Mercurial > repos > greg > bam_to_scidx
comparison bam_to_scidx.py @ 0:962b0366a980 draft
Uploaded
| author | greg |
|---|---|
| date | Tue, 24 Nov 2015 08:13:56 -0500 |
| parents | |
| children |
comparison
equal
deleted
inserted
replaced
| -1:000000000000 | 0:962b0366a980 |
|---|---|
| 1 """ | |
| 2 bam_to_scidx.py | |
| 3 | |
| 4 Input: BAM file | |
| 5 Output: Converted scidx file | |
| 6 """ | |
| 7 import optparse | |
| 8 import os | |
| 9 import shutil | |
| 10 import subprocess | |
| 11 import sys | |
| 12 import tempfile | |
| 13 | |
| 14 BUFF_SIZE = 1048576 | |
| 15 | |
| 16 | |
| 17 def get_stderr_exception(tmp_err, tmp_stderr): | |
| 18 """ | |
| 19 Return a stderr string of reasonable size. | |
| 20 """ | |
| 21 tmp_stderr.close() | |
| 22 # Get stderr, allowing for case where it's very large. | |
| 23 tmp_stderr = open(tmp_err, 'rb') | |
| 24 stderr_str = '' | |
| 25 buffsize = BUFF_SIZE | |
| 26 try: | |
| 27 while True: | |
| 28 stderr_str += tmp_stderr.read(buffsize) | |
| 29 if not stderr_str or len(stderr_str) % buffsize != 0: | |
| 30 break | |
| 31 except OverflowError: | |
| 32 pass | |
| 33 tmp_stderr.close() | |
| 34 return stderr_str | |
| 35 | |
| 36 | |
| 37 def stop_err(msg): | |
| 38 sys.stderr.write(msg) | |
| 39 sys.exit(1) | |
| 40 | |
| 41 | |
| 42 parser = optparse.OptionParser() | |
| 43 parser.add_option('-j', '--jar_file', dest='jar_file', type='string', help='BAMtoIDX.jar') | |
| 44 parser.add_option('-b', '--input_bam', dest='input_bam', type='string', help='Input dataset in BAM format') | |
| 45 parser.add_option('-i', '--input_bai', dest='input_bai', type='string', help='Input dataset index') | |
| 46 parser.add_option('-r', '--read', dest='read', type='int', default=0, help='Reads.') | |
| 47 parser.add_option('-o', '--output', dest='output', type='string', help='Output dataset in scidx format') | |
| 48 options, args = parser.parse_args() | |
| 49 | |
| 50 tmp_dir = tempfile.mkdtemp(prefix='tmp-scidx-') | |
| 51 tmp_out = tempfile.NamedTemporaryFile().name | |
| 52 tmp_stdout = open(tmp_out, 'wb') | |
| 53 tmp_err = tempfile.NamedTemporaryFile().name | |
| 54 tmp_stderr = open(tmp_err, 'wb') | |
| 55 | |
| 56 # Link input BAM file and associated index file into the working directory. | |
| 57 input_data_file_name = "input.bam" | |
| 58 os.link(options.input_bam, os.path.join(tmp_dir, input_data_file_name)) | |
| 59 os.link(options.input_bai, os.path.join(tmp_dir, "%s.bai" % input_data_file_name)) | |
| 60 | |
| 61 cmd = "java -jar %s -b %s -s %d -o %s" % (options.jar_file, input_data_file_name, options.read, options.output) | |
| 62 proc = subprocess.Popen(args=cmd, stdout=tmp_stdout, stderr=tmp_stderr, shell=True, cwd=tmp_dir) | |
| 63 return_code = proc.wait() | |
| 64 if return_code != 0: | |
| 65 error_message = get_stderr_exception(tmp_err, tmp_stderr) | |
| 66 stop_err(error_message) | |
| 67 | |
| 68 if tmp_dir and os.path.exists(tmp_dir): | |
| 69 shutil.rmtree(tmp_dir) |
