|
0
|
1 #!/usr/bin/env python
|
|
|
2
|
|
|
3 """
|
|
|
4 xena_import.py: import a dataset into Xena
|
|
|
5
|
|
|
6 Given a cmdline-specified genomic data file and a cmdline-specified Xena
|
|
|
7 directory, import the genomic data fle into Xena. This requires assembling
|
|
|
8 the necessary json file, based on cmdline input.
|
|
|
9 """
|
|
|
10
|
|
|
11 import argparse
|
|
|
12 import json
|
|
|
13 import os
|
|
|
14 import re
|
|
|
15 import shutil
|
|
|
16 import subprocess
|
|
|
17 import sys
|
|
|
18 import traceback
|
|
|
19 import xena_utils as xena
|
|
|
20
|
|
|
21 def updateColNormalization(jsonMetadata):
|
|
|
22 """Set colNormalization to TRUE if the data is of some expression-related
|
|
|
23 subtype and if colNormalization has not been set"""
|
|
|
24 if not jsonMetadata.has_key("colNormalization"):
|
|
|
25 if jsonMetadata.has_key("dataSubType"):
|
|
|
26 if re.search("expression", jsonMetadata['dataSubType'],
|
|
|
27 re.IGNORECASE):
|
|
|
28 jsonMetadata["colNormalization"] = "true"
|
|
|
29
|
|
|
30 def verifyAndImportExistingMetadata(inputJsonPath, outputJsonPath):
|
|
|
31 """Take an existing JSON file. Read the contents, and check for any
|
|
|
32 content that might be invalid in the local installation. Delete that
|
|
|
33 content. Write the rest to the indicated output pathname"""
|
|
|
34 problemFields = [":probeMap"]
|
|
|
35 fileContents = open(inputJsonPath).read()
|
|
|
36 jsonMetadata = json.loads(fileContents)
|
|
|
37 for thisProblem in problemFields:
|
|
|
38 if jsonMetadata.has_key(thisProblem):
|
|
|
39 del jsonMetadata[thisProblem]
|
|
|
40 updateColNormalization(jsonMetadata)
|
|
|
41 fp = open(outputJsonPath, "w")
|
|
|
42 fp.write("%s\n" % json.dumps(jsonMetadata, indent=2))
|
|
|
43 fp.close()
|
|
|
44
|
|
|
45
|
|
|
46 def main():
|
|
|
47 parser = argparse.ArgumentParser()
|
|
|
48 parser.add_argument("genomicDataPath", type=str)
|
|
|
49 parser.add_argument("outfile", type=str)
|
|
|
50 parser.add_argument("--json", type=str, default=None)
|
|
|
51 parser.add_argument("--cohort", type=str)
|
|
|
52 parser.add_argument("--type", type=str)
|
|
|
53 parser.add_argument("--dataSubType", type=str, default=None)
|
|
|
54 parser.add_argument("--label", type=str, default=None)
|
|
|
55 args = parser.parse_args()
|
|
|
56
|
|
|
57 fp2 = open(args.outfile, "w")
|
|
|
58 fp2.write("Importing data to Xena\n")
|
|
|
59 xenaFileDir = xena.fileDir()
|
|
|
60 genomicDataFile = args.genomicDataPath.split("/")[-1]
|
|
|
61 jsonMetadataTargetPathname = "%s/%s.json" % (xenaFileDir,
|
|
|
62 genomicDataFile)
|
|
|
63
|
|
|
64
|
|
|
65 # The metadata either came as the name of a JSON file or a series of
|
|
|
66 # command line arguments.
|
|
|
67 if args.json is not None:
|
|
|
68 # In this case, the metadata came in the form of a JSON file.
|
|
|
69 # Verify that the metadata is valid on the current system, which
|
|
|
70 # might mean altering it. Import the stuff that will validate.
|
|
|
71 verifyAndImportExistingMetadata(args.json, jsonMetadataTargetPathname)
|
|
|
72 else:
|
|
|
73 # In this case, the metadata came in the form of a series of
|
|
|
74 # command line arguments. Assemble them into JSON format,
|
|
|
75 # and write a JSON file into the Xena file directory.
|
|
|
76 metadata = { 'cohort': args.cohort, 'type': args.type }
|
|
|
77 if args.dataSubType is not None:
|
|
|
78 metadata['dataSubType'] = args.dataSubType
|
|
|
79 if args.label is not None:
|
|
|
80 metadata['label'] = args.label
|
|
|
81 jsonMetadata = json.dumps(metadata, indent=2)
|
|
|
82 fp = open(jsonMetadataTargetPathname, "w")
|
|
|
83 fp.write("%s\n" % (jsonMetadata))
|
|
|
84 fp.close()
|
|
|
85
|
|
|
86 # Finally, copy the genomic data into the Xena directory
|
|
|
87 shutil.copy(args.genomicDataPath, xenaFileDir)
|
|
|
88
|
|
|
89 # Set up the xena load comamnd and try to execute it. If an exception
|
|
|
90 # is generated, output a traceback and exit with nonzero status. If
|
|
|
91 # no exception was generated, indicate a successful import and exit
|
|
|
92 # with zero status.
|
|
|
93 xenaLoadCmd = "java -jar %s -l --force %s/%s -p %s" % (xena.jarPath(),
|
|
|
94 xenaFileDir,
|
|
|
95 genomicDataFile,
|
|
|
96 xena.port())
|
|
|
97 try:
|
|
|
98 subprocess.call(xenaLoadCmd, shell=True)
|
|
|
99 except:
|
|
|
100 exc_type, exc_value, exc_traceback = sys.exc_info()
|
|
|
101 lines = traceback.format_exception(exc_type, exc_value, exc_traceback)
|
|
|
102 allLines = ''.join('!! ' + line for line in lines)
|
|
|
103 fp2.write("Unsuccessful: error %s\n" % allLines)
|
|
|
104 fp2.close()
|
|
|
105 sys.exit(-1)
|
|
|
106 else:
|
|
|
107 fp2.write( "Import successful\n")
|
|
|
108 fp2.close()
|
|
|
109 sys.exit(0)
|
|
|
110
|
|
|
111
|
|
|
112 if __name__ == "__main__":
|
|
|
113 main()
|
