "%lf", &doubleArg);
if ((argumentExists != 1) || (doubleArg <= (double) 0.0) ||
(doubleArg >= (double) 1.0))
{
fprintf(stderr, "%s: Information Theoretic (MDL) threshold must be greater than 0.0 and less than 1.0\n", argv[0]);
exit(1);
}
parameters->mdl = TRUE;
parameters->mdlThreshold = doubleArg;
}
else if (strcmp(argv[i], "-prob") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg < 2)
{
fprintf(stderr, "%s: you must include a value greater than 1 as a parameter to the probabilistic anomaly detection method.\n", argv[0]);
exit(1);
}
parameters->prob = TRUE;
parameters->iterations = ulongArg; // overrides -iterations specification
parameters->maxAnomalousScore = 1.0; // overrides default of MAX_DOUBLE
}
else if (strcmp(argv[i], "-mps") == 0)
{
i++;
argumentExists = sscanf(argv[i], "%lf", &doubleArg);
if ((argumentExists != 1) || (doubleArg <= (double) 0.0) ||
(doubleArg >= (double) 1.0))
{
fprintf(stderr, "%s: Maximum Partial Substructure (MPS) threshold must be greater than 0.0 and less than 1.0\n", argv[0]);
exit(1);
}
parameters->mps = TRUE;
parameters->mpsThreshold = doubleArg;
}
else if (strcmp(argv[i], "-maxAnomalousScore") == 0)
{
i++;
argumentExists = sscanf(argv[i], "%lf", &doubleArg);
if ((argumentExists != 1) || (doubleArg <= (double) 0.0) ||
(doubleArg >= (double) MAX_DOUBLE))
{
fprintf(stderr, "%s: maximum anomalous score must be greater than 0.0 and less than %lf\n",
argv[0],MAX_DOUBLE);
exit(1);
}
//
// NOTE: This check assumes that the user has specified a max
// anomalous score AFTER specifying they want to run the GBAD-P
// algorithm.
//
if (((doubleArg <= (double) 0.0) || (doubleArg >= 1.0)) &&
(parameters->prob))
{
fprintf(stderr, "%s: maximum anomalous score must be greater than 0.0 and less than 1.0\n",argv[0]);
exit(1);
}
parameters->maxAnomalousScore = doubleArg;
}
else if (strcmp(argv[i], "-minAnomalousScore") == 0)
{
i++;
argumentExists = sscanf(argv[i], "%lf", &doubleArg);
if ((argumentExists != 1) || (doubleArg < (double) 0.0) ||
(doubleArg >= (double) MAX_DOUBLE))
{
fprintf(stderr, "%s: minimum anomalous score must be greater than or equal to 0.0 and less than %lf\n", argv[0],MAX_DOUBLE);
exit(1);
}
parameters->minAnomalousScore = doubleArg;
}
else if (strcmp(argv[i], "-norm") == 0)
{
i++;
sscanf(argv[i], "%lu", &ulongArg);
if (ulongArg < 1)
{
fprintf(stderr, "%s: you must specify a value of 1 or greater.\n", argv[0]);
exit(1);
}
parameters->norm = ulongArg;
}
else if (strcmp(argv[i], "-sim") == 0)
{
i++;
sscanf(argv[i], "%lf", &doubleArg);
if ((doubleArg < (double) 0.0) || (doubleArg > (double) 1.0))
{
fprintf(stderr, "%s: similarity measurement must be between 0.0 and 1.0.\n", argv[0]);
exit(1);
}
parameters->similarity = doubleArg;
}
else
{
fprintf(stderr, "%s: unknown option %s\n", argv[0], argv[i]);
exit(1);
}
i++;
}
if ((parameters->mdl) || (parameters->prob) || (parameters->mps))
parameters->noAnomalyDetection = FALSE;
if (parameters->iterations == 0)
parameters->iterations = MAX_UNSIGNED_LONG; // infinity
// initialize log2Factorial[0..1]
parameters->log2Factorial = (double *) malloc(2 * sizeof(double));
if (parameters->log2Factorial == NULL)
OutOfMemoryError("GetParameters:parameters->log2Factorial");
parameters->log2FactorialSize = 2;
parameters->log2Factorial[0] = 0; // lg(0!)
parameters->log2Factorial[1] = 0; // lg(1!)
