Skip to content

Commit

Permalink
Version 0.4
Browse files Browse the repository at this point in the history
  • Loading branch information
jltsiren committed Nov 22, 2015
1 parent d63ec4b commit f3386ec
Showing 1 changed file with 2 additions and 2 deletions.
4 changes: 2 additions & 2 deletions README.md
Original file line number Diff line number Diff line change
Expand Up @@ -6,15 +6,15 @@ This is a reimplementation of the Generalized Compressed Suffix Array (GCSA), a

The input to index construction is a set of paths of length *k* in the input graph. The prefix-doubling algorithm transforms the input into an order-*8k* (order-*2k*, order-*4k*) pruned de Bruijn graph for paths in the input graph. A pruned de Bruijn graph differs from a de Bruijn graph in that its nodes may have shorter labels than the order of the graph, if the shorter labels uniquely determine the start nodes of the corresponding paths in the input graph. As such, pruned de Bruijn graphs are usually smaller than proper de Bruijn graphs.

At the moment, GCSA2 is being developed as a plugin to Erik Garrison's [variation graph tools](https://github.com/ekg/vg). The only implemented construction option is based on extracting *k*-mers from vg. Later, GCSA2 should become a more general graph indexing library.
At the moment, GCSA2 is being developed as a part of Erik Garrison's [variation graph tools](https://github.com/ekg/vg). The only implemented construction option is based on extracting *k*-mers from vg. Later, GCSA2 should become a more general graph indexing library.

See [the wiki](https://github.com/jltsiren/gcsa2/wiki) for further documentation.

## Compilation options

The maximum resident size reported by `getrusage()` is in kilobytes in Linux and in bytes in OS X. By default, the implementation assumes Linux-like behavior. To get the correct memory usage reports in OS X, uncomment the line `RUSAGE_FLAGS=-DRUSAGE_IN_BYTES` in the makefile.

There are some verification/debugging options in `build_gcsa`. To disable them, comment out the line `VERIFY_FLAGS=-DVERIFY_CONSTRUCTION` in the makefile.
There are some verification/debugging options in `build_gcsa`. To disable them, comment out the line `VERIFY_FLAGS=-DVERIFY_CONSTRUCTION` in the makefile. This may reduce the memory usage of index construction significantly (see [the wiki](https://github.com/jltsiren/gcsa2/wiki/Construction-Benchmarks)).

Index construction can be set to output some status information to `stderr` by uncommenting the line `OUTPUT_FLAGS=-DVERBOSE_STATUS_INFO` in the makefile.

Expand Down

0 comments on commit f3386ec

Please sign in to comment.