void ProcessPairedEndReads(const string& command, const string& index_file,
const string& reads_file_p1,
const string& reads_file_p2,
const string& output_file,
const uint32_t& n_reads_to_process,
const uint32_t& max_mismatches,
const string& adaptor, const uint32_t& top_k,
const int& frag_range, const bool& ambiguous,
const bool& unmapped, const bool& SAM,
const int& num_of_threads) {
// LOAD THE INDEX HEAD INFO
Genome genome;
HashTable hash_table;
uint32_t size_of_index;
ReadIndexHeadInfo(index_file, genome, size_of_index);
genome.sequence.resize(genome.length_of_genome);
hash_table.counter.resize(power(4, F2SEEDKEYWIGTH) + 1);
hash_table.index.resize(size_of_index);
vector<vector<string> > index_names(2, vector<string>(2));
index_names[0][0] = index_file + "_CT00";
index_names[0][1] = index_file + "_CT01";
index_names[1][0] = index_file + "_GA10";
index_names[1][1] = index_file + "_GA11";
vector<vector<string> > read_names(2, vector<string>(n_reads_to_process));
vector<vector<string> > read_seqs(2, vector<string>(n_reads_to_process));
vector<vector<string> > read_scores(2, vector<string>(n_reads_to_process));
vector<int> ranked_results_size(2);
vector<vector<CandidatePosition> > ranked_results(2,
vector<CandidatePosition>(MAX_NUM_EXACT_MAPPED));
vector<vector<TopCandidates> > top_results(2,
vector<TopCandidates>(n_reads_to_process));
FILE * fin[2];
fin[0] = fopen(reads_file_p1.c_str(), "r");
if (!fin[0]) {
throw SMITHLABException("cannot open input file " + reads_file_p1);
}
fin[1] = fopen(reads_file_p2.c_str(), "r");
if (!fin[1]) {
throw SMITHLABException("cannot open input file " + reads_file_p2);
}
string adaptors[2];
extract_adaptors(adaptor, adaptors[0], adaptors[1]);
clock_t start_t = clock();
FILE * fout = fopen(output_file.c_str(), "w");
if (!fout) {
throw SMITHLABException("cannot open input file " + output_file);
}
uint32_t num_of_reads[2];
StatPairedReads stat_paired_reads(ambiguous, unmapped, output_file, SAM);
bool AG_WILDCARD = true;
fprintf(stderr, "[MAPPING PAIRED-END READS FROM THE FOLLOWING TWO FILES]\n");
fprintf(stderr, " %s (AND)\n %s\n", reads_file_p1.c_str(),
reads_file_p2.c_str());
fprintf(stderr, "[OUTPUT MAPPING RESULTS TO %s]\n", output_file.c_str());
if (SAM) {
SAMHead(index_file, command, fout);
}
omp_set_dynamic(0);
omp_set_num_threads(num_of_threads);
for (uint32_t i = 0;; i += n_reads_to_process) {
num_of_reads[0] = num_of_reads[1] = 0;
for (uint32_t pi = 0; pi < 2; ++pi) { // paired end reads _1 and _2
AG_WILDCARD = pi == 1 ? true : false;
LoadReadsFromFastqFile(fin[pi], i, n_reads_to_process, adaptors[pi],
num_of_reads[pi], read_names[pi], read_seqs[pi],
read_scores[pi]);
if (num_of_reads[pi] == 0)
break;
//Initialize the paired results
for (uint32_t j = 0; j < num_of_reads[pi]; ++j) {
top_results[pi][j].Clear();
top_results[pi][j].SetSize(top_k);
}
for (uint32_t fi = 0; fi < 2; ++fi) {
ReadIndex(index_names[pi][fi], genome, hash_table);
char strand = fi == 0 ? '+' : '-';
#pragma omp parallel for
for (uint32_t j = 0; j < num_of_reads[pi]; ++j) {
PairEndMapping(read_seqs[pi][j], genome, hash_table, strand,
AG_WILDCARD, max_mismatches, top_results[pi][j]);
}
}
}
if (num_of_reads[0] != num_of_reads[1]) {
fprintf(stderr,
"The number of reads in paired-end files should be the same.\n");
exit( EXIT_FAILURE);
}
if (num_of_reads[0] == 0) {
break;
}
stat_paired_reads.total_read_pairs += num_of_reads[0];
///////////////////////////////////////////////////////////
// Merge Paired-end results
for (uint32_t j = 0; j < num_of_reads[0]; ++j) {
for (uint32_t pi = 0; pi < 2; ++pi) {
ranked_results_size[pi] = 0;
while (!top_results[pi][j].candidates.empty()) {
ranked_results[pi][ranked_results_size[pi]++] =
top_results[pi][j].Top();
top_results[pi][j].Pop();
}
}
MergePairedEndResults(genome, read_names[0][j], read_seqs[0][j],
read_scores[0][j], read_seqs[1][j],
read_scores[1][j], ranked_results,
ranked_results_size, frag_range, max_mismatches,
SAM, stat_paired_reads, fout);
}
if (num_of_reads[0] < n_reads_to_process)
break;
}
fclose(fin[0]);
fclose(fin[1]);
fclose(fout);
OutputPairedStatInfo(stat_paired_reads, output_file);
