fix bug for false positive

README.md

@@ -13,6 +13,7 @@ To compile the SeqOthello tool chain, decompress the source code and execute the
 ```
 ./build.sh
 ```
+Then you can find the compile resutls in the ```build/bin``` folder. 
 
 ##Build
 To build the SeqOthello structure, please first prepare the Jellyfish-generated Kmer files. For test purpose, 
@@ -34,6 +35,11 @@ For all prompted input questions, just press Enter to use the default value for
 ```
 And the generated SeqOthello file can be found in _out_ folder.
 
+###Prepare the kmers in parallel
+Each line of the generated scripts contains a command to prepare the files. The commands in each of the generated scripts can be executed in parallel. For example, with GNU Parallel, you can run 
+
+``` cat ConvertToBinary.sh | parallel ```
+
 ##Query
 
 Containment Query
@@ -45,7 +51,7 @@ Coverage Query
 ```build/bin/Query --map-folder=out/ --transcript=test.fa --detail --output=queryresult --lthread=16 > querylog```
 
 ##OnlineQuery
-use the following command to start a server on the machine, (e.g, on TCP port 3322). The service will run as a deamon.
+Use the following command to start a server on the machine, (e.g., on TCP port 3322). The service will run as a deamon.
 
 ```build/bin/Query --map-folder=out/ --start-server-port 3322```
 

seqothlib/L2Node.cpp

@@ -154,6 +154,7 @@ bool L2EncodedValueListNode::smartQuery(const keyType *k, vector<uint32_t> &ret,
     if (encodetype == L2NodeTypes::VALUE_INDEX_ENCODED) {
         ret.clear();
         if (IOLengthInBytes*index >= lines.size()) return true;
+        if (index==0) return true;
         vector<uint32_t> decode;
         valuelistDecode(&lines[IOLengthInBytes*index], decode, IOLengthInBytes);
         if (decode.size()==0) return true;
@@ -203,8 +204,10 @@ void L2EncodedValueListNode::add(keyType &k, vector<uint32_t> & valuelist) { //
         throw invalid_argument("can not add value list L2EncodedValueListNode");
     keys.push_back(k);
     vector<uint8_t> buff(IOLengthInBytes);
-    if (lines.size() == 0)
+    if (lines.size() == 0) {
         lines.resize(IOLengthInBytes);
+        entrycnt++;
+    }
     uint32_t curr = lines.size();
     lines.resize(lines.size() + IOLengthInBytes);
     valuelistEncode(&lines[curr], valuelist, true);
@@ -218,13 +221,18 @@ void L2EncodedValueListNode::addMAPP(keyType &k, vector<uint8_t> &mapp) {
     if (encodetype!= L2NodeTypes::MAPP)
         throw invalid_argument("can not add bitmap to L2EncodedValueListNode");
     keys.push_back(k);
-    values.push_back(keycnt);
+    //TODO :: Need to pre-pend a record for false positives!
     keycnt++;
     entrycnt++;
+    if (lines.size() == 0) {
+        lines.resize(mapp.size());
+        entrycnt++;
+    }
     if (mapp.size() != IOLengthInBytes) {
         throw invalid_argument("can not add bitmap to L2ShortValuelist type");
     }
     lines.insert(lines.end(), mapp.begin(), mapp.end());
+    values.push_back(keycnt);
 }
 
 void L2ShortValueListNode::writeDataToGzipFile(string gzfname) {

src/query.cc

@@ -50,7 +50,6 @@ void getL1Result(SeqOthello * seqoth, const vector<string> & seq, const vector<i
     printf("%s: L1 finished. Got %lu kmers. \n", get_thid().c_str(), TID.size());
 };
 void getL2Result(uint32_t high, const vector<L2Node *> &pvNodes, const vector<shared_ptr<vector<uint64_t>>> & vpvkmer, const vector<shared_ptr<vector<uint32_t>>> &vTID, unordered_map<int, vector<int>> *pans) {
-
     int myid = workers.fetch_add(1);
     unsigned int totkmer = 0;
     for (auto const & p:vpvkmer)
@@ -89,14 +88,13 @@ void getL2Result(uint32_t high, const vector<L2Node *> &pvNodes, const vector<sh
             /*
             printf("Update pans %d:", TID[i]);
 
-            ConstantLengthKmerHelper<uint64_t, uint16_t> helper(D_KMERLENGTH,0);
+            ConstantLengthKmerHelper<uint64_t, uint16_t> helper(20,0);
                 printf( "%12llx:", kmers[i]);
                 char buf[30];
                 memset(buf,0,sizeof(buf));
                 uint64_t k = kmers[i];
                 helper.convertstring(buf,&k);
                 printf( "%s ", buf);
-
             for (int v = 0 ; v < high; v++)
                 printf("%d ", pans->at(TID[i]).at(v));
             printf("\n");