Add files via upload

src/Blast.py

@@ -4,7 +4,7 @@
 from typing import List
 import copy
 
-__version__ = "0.2.2"
+__version__ = "0.2.3"
 
 class BLAST:
     def __init__(self,

src/Step1_preparation.cpp

@@ -45,9 +45,9 @@ int main(int argc, char** argv) {
         exit(0);
     }
 
-    std::ofstream pre_res_output(pre_res_path, std::ios::trunc);
+    std::ofstream pre_res_output(pre_res_path, std::ios::trunc); // m = overwrite
 
-    int start_c = 10000;
+    int start_c = 0;
     std::string strain_naam;
 
     for (const auto& entry : fs::directory_iterator(raw_annotated_dir_path)) { // range iteration to get abs path

src/Step2_simple_derep.cpp

@@ -98,7 +98,7 @@ int main(int argc, char** argv) {
 
             std::string a_line, info;
             std::string a_seq = "";
-            int seq_id = 10000; // start
+            int seq_id = 0; // start
 
             std::unordered_set<std::string> added_seq = {"added_seq"};
 

src/Step3_pre_cluster.cpp

@@ -78,7 +78,7 @@ int main(int argc, char** argv) {
     if (!(fs::exists(parent_dereped_cated_fasta_pth))) {
         std::cerr << "Error: parent path provided to '-d or --concatenated_fasta' does not exist. 路径不存在\n";
         exit(0);
-    }
+    } 
     else if (!(fs::exists(parent_nr_dir))) {
         std::cerr << "Error: parent path provided to '-n or --nr_genomes' does not exist. 路径不存在\n";
         exit(0);
@@ -119,14 +119,17 @@ int main(int argc, char** argv) {
 
         // first
         getline (a_dereped_fasta, a_line);
-        seq_id = a_line.substr(1, 11);
-        info = a_line.substr(13);
+
+        size_t  space_indx = a_line.find(" ");
+        seq_id = a_line.substr(1, space_indx - 1);
+        info = a_line.substr(space_indx + 1);
 
         while (getline (a_dereped_fasta, a_line)) {
 
             if (a_line[0] == '>') { // if the row of id 
                 // save id info (a previous info)
                 id_info << seq_id + "\t" + info + "\n";
+
                 // operate on previous SeqRecord using previous seq_id
                 if (pre_cluster.count(a_seq) == 0) { // if no prescence before, need comparison with find
                     pre_cluster[a_seq].push_back(seq_id);
@@ -145,16 +148,17 @@ int main(int argc, char** argv) {
 
                     dereped_cated_dict[seq_id] = rec_id + a_seq_to_write;
 
-                    each_d[seq_id.substr(0, 5)].push_back(seq_id);
+                    each_d[seq_id.substr(0, seq_id.find("-"))].push_back(seq_id);
 
                 } else { // if presence before
                     (pre_cluster.at(a_seq)).push_back(seq_id);
 
                 }
 
                 // cover previous seq_id
-                info = a_line.substr(13);
-                seq_id = a_line.substr(1, 11);
+                size_t space_indx = a_line.find(" ");
+                seq_id = a_line.substr(1, space_indx - 1);
+                info = a_line.substr(space_indx + 1);
 
                 // reset previous a_seq
                 a_seq = ""; // reset a_seq
@@ -184,7 +188,7 @@ int main(int argc, char** argv) {
 
             dereped_cated_dict[seq_id] = rec_id + a_seq_to_write;
 
-            each_d[seq_id.substr(0, 5)].push_back(seq_id);
+            each_d[seq_id.substr(0, seq_id.find("-"))].push_back(seq_id);
 
         } else {
             (pre_cluster.at(a_seq)).push_back(seq_id);

src/Step5_query_binning.cpp

@@ -114,7 +114,7 @@ int main(int argc, char** argv) {
             while (std::getline(a_blast_op, a_line)) {
                 a_line = a_line + "\n";
 
-                std::string query = a_line.substr(0, 11);
+                std::string query = a_line.substr(0, a_line.find("\t"));
 
                 size_t h = hash_fn(query);
                 int b = h % b_level;

src/Step6_filter_n_bin.cpp

@@ -152,7 +152,7 @@ int main(int argc, char** argv) {
                 key = element;
             }
             key_++;
-            bin_spe[key].insert(element.substr(0, 5));
+            bin_spe[key].insert(element.substr(0, element.find("-")));
         }
 
     }
@@ -203,7 +203,7 @@ int main(int argc, char** argv) {
                 std::getline(ss_, subject, '\t');
                 std::getline(ss_, score);
 
-                if (query.substr(0, 5) == subject.substr(0, 5)) {
+                if (query.substr(0, query.find("-")) == subject.substr(0, subject.find("-"))) {
                     continue;
                 } else {
                     int score_ = stoi(score);
@@ -227,7 +227,7 @@ int main(int argc, char** argv) {
 
                 for (const auto& [sub, its_score] : nested_um) {
 
-                    std::string s_spe = sub.substr(0, 5);
+                    std::string s_spe = sub.substr(0, sub.find("-"));
 
                     if (bin_spe_->at(qs).find(s_spe) != bin_spe_->at(qs).end()) { // if qs pre cluster has that spe
                         continue;

src/Step9_write_clusters.cpp

@@ -207,15 +207,15 @@ int main(int argc, char** argv) {
 
     // first one
     getline(dereped_cat_file, a_line);
-    std::string key = a_line.substr(1, 11);
+    std::string key = a_line.substr(1, a_line.find(" ") - 1);
     std::string seq = "";
 
     while (getline(dereped_cat_file, a_line)) {
         if (a_line[0] == '>') {
             // save previous one
             fasta_dict[key] = seq;
 
-            key = a_line.substr(1, 11);
+            key = a_line.substr(1, a_line.find(" ") - 1);
             seq = "";
         } else {
             seq = seq + a_line + "\n";
@@ -243,7 +243,7 @@ int main(int argc, char** argv) {
             std::unordered_set<std::string> spe_count;
 
             for (const std::string& s : the_cluster) {
-                spe_count.insert(s.substr(0, 5));
+                spe_count.insert(s.substr(0, s.find("-")));
             }
 
             fs::path file_naam;

src/ThreadPool.h

@@ -12,7 +12,7 @@
 #include <stdexcept>
 #include <string>
 
-extern const std::string __version__ = "0.2.2";
+extern const std::string __version__ = "0.2.3";
 
 class ThreadPool {
 public: