Closed henrikstranneheim closed 11 years ago
Hej Henrik,
Please 'git pull' it first. The reason that you still get this stuff is that you are using the old binary file.
You have to 'make clean' and 'make all' after pulling.
And be advised... The new binary file doesn't support bloom file list any more. Roman thought we should keep it simple and use shell script to do the multiple processing job. But we can discuss a little bit if you want. I still keep the old list supporting version
[henriks@kalkyl1 DRASS]$ make clean rm -f core.* _.o bloom_build simple_check simple_remove [henriks@kalkyl1 DRASS]$ git pull Already up-to-date. [henriks@kalkyl1 DRASS]$ make all Make sure you have MPI support on your cluster hint: module load openmpi mpicc -c .c -O3 -D_FILE_OFFSET_BITS=64 -D_LARGE_FILE -fopenmp bloom.c: I funktion "help": bloom.c:562: fel: expected ";" before "ptions" bloom.c:562: fel: överblivet "\" i program bloom.c:562:10: varning: avslutande "-tecken saknas bloom.c:562: fel: avslutande "-tecken saknas bloom.c: I funktion "build_help": bloom.c:596: fel: expected ";" before "printf" make: _* [all] Fel 1
Henrik Stranneheim Ph.D. Department of Molecular Medicine and Surgery Karolinska Institute Science for Life Laboratory, KISP SE-171 65 Solna, Sweden
E-mail: henrik.stranneheim@scilifelab.se Phone: +46 (0)8 524 81487 (Office) Phone: +46 (0)736251487 (Mobile) Visiting address: Tometebodavägen 23A
On 7 sep 2012, at 10:09, tzcoolman wrote:
Hej Henrik,
Please 'git pull' it first. The reason that you still get this stuff is that you are using the old binary file.
You have to 'make clean' and 'make all' after pulling.
— Reply to this email directly or view it on GitHub.
That will make it a lot harder to use for the general public.
Not everyone knows shell-scripting but anyone can compile a list if the format is well described.
Henrik Stranneheim Ph.D. Department of Molecular Medicine and Surgery Karolinska Institute Science for Life Laboratory, KISP SE-171 65 Solna, Sweden
E-mail: henrik.stranneheim@scilifelab.se Phone: +46 (0)8 524 81487 (Office) Phone: +46 (0)736251487 (Mobile) Visiting address: Tometebodavägen 23A
On 7 sep 2012, at 10:13, tzcoolman wrote:
And be advised... The new binary file doesn't support bloom file list any more. Roman thought we should keep it simple and use shell script to do the multiple processing job. But we can discuss a little bit if you want. I still keep the old list supporting version
— Reply to this email directly or view it on GitHub.
Must be some errors on GITHUB. I updated something yesterday and GITHUB didnt get it. I have pushed again.. Plz go pull.
:-/
everybody got their point. Maybe I should add a duplicated list supported version?
Let me talk to lars 1st.
Shit.... Happened again??
I ll come to the lab when I finish FIFO supporting.
Enze
Alright. I got the updated version working.
There is no flag for supplying the query data set in simple_remove.
In simple_remove: If you supply: -o analysis/henriks/Sim_HC/21_0.005/Homo_sapiens_51511750_NC_000021.7_chr21
you get match->analysis/henriks/Sim_HC/21_0.005/Homo_sapiens_51511750_NC_000021.7_chr21SimHC-454.7b8911bf_contam_Homo_sapiens_51511750_NC_000021.7_chr21.fasta mis->analysis/henriks/Sim_HC/21_0.005/Homo_sapiens_51511750_NC_000021.7_chr21SimHC-454.7b8911bf_clean_Homo_sapiens_51511750_NC_000021.7_chr21.fasta
whcih does not look very nice.
Okay... I still have the list supported version. So if everyone all agrees with the list supported feature, I can reverse my program a little bit. So dont worry. And For the 'ugly' match/mis filename, I can change it if you have any good idea.
Enze
I think that the default behavior should be to take input files on the command line, this is standard Unix practice, and it would be nice to have Henrik's input list (file containing filenames, one per line?) as alternative input feature. Internally, it is not hard to move command line filenames OR filenames read from a file into a simple list which is then iterated over.
Lasse
On Sep 7, 2012, at 13:59 , tzcoolman wrote:
Okay... I still have the list supported version. So if everyone all agrees with the list supported feature, I can reverse my program a little bit. So dont worry. And For the 'ugly' match/mis filename, I can change it if you have any good idea.
Enze
— Reply to this email directly or view it on GitHub.
I think we should support file, dir , or list as input. then it offers a lot of flexibility and its up to the user what to use.
Example: -o analysis/henriks/Sim_HC/21_0.005/Homo_sapiens_51511750_NC_000021.7_chr21
yields: analysis/henriks/Sim_HC/21_0.005/Homo_sapiens_51511750_NC_000021.7_chr21_SimHC-454.7b8911bf_clean.fasta (it should be the same .ending as the query data set infile ending)
Okay...I ll try to make it both file supported and list supported.
Excellent!
This issue should be closed in favor of pull request #29. I cannot close it, Enze, could you verify that the requested functionality is in the pull request and close it?
Thanks!
Hi,
USAGE
For Bloom build:
./simple_check -m 1 -k 21 -e 0.005 -l file_list -p test/
Arguments:
-m mode (default 1)
1--> build one filter for each file in the list
2--> build one filter for every file in the list
-k k_mer size (default 1)
-e error rate (default 0.0005)
-l list containing all references files. One file per line
-p prefix (default same path as the script)