Auxiliary methods for taxonomic classification #343
Conversation
Signed-off-by: Radu Muntean <radu.muntean.2@gmail.com>
| throw std::runtime_error( | ||
| folly::to<string>("The current 'call_annotated_rows' call has returned, ", unique_matrix_rows.size(), | ||
| "rows. The maximum number of rows that can be returned is ", | ||
| std::numeric_limits<uint32_t>::max(), | ||
| ". Please reduce the query batch size")); |
| if (prefix.size() > str.size()) { | ||
| return false; | ||
| } | ||
| return prefix == str.substr(0, static_cast<int>(prefix.size())); |
There was a problem hiding this comment.
| return prefix == str.substr(0, static_cast<int>(prefix.size())); | |
| return prefix == std::string_view(str).substr(0, prefix.size()); |
| void AnnotatedDBG::call_annotated_rows(const std::vector<node_index> &rows, | ||
| std::function<void(const std::string&)> callback_cell, | ||
| std::function<void()> callback_row) const { | ||
| assert(check_compatibility()); | ||
|
|
||
| auto unique_matrix_rows = annotator_->get_matrix().get_rows(rows); | ||
|
|
||
| //TODO make sure that this function works even if we have duplications in 'rows'. Then, delete this error catch. | ||
| if (rows.size() != unique_matrix_rows.size()) { | ||
| throw std::runtime_error("The current 'call_annotated_rows' call contains duplication."); | ||
| } | ||
|
|
||
| if (unique_matrix_rows.size() >= std::numeric_limits<uint32_t>::max()) { | ||
| throw std::runtime_error( | ||
| folly::to<string>("The current 'call_annotated_rows' call has returned, ", unique_matrix_rows.size(), | ||
| "rows. The maximum number of rows that can be returned is ", | ||
| std::numeric_limits<uint32_t>::max(), | ||
| ". Please reduce the query batch size")); | ||
| } | ||
| const auto &label_encoder = annotator_->get_label_encoder(); | ||
| for (auto row : unique_matrix_rows) { | ||
| for (auto cell : row) { | ||
| callback_cell(label_encoder.decode(cell)); | ||
| } | ||
| callback_row(); | ||
| } | ||
| } |
There was a problem hiding this comment.
I don't quite see a reason to add this function. You have access to all members, so can implement this directly where needed.
There was a problem hiding this comment.
thanks! answered in the other comment
| ". Please reduce the query batch size")); | ||
| } | ||
| const auto &label_encoder = annotator_->get_label_encoder(); | ||
| for (auto row : unique_matrix_rows) { |
There was a problem hiding this comment.
this is unnecessarily making a copy
| for (auto row : unique_matrix_rows) { | |
| for (const auto &row : unique_matrix_rows) { |
| @@ -156,6 +156,10 @@ class AnnotatedDBG : public AnnotatedSequenceGraph { | |||
| int32_t match_score = 1, | |||
| int32_t mismatch_score = 2) const; | |||
|
|
|||
| void call_annotated_rows(const std::vector<node_index> &rows, | |||
| std::function<void(const std::string&)> callback_cell, | |||
| std::function<void()> callback_row) const; | |||
There was a problem hiding this comment.
What's the second callback supposed to be used for? It's not clear...
There was a problem hiding this comment.
I suggest to remove this function
There was a problem hiding this comment.
Thanks for the comments! I need to have access to annotator_ field and I can't implement it directly where needed.
There was a problem hiding this comment.
It is using 2 callback functions: one for doing a preprocessing per cell (i.e. get the taxid from label) and one for a final computation on an annotated row (i.e. get the LCA taxid). If you think that it would be more clear, I can reimplement it to use only one callback that receives the list of labels on one row (but this will use more memory -> store list of labels instead of list of integers)
There was a problem hiding this comment.
But do you need to have a non-const reference to the annotator? If not, you can just call get_annotator().get_matrix()
There was a problem hiding this comment.
Cool, I missed that get_annotator(). Did the changes now, thanks for the feedback
heracle
force-pushed
the
mtg_taxclass_1
branch
2 times, most recently
from
78ad12f
to
b92912e
Compare
Add some small methods that will be further used for tax class
Signed-off-by: Radu Muntean radu.muntean.2@gmail.com