Changeset: 4b58b4d3e908 for MonetDB
URL: http://dev.monetdb.org/hg/MonetDB?cmd=changeset;node=4b58b4d3e908
Modified Files:
monetdb5/extras/rdf/rdflabels.c
monetdb5/extras/rdf/rdfretrieval.c
monetdb5/extras/rdf/rdfschema.c
monetdb5/extras/rdf/rdfschema.h
Branch: rdf
Log Message:
Fix the warning not-protecting-local-variables-variable-length-buffer caused by
Linnea's implementation in rdflabel and rdfretrieval
diffs (236 lines):
diff --git a/monetdb5/extras/rdf/rdflabels.c b/monetdb5/extras/rdf/rdflabels.c
--- a/monetdb5/extras/rdf/rdflabels.c
+++ b/monetdb5/extras/rdf/rdflabels.c
@@ -855,7 +855,7 @@ void createTypeAttributesHistogram(BAT *
int i, j, k;
int fit;
- oid typeAttributesOids[typeAttributesCount];
+ oid *typeAttributesOids = malloc(sizeof(oid) *
typeAttributesCount);
if (BATcount(sbat) == 0) {
fprintf(stderr, "sbat must not be empty");
@@ -944,6 +944,8 @@ void createTypeAttributesHistogram(BAT *
}
}
}
+
+ free(typeAttributesOids);
}
#endif
diff --git a/monetdb5/extras/rdf/rdfretrieval.c
b/monetdb5/extras/rdf/rdfretrieval.c
--- a/monetdb5/extras/rdf/rdfretrieval.c
+++ b/monetdb5/extras/rdf/rdfretrieval.c
@@ -129,14 +129,18 @@ void bfs1(int root, int* table_id, int t
static
void addNode1(int* adjacency_from, int* adjacency_to, int adjacencyCount,
NodeStat* nodeStats, int* table_id, int tableCount, int root, char initial) {
- int queue[tableCount]; // cyclic array
- int visited[tableCount];
- int isInQueue[tableCount];
+ int *queue; // cyclic array
+ int *visited;
+ int *isInQueue;
int queuePosition; // next element in queue to view at
int queueLength;
int pathId, pathIdTmp;
int i;
+ queue = (int*)malloc(sizeof(int) * tableCount);
+ visited = (int*)malloc(sizeof(int) * tableCount);
+ isInQueue = (int*)malloc(sizeof(int) * tableCount);
+
// init
for (i = 0; i < tableCount; ++i) {
queue[i] = -1;
@@ -180,6 +184,10 @@ void addNode1(int* adjacency_from, int*
}
bfs1(root, table_id, tableCount, adjacency_from, adjacency_to,
adjacencyCount, queue, visited, isInQueue, &queuePosition, &queueLength,
nodeStats);
+
+ free(queue);
+ free(visited);
+ free(isInQueue);
}
static
@@ -368,8 +376,9 @@ int* retrieval2(int root, int numNodesMa
static
void assignWeightToChildren3(int* adjacency_from, int* adjacency_to, int
adjacencyCount, NodeStat* nodeStats, int* table_id, int tableCount, int root) {
int i, j, k;
- char visited[tableCount];
-
+ char *visited;
+
+ visited = (char*)malloc(sizeof(char) * tableCount);
// mark root as a "chosen node"
nodeStats[root].steps = 0;
nodeStats[root].weight = 0;
@@ -406,6 +415,8 @@ void assignWeightToChildren3(int* adjace
nodeStats[i].weight += nodeStats[i].origWeight;
}
}
+
+ free(visited);
}
static
@@ -607,14 +618,18 @@ int* retrievalOverview(int* numNodesActu
int csCount = 0;
int sumChosenSubjects = 0;
- int queue[tableCount]; // cyclic array
- int isInQueue[tableCount];
+ int *queue; // cyclic array
+ int *isInQueue;
