More instructions to build using nim/nimble
Opened this issue · 5 comments
If you get a chance to include more instructions on how to build using nim/nimble as there are no instructions and would like to know in detail how to get it to recognize the "import hts" when compiling src/indelope.nim
Probably get hit two birds/issues and happy to provide a binary that will work with libc-2.12.so if you provide more details to build. Thanks!
Hi Monkol. I am happy to have interest in the tool. That would be great if you could provide a binary.
You can follow the recipe in mosdepth, that bioconda uses: https://github.com/brentp/mosdepth/blob/master/scripts/install.sh
after you follow those instructions, you can use nim c -d:release src/indelope to get a binary for you system.
For now, I am focusing on the accuracy. As you can see from the plot in the README, there is some work to go on that front--though it's unclear how well I can rely on the GIAB truthset for large indels.
I will update the README with more cautionary information and a hopeful timeline.
Thanks for the fast response and tip. I had an old version of nim/nimble install but installing 0.17.3 from source. I'm still getting compile errors and will open two issues.
I've created a binary for older systems based on libc-2.12.so but what a huge pain in the ass! I'm gonna rant, though it has nothing to do with indelope but thought it would be helpful to others and I'm waiting for Walking Dead to start on AMC. 🍣😂
So the compute nodes that I use at Broad Institute are so damn old I had to install other libraries to get this to work. The libpcre library was missing so compiled from source
https://sourceforge.net/projects/pcre/files/pcre/8.41
The new hts library because of cram support required two other compression libraries libbz2 and liblzma which I compiled from source
https://github.com/kobolabs/liblzma
http://www.bzip.org/1.0.6/bzip2-1.0.6.tar.gz
liblzma was a bastard to get source code that had proper make files within them so lucky to stumble upon the repo!
I LD_LIBRARY_PATH those babies up and the binary works and tested on a small hg19 WGS bam. I'm yet to test on our hg38 WGS cram files.
Blah. Thanks for persisting. It should not be this hard.
Thankfully, bioconda makes a lot of this stuff less of a PITA.
I'll also see if I can get a docker container based on an old centOS to build a static binary once I have worked more on the accuracy of this tool.
Now that you've made it this far, I'd be interested to hear any feedback on the tool itself. Its current intended use is on exomes, but I've run it on a few whole genomes and it works fine AFAICT.