diff options
author | Zachary Sloan | 2013-07-19 16:13:47 -0500 |
---|---|---|
committer | Zachary Sloan | 2013-07-19 16:13:47 -0500 |
commit | 6aaefdaae3a9fb068278d9b94d8cdf25d4f8d852 (patch) | |
tree | f002469b1ef94ec5de6e379f1bc4aa0e182a3849 /wqflask/maintenance/get_group_samplelists.py | |
parent | 0fcadee805a6d4a2007e57a69ab130eb9b1c1a3c (diff) | |
download | genenetwork2-6aaefdaae3a9fb068278d9b94d8cdf25d4f8d852.tar.gz |
Created file gen_group_samplelists that iterates through all genofiles
and builds each groups' samplelist
Diffstat (limited to 'wqflask/maintenance/get_group_samplelists.py')
-rw-r--r-- | wqflask/maintenance/get_group_samplelists.py | 43 |
1 files changed, 43 insertions, 0 deletions
diff --git a/wqflask/maintenance/get_group_samplelists.py b/wqflask/maintenance/get_group_samplelists.py new file mode 100644 index 00000000..2434038e --- /dev/null +++ b/wqflask/maintenance/get_group_samplelists.py @@ -0,0 +1,43 @@ +from __future__ import absolute_import, print_function, division + +import os +import glob +import gzip + +from base import webqtlConfig + + +def get_sample_list_dir(geno_dir="/home/zas1024/gene/web/genotypes/"): + os.chdir(geno_dir) + + for group_file in glob.glob("*"): + if group_file.lower().endswith(('.geno', '.geno.gz')): + #group_name = genofilename.split('.')[0] + sample_list = get_sample_list(group_file) + print("\n\n{}\n\n".format(sample_list)) + + +def get_sample_list(group_file): + print(group_file) + genofilename = str(os.path.join(webqtlConfig.GENODIR, group_file)) + if genofilename.lower().endswith('.geno.gz'): + genofile = gzip.open(genofilename) + else: + genofile = open(genofilename) + for line in genofile: + line = line.strip() + if not line: + continue + if line.startswith(("#", "@")): + continue + headline = line + break + headers = headline.split("\t") + if headers[3] == "Mb": + samplelist = headers[4:] + else: + samplelist = headers[3:] + return samplelist + +if __name__ == '__main__': + get_sample_list_dir() |