fprintf(stderr, "[MAPPING TAKES %.0lf SECONDS]\n",
(double(clock() - start_t) / CLOCKS_PER_SEC));
}
Info API::DeleteFrom(SqlCommandDeleteFrom* command){
CatalogManager catalog_manager;
IndexManager index_manager;
RecordManager record_manager;
bool delete_all_records = command->delete_all_records();
std::string table_name = command->table_name();
std::vector<WhereClause> where_clause = command->where_clause();
if (!catalog_manager.HasTable(table_name)){
std::string error_info;
error_info = "Table \"" + table_name + "\" not exists.";
return Info(error_info);
}
auto table = catalog_manager.GetTableInfo(table_name);
for (auto it: where_clause){
if(!table.HasAttribute(it.kColumnName)){
std::string error_info;
error_info = "attribute \"" + it.kColumnName + "\"in where clause \""+ it.kColumnName + it.kOperator + it.kCondition +"\" not exists.";
return Info(error_info);
}
}
if (delete_all_records){
if(record_manager.DeleteAllRecords(table)){
//update index
for (auto it : table.index_names()){
IndexInfo index_info = catalog_manager.GetIndexInfo(it);
index_manager.EmptyIndex(index_info);
}
return Info();
}
else{
return Info("Delete records failed");
}
}
else{
std::vector<WhereClause> where_clause_with_index;
std::vector<WhereClause> where_clause_without_index;
for (auto it : where_clause){
if (!table.attribute(it.kColumnName).index_names().empty() && it.kOperator != "!="){
where_clause_with_index.push_back(it);
}
else{
where_clause_without_index.push_back(it);
}
}
std::vector<std::pair<int,int>> results;
if (!where_clause_with_index.empty()) //条件里有属性有 index
{
for (auto it : where_clause_with_index){
std::string index_name = table.attribute(it.kColumnName).index_names().at(0);
IndexInfo index = catalog_manager.GetIndexInfo(index_name);
std::vector<int> offsets_of_a_clause = index_manager.FindRecords(table,index, it);
if(offsets_of_a_clause.empty()){
std::string error_info;
error_info = "record not found with index, deletion failed";
return Info(error_info);
}
std::sort(offsets_of_a_clause.begin(), offsets_of_a_clause.end());
offsets_of_a_clause.erase(std::unique(offsets_of_a_clause.begin(), offsets_of_a_clause.end()), offsets_of_a_clause.end());
std::vector<std::pair<int,int>> results_of_a_clause = record_manager.FindRecordsWithIndex(offsets_of_a_clause, table, it);
if (results.empty()){
results = results_of_a_clause;
}
else{
std::vector<std::pair<int,int> > results_temp;
std::set_intersection(results.begin(),results.end(),
results_of_a_clause.begin(),results_of_a_clause.end(),
std::back_inserter(results_temp));
results = results_temp;
}
}
for(auto it : where_clause_without_index){
if(results.empty()){
std::string error_info;
error_info = "record not found with index, deletion failed";
return Info(error_info);
}
results = record_manager.RecordsFilter(results, table, it);
}
}
else{ //条件里属性都没有 index
for (auto it: where_clause_without_index){
if (results.empty()){
results = record_manager.FindRecordsWithNoIndex(table, it);
}
else{
results = record_manager.RecordsFilter(results,table, it);
}
}
if(results.empty()){
std::string error_info;
error_info = "record not found with no index, deletion failed";
return Info(error_info);
}
}
if(results.empty()){
std::string error_info;
error_info = "record not found, deletion failed";
return Info(error_info);
}
auto records = record_manager.SelectRecords(results, table);
int ret = record_manager.DeleteRecords(results, table);
if(!ret){
return Info("Delete records failed.");
}
else{
auto index_names = table.index_names();
int pair_index = 0;
for (auto it : records){
for (auto i : index_names){
IndexInfo index_info = catalog_manager.GetIndexInfo(i);
auto index = table.attribute_index(index_info.attribute_name());
std::string value = it.at(index);
int offset_i = results.at(pair_index).first;
bool ret = index_manager.DeleteRecord(table, index_info, value, offset_i);
if(!ret){
return Info("Delete record in index \"" + index_info.name() + "on attribute \"" + index_info.attribute_name()
+"\" of table \"" + index_info.table_name() + "\"failed.");
}
}
pair_index++;
}
return Info();
}
}
}
