jnalanko / vomm Goto Github PK

reconstruct.cpp:134:75: error: no matching function for call to 'Global_Data::load_all_from_disk(std::__cxx11::string&, std::__cxx11::string&, bool&, bool)'
G.load_all_from_disk(C.modeldir, C.filename, C.run_length_coding, true);
^
In file included from Parent_Support.hh:10:0,
from reconstruct.cpp:18:
globals.hh:120:10: note: candidate: void Global_Data::load_all_from_disk(std::__cxx11::string, std::__cxx11::string, bool)
void load_all_from_disk(string directory, string filename_prefix, bool load_bibwt) {

could you please also offer a flag with which the user can print info of all contexts? e.g.:

• number of contexts, versus total number of candidate contexts
• string depth of every context
• tree depth of every context
• the score of every context (if it makes sense)

if i don't specify --escapeprob with --lin-scoring i get:

score_string_optimized: score_string.cpp:85: void Scoring_Config::assert_all_ok(): Assertion `escapeprob != -1' failed.

but i don't think the competitors' paper uses escape probability?

i pulled the latest version from git and i started to experiment with the --store-depths flag. but i get the following error in many datasets:

Sat Mar 31 17:54:45 2018 Starting to build the model
Sat Mar 31 17:54:45 2018 Building the BiBWT
Sat Mar 31 17:55:10 2018 Building reverse suffix tree BPR and pruning marks
Sat Mar 31 17:57:03 2018 Pruning marks is all ones
Sat Mar 31 17:58:04 2018 Marking contexts and storing string depths of maxreps
Sat Mar 31 17:59:24 2018 Building the BPR of contexts only
Sat Mar 31 17:59:30 2018 Computing reverse BWT (todo: reuse the bibwt)
Sat Mar 31 17:59:42 2018 Writing model to directory: /tmp/cunial
terminate called after throwing an instance of 'std::runtime_error'
what(): Error writing to disk: /tmp/cunial/protein.txt.rev_bwt_bwt.dat

Don't do "return next();"

if there is no pruning, there is no need for pruning marks and the support structures at all.
The pruning marks will be a bit vector contain only ones. It's still built because the code is simpler
that way, but it should not be a big deal to change that.

this is important since its scoring function if very simple, and it will be probably faster than ours.

since its scoring is so simple, we might also end up noticing that we don't need the context data structures and probably just revST topology and rlbwt to implement it.

last week i was able to solve the problem with the cluster, and i started building the indexes on all bacteria. after five days and 13 hours, one of the indexes is still not built (the one with depth pruning at 8, the smallest i tried), and it seems stuck to "Building the BiBWT".

could you please provide a switch that allows one to reuse parts of an index at depth, say, 16, which i already have, without recomputing everything from scratch?

i think this is really crucial in order to index large files in practice.

I thought that changing the wavelet tree fixed this but nope.

Reproducable with:

./build_model_optimized --reference-raw ../allBacteria/GCF_000325745.1_ASM32574v1_genomic.txt --four-thresholds 0.0001 0.001 0.952 1.05 --outputdir ./models

Buffer input reading and optimize scoring itself.