// read graphs from input file
strcpy(parameters->inputFileName, argv[argc - 1]);
parameters->labelList = AllocateLabelList();
parameters->posGraph = NULL;
parameters->numPosEgs = 0;
parameters->posEgsVertexIndices = NULL;
ReadInputFile(parameters);
if (parameters->evalMethod == EVAL_MDL)
{
parameters->posGraphDL = MDL(parameters->posGraph,
parameters->labelList->numLabels, parameters);
}
// read predefined substructures
parameters->numPreSubs = 0;
if (parameters->predefinedSubs)
ReadPredefinedSubsFile(parameters);
parameters->incrementList = malloc(sizeof(IncrementList));
parameters->incrementList->head = NULL;
// create output file, if given
if (parameters->outputToFile)
{
outputFile = fopen(parameters->outFileName, "w");
if (outputFile == NULL)
{
printf("ERROR: unable to write to output file %s\n",
parameters->outFileName);
exit(1);
}
fclose(outputFile);
}
if (parameters->numPosEgs == 0)
{
fprintf(stderr, "ERROR: no positive graphs defined\n");
exit(1);
}
// Check bounds on discovered substructures' number of vertices
if (parameters->maxVertices == 0)
parameters->maxVertices = parameters->posGraph->numVertices;
if (parameters->maxVertices < parameters->minVertices)
{
fprintf(stderr, "ERROR: minsize exceeds maxsize\n");
exit(1);
}
// Set limit accordingly
if (parameters->limit == 0)
{
parameters->limit = parameters->posGraph->numEdges / 2;
}
return parameters;
}
Parameters *GetParameters(int argc, char *argv[])
{
Parameters *parameters;
int i;
double doubleArg;
if (argc < 3)
{
fprintf(stderr, "%s: not enough arguments\n", argv[0]);
exit(1);
}
parameters = (Parameters *) malloc(sizeof(Parameters));
if (parameters == NULL)
OutOfMemoryError("GetParameters:parameters");
// initialize default parameter settings
parameters->directed = TRUE;
parameters->allowInstanceOverlap = FALSE;
parameters->threshold = 0.0;
parameters->outputToFile = FALSE;
parameters->posGraph = NULL;
parameters->negGraph = NULL;
// process command-line options
i = 1;
while (i < (argc - 2))
{
if (strcmp(argv[i], "-dot") == 0)
{
i++;
strcpy(parameters->outFileName, argv[i]);
parameters->outputToFile = TRUE;
}
else if (strcmp(argv[i], "-overlap") == 0)
{
parameters->allowInstanceOverlap = TRUE;
}
else if (strcmp(argv[i], "-threshold") == 0)
{
i++;
sscanf(argv[i], "%lf", &doubleArg);
if ((doubleArg < 0.0) || (doubleArg > 1.0))
{
fprintf(stderr, "%s: threshold must be 0.0-1.0\n", argv[0]);
exit(1);
}
parameters->threshold = doubleArg;
}
else
{
fprintf(stderr, "%s: unknown option %s\n", argv[0], argv[i]);
exit(1);
}
i++;
}
// initialize log2Factorial[0..1]
parameters->log2Factorial = (double *) malloc(2 * sizeof(double));
if (parameters->log2Factorial == NULL)
OutOfMemoryError("GetParameters:paramters->log2Factorial");
parameters->log2FactorialSize = 2;
parameters->log2Factorial[0] = 0; // lg(0!)
parameters->log2Factorial[1] = 0; // lg(1!)
parameters->labelList = AllocateLabelList();
return parameters;
}
