[FEAT] option to not ignore blanks before the id in fasta files

seqan · Feb 9, 2022 · 0345900 · 0345900
1 parent 23d2bca
commit 0345900
Show file tree

Hide file tree

Showing 2 changed files with 50 additions and 12 deletions.
diff --git a/include/seqan3/io/sequence_file/format_fasta.hpp b/include/seqan3/io/sequence_file/format_fasta.hpp
@@ -194,8 +194,16 @@ class format_fasta
             #if SEQAN3_WORKAROUND_VIEW_PERFORMANCE
                 auto it = stream_view.begin();
                 auto e = stream_view.end();
-                for (; (it != e) && (is_id || is_blank)(*it); ++it)
-                {}
+                if (options.fasta_ignore_blanks_before_id)
+                {
+                    for (; (it != e) && (is_id || is_blank)(*it); ++it)
+                    {}
+                }
+                else
+                {
+                    for (; (it != e) && (is_id)(*it); ++it)
+                    {}
+                }
 
                 bool at_delimiter = false;
                 for (; it != e; ++it)
@@ -215,11 +223,21 @@ class format_fasta
                 {}
 
             #else // ↑↑↑ WORKAROUND | ORIGINAL ↓↓↓
+                if (options.fasta_ignore_blanks_before_id)
+                {
+                    std::ranges::copy(stream_view | std::views::drop_while(is_id || is_blank)        // skip leading >
+                                                  | detail::take_until_or_throw(is_cntrl || is_blank) // read ID until delimiter…
+                                                  | views::char_to<std::ranges::range_value_t<id_type>>,
+                                      std::cpp20::back_inserter(id));                               // … ^A is old delimiter
+                }
+                else
+                {
+                    std::ranges::copy(stream_view | std::views::drop_while(is_id)
+                                                  | detail::take_until_or_throw(is_cntrl || is_blank) // read ID until delimiter…
+                                                  | views::char_to<std::ranges::range_value_t<id_type>>,
+                                      std::cpp20::back_inserter(id));                               // … ^A is old delimiter
 
-                std::ranges::copy(stream_view | std::views::drop_while(is_id || is_blank)        // skip leading >
-                                              | detail::take_until_or_throw(is_cntrl || is_blank) // read ID until delimiter…
-                                              | views::char_to<std::ranges::range_value_t<id_type>>,
-                                  std::cpp20::back_inserter(id));                               // … ^A is old delimiter
+                }
 
                 // consume rest of line
                 detail::consume(stream_view | detail::take_line_or_throw);
@@ -231,8 +249,16 @@ class format_fasta
             #if SEQAN3_WORKAROUND_VIEW_PERFORMANCE
                 auto it = stream_view.begin();
                 auto e = stream_view.end();
-                for (; (it != e) && (is_id || is_blank)(*it); ++it)
-                {}
+                if (options.fasta_ignore_blanks_before_id)
+                {
+                    for (; (it != e) && (is_id || is_blank)(*it); ++it)
+                    {}
+                }
+                else
+                {
+                    for (; (it != e) && (is_id)(*it); ++it)
+                    {}
+                }
 
                 bool at_delimiter = false;
                 for (; it != e; ++it)
@@ -249,11 +275,21 @@ class format_fasta
                     throw unexpected_end_of_input{"FASTA ID line did not end in newline."};
 
             #else // ↑↑↑ WORKAROUND | ORIGINAL ↓↓↓
+                if (options.fasta_ignore_blanks_before_id)
+                {
+                    std::ranges::copy(stream_view | detail::take_line_or_throw                   // read line
+                                                  | std::views::drop_while(is_id || is_blank)    // skip leading >
+                                                  | views::char_to<std::ranges::range_value_t<id_type>>,
+                                      std::cpp20::back_inserter(id));
+                }
+                else
+                {
+                    std::ranges::copy(stream_view | detail::take_line_or_throw                   // read line
+                                                  | std::views::drop_while(is_id)                // skip leading >
+                                                  | views::char_to<std::ranges::range_value_t<id_type>>,
+                                      std::cpp20::back_inserter(id));
 
-                std::ranges::copy(stream_view | detail::take_line_or_throw                   // read line
-                                              | std::views::drop_while(is_id || is_blank)    // skip leading >
-                                              | views::char_to<std::ranges::range_value_t<id_type>>,
-                                  std::cpp20::back_inserter(id));
+                }
             #endif // SEQAN3_WORKAROUND_VIEW_PERFORMANCE
             }
         }

diff --git a/include/seqan3/io/sequence_file/input_options.hpp b/include/seqan3/io/sequence_file/input_options.hpp
@@ -29,6 +29,8 @@ struct sequence_file_input_options
     bool truncate_ids = false;
     //!\brief Read the complete_header into the seqan3::field::id for embl or genbank format.
     bool embl_genbank_complete_header = false;
+    //!\brief Remove spaces after ">" (or ";") before the actual ID.
+    bool fasta_ignore_blanks_before_id = true;
 };
 
 } // namespace seqan3