int queuePosition; // next element in queue to view at
int queueLength;
- char visited[tableCount];
+ char *visited;
int subgraphSize;
Groups groups;
int *chosenNodes = NULL;
+
+ queue = (int*)malloc(sizeof(int) * tableCount);
+ isInQueue = (int*)malloc(sizeof(int) * tableCount);
+ visited = (char*)malloc(sizeof(char) * tableCount);
groups.count = 0;
groups.groups = NULL;
@@ -695,6 +710,10 @@ int* retrievalOverview(int* numNodesActu
assert(j == subgraphSize);
}
+ free(queue);
+ free(visited);
+ free(isInQueue);
+
// transitive closure (Floyd-Warshall-Algorithm)
for (k = 0; k < tableCount; ++k) {
for (i = 0; i < tableCount; ++i) {
@@ -732,7 +751,7 @@ int* retrievalOverview(int* numNodesActu
}
if (!found) {
int node;
- char reachability[tableCount];
+ char *reachability = malloc(sizeof(char) * tableCount);
int reachabilityCount = 0;
int nextNode; // position in the (sorted) list of nodes
to look at next
@@ -778,6 +797,8 @@ int* retrievalOverview(int* numNodesActu
}
}
}
+
+ free(reachability);
}
}
diff --git a/monetdb5/extras/rdf/rdfschema.c b/monetdb5/extras/rdf/rdfschema.c
--- a/monetdb5/extras/rdf/rdfschema.c
+++ b/monetdb5/extras/rdf/rdfschema.c
@@ -3527,6 +3527,27 @@ void mergeMaxFreqCSByS6(CSrel *csrelMerg
}
+/*
+static
+str printTKNZStringFromOid(oid id){
+ int ret;
+ char* schema = "rdf";
+ str propStr;
+
+ if (TKNZRopen (NULL, &schema) != MAL_SUCCEED) {
+ throw(RDF, "rdf.rdfschema",
+ "could not open the tokenizer\n");
+ }
+
+ takeOid(id, &propStr);
+ printf("String for "BUNFMT": %s\n", id, propStr);
+
+ TKNZRclose(&ret);
+
+ return MAL_SUCCEED;
+}
+*/
+
static
char isSemanticSimilar(int freqId1, int freqId2, CSlabel* labels,
OntoUsageNode *tree, int numOrigFreqCS, oid *ancestor){ /*Rule S1 S2 S3*/
int i, j;
@@ -3599,9 +3620,18 @@ char isSemanticSimilar(int freqId1, int
}
level++;
}
- //printf("The common ancestor of freqCS %d (%s) and freqCS %d
(%s) is: %s --- %f \n", freqId1, labels[freqId1].name, freqId2,
labels[freqId2].name, tmpNode->uri, tmpNode->percentage);
+ /*
+ printf("The common ancestor of freqCS %d ("BUNFMT") and freqCS
%d ("BUNFMT") is: "BUNFMT" --- %f \n", freqId1, labels[freqId1].name, freqId2,
labels[freqId2].name, tmpNode->uri, tmpNode->percentage);
+
+ printTKNZStringFromOid(labels[freqId1].name);
+ printTKNZStringFromOid(labels[freqId2].name);
+ printTKNZStringFromOid(tmpNode->uri);
+ */
+
if (tmpNode->percentage < IMPORTANCE_THRESHOLD) {
- //printf("Merge two CS's %s and %s using the common
ancestor (%s) at level %d (score:
%f)\n",labels[freqId1].name,labels[freqId2].name,tmpNode->uri,
i,tmpNode->percentage);
+ printf("Merge two CS's %d (Label: "BUNFMT") and %d
(Label: "BUNFMT") using the common ancestor ("BUNFMT") at level %d (score:
%f)\n",
+ freqId1, labels[freqId1].name, freqId2,
labels[freqId2].name,tmpNode->uri, i,tmpNode->percentage);
+
(*ancestor) = tmpNode->uri;
return 1;
}
@@ -3722,7 +3752,7 @@ void mergeCSByS3S5(CSset *freqCSset, CSl
#else
if (simscore > SIM_THRESHOLD) {
#endif
- //printf("S3S5: merge freqCS %d and freqCS %d
\n", freqId1, freqId2);
+ //printf("S3S5: merge freqCS %d and freqCS %d
(sim: %f)\n", freqId1, freqId2,simscore);
//Check whether these CS's belong to any mergeCS
if (cs1->parentFreqIdx == -1 &&
cs2->parentFreqIdx == -1){ /* New merge */
mergecs = mergeTwoCSs(*cs1,*cs2,
freqId1,freqId2, *mergecsId);
@@ -5774,14 +5804,17 @@ RDFextractCSwithTypes(int *ret, bat *sba
mergeCSFreqCSMap = (oid*) malloc(sizeof(oid) * curNumMergeCS);
initMergeCSFreqCSMap(freqCSset, mergeCSFreqCSMap);
+
+ /* S6: Merged CS referred from the same CS via the same property */
+ if (1){
tmpCSrelToMergeCS = generateCsRelToMergeFreqSet(csrelSet, freqCSset);
tmpNumRel = freqCSset->numCSadded;
- /* S6: Merged CS referred from the same CS via the same property */
- if (1) mergeMaxFreqCSByS6(tmpCSrelToMergeCS, freqCSset, labels,
mergeCSFreqCSMap, curNumMergeCS, &mergecsId, ontmetadata, ontmetadataCount);
+ mergeMaxFreqCSByS6(tmpCSrelToMergeCS, freqCSset, labels,
mergeCSFreqCSMap, curNumMergeCS, &mergecsId, ontmetadata, ontmetadataCount);
//printf("DISABLE S6 (For Testing) \n");
freeCSrelSet(tmpCSrelToMergeCS,tmpNumRel);
+ }
curNumMergeCS = countNumberMergeCS(freqCSset);
curT = clock();
diff --git a/monetdb5/extras/rdf/rdfschema.h b/monetdb5/extras/rdf/rdfschema.h
--- a/monetdb5/extras/rdf/rdfschema.h
+++ b/monetdb5/extras/rdf/rdfschema.h
@@ -125,11 +125,11 @@ typedef struct PropStat {
/* ---- For detecting dimension table */
#define NUM_ITERATION_FOR_IR 3 /* Number of iteration for
indirect referrences to a CS (table) */
-#define IR_DIMENSION_THRESHOLD_PERCENTAGE 0.2 /* Score of indirect
references that the CS can be considered as a dimension CS
+#define IR_DIMENSION_THRESHOLD_PERCENTAGE 0.02 /* Score of indirect
references that the CS can be considered as a dimension CS
IR_DIMENSION_THRESHOLD_PERCENTAGE * totalFrequency
- Number of IR
references should be several times larger than the CS frequency
+ Number of IR
references should be several times larger than the CS frequenc
*/
-//#define IR_DIMENSION_THRESHOLD_PERCENTAGE 0.02 //Value 0.2 is for
example data only
+//#define IR_DIMENSION_THRESHOLD_PERCENTAGE 0.2 //Value 0.2 is for
example data only
#define NOT_MERGE_DIMENSIONCS 1 /* Default: 1, 0: Is for
example data */
@@ -214,6 +214,7 @@ typedef struct SubCSSet{
// the CS's to-be-merged in this rule must
cover > MIN_FROMTABLE_SIZE_S6 / MIN_PERCETAGE_S6 triples
//#define MIN_FROMTABLE_SIZE_S6 1 /* For example data */
#define MINIMUM_TABLE_SIZE 10000 //The minimum number of triples coverred by
a table (i.e., a final CS)
+//#define MINIMUM_TABLE_SIZE 1 // For example dataset only
#define SAMPLE_FILTER_THRESHOLD 1 // SAMPLE_FILTER_THRESHOLD/ 100
#define HIGH_REFER_THRESHOLD 5
_______________________________________________
checkin-list mailing list
checkin-list@monetdb.org
https://www.monetdb.org/mailman/listinfo/checkin-list
