Mercurial > repos > peterjc > blast2go

comparison tools/blast2go/blast2go.py @ 9:887adf823bc0 draft

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
v0.0.10 - Python 3 compatiblity etc (overdue upload)
author peterjc
date Tue, 06 Dec 2022 16:26:16 +0000
parents e23b621eb7bb
children 8664c4c94764
comparison
equal deleted inserted replaced
8:e23b621eb7bb 9:887adf823bc0
28 """ 28 """
29 import sys 29 import sys
30 import os 30 import os
31 import subprocess 31 import subprocess
32 32
33 #You may need to edit this to match your local setup, 33 # You may need to edit this to match your local setup,
34 blast2go_dir = os.environ.get("B2G4PIPE", "/opt/b2g4pipe_v2.5/") 34 blast2go_dir = os.environ.get("B2G4PIPE", "/opt/b2g4pipe_v2.5/")
35 blast2go_jar = os.path.join(blast2go_dir, "blast2go.jar") 35 blast2go_jar = os.path.join(blast2go_dir, "blast2go.jar")
36 36
37 def stop_err(msg, error_level=1):
38 """Print error message to stdout and quit with given error level."""
39 sys.stderr.write("%s\n" % msg)
40 sys.exit(error_level)
41 37
42 try: 38 try:
43 from massage_xml_for_blast2go import prepare_xml 39 from massage_xml_for_blast2go import prepare_xml
44 except ImportError: 40 except ImportError:
45 stop_err("Missing sister file massage_xml_for_blast2go.py") 41 sys.exit("Missing sister file massage_xml_for_blast2go.py")
46 42
47 if len(sys.argv) != 4: 43 if len(sys.argv) != 4:
48 stop_err("Require three arguments: XML filename, properties filename, output tabular filename") 44 sys.exit("Require three arguments: XML filename, properties filename, output tabular filename")
49 45
50 xml_file, prop_file, tabular_file = sys.argv[1:] 46 xml_file, prop_file, tabular_file = sys.argv[1:]
51 47
52 #We should have write access here: 48 # We should have write access here:
53 tmp_xml_file = tabular_file + ".tmp.xml" 49 tmp_xml_file = tabular_file + ".tmp.xml"
54 50
55 if not os.path.isfile(blast2go_jar): 51 if not os.path.isfile(blast2go_jar):
56 stop_err("Blast2GO JAR file not found: %s" % blast2go_jar) 52 sys.exit("Blast2GO JAR file not found: %s" % blast2go_jar)
57 53
58 if not os.path.isfile(xml_file): 54 if not os.path.isfile(xml_file):
59 stop_err("Input BLAST XML file not found: %s" % xml_file) 55 sys.exit("Input BLAST XML file not found: %s" % xml_file)
60 56
61 if not os.path.isfile(prop_file): 57 if not os.path.isfile(prop_file):
62 tmp = os.path.join(os.path.split(blast2go_jar)[0], prop_file) 58 tmp = os.path.join(os.path.split(blast2go_jar)[0], prop_file)
63 if os.path.isfile(tmp): 59 if os.path.isfile(tmp):
64 #The properties file seems to have been given relative to the JAR 60 # The properties file seems to have been given relative to the JAR
65 prop_file = tmp 61 prop_file = tmp
66 else: 62 else:
67 stop_err("Blast2GO configuration file not found: %s" % prop_file) 63 sys.exit("Blast2GO configuration file not found: %s" % prop_file)
68 del tmp 64 del tmp
69 65
70 66
71 def run(cmd): 67 def run(cmd):
72 #Avoid using shell=True when we call subprocess to ensure if the Python 68 # Avoid using shell=True when we call subprocess to ensure if the Python
73 #script is killed, so too is the child process. 69 # script is killed, so too is the child process.
74 try: 70 try:
75 child = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE) 71 child = subprocess.Popen(cmd, stdout=subprocess.PIPE, stderr=subprocess.PIPE)
76 except Exception, err: 72 except Exception, err:
77 stop_err("Error invoking command:\n%s\n\n%s\n" % (" ".join(cmd), err)) 73 sys.exit("Error invoking command:\n%s\n\n%s\n" % (" ".join(cmd), err))
78 #Use .communicate as can get deadlocks with .wait(),
79 stdout, stderr = child.communicate() 74 stdout, stderr = child.communicate()
80 return_code = child.returncode 75 return_code = child.returncode
81 76
82 #keep stdout minimal as shown prominently in Galaxy 77 # keep stdout minimal as shown prominently in Galaxy
83 #Record it in case a silent error needs diagnosis 78 # Record it in case a silent error needs diagnosis
84 if stdout: 79 if stdout:
85 sys.stderr.write("Standard out:\n%s\n\n" % stdout) 80 sys.stderr.write("Standard out:\n%s\n\n" % stdout)
86 if stderr: 81 if stderr:
87 sys.stderr.write("Standard error:\n%s\n\n" % stderr) 82 sys.stderr.write("Standard error:\n%s\n\n" % stderr)
88 83
89 error_msg = None 84 error_msg = None
90 if return_code: 85 if return_code:
91 cmd_str = " ".join(cmd) 86 cmd_str = " ".join(cmd)
92 error_msg = "Return code %i from command:\n%s" % (return_code, cmd_str) 87 error_msg = "Return code %i from command:\n%s" % (return_code, cmd_str)
93 elif "Database or network connection (timeout) error" in stdout+stderr: 88 elif "Database or network connection (timeout) error" in stdout + stderr:
94 error_msg = "Database or network connection (timeout) error" 89 error_msg = "Database or network connection (timeout) error"
95 elif "Annotation of 0 seqs with 0 annots finished." in stdout+stderr: 90 elif "Annotation of 0 seqs with 0 annots finished." in stdout + stderr:
96 error_msg = "No sequences processed!" 91 error_msg = "No sequences processed!"
97 92
98 if error_msg: 93 if error_msg:
99 print error_msg 94 print error_msg
100 stop_err(error_msg) 95 sys.exit(error_msg)
101 96
102 97
103 blast2go_classpath = os.path.split(blast2go_jar)[0] 98 blast2go_classpath = os.path.split(blast2go_jar)[0]
104 assert os.path.isdir(blast2go_classpath) 99 assert os.path.isdir(blast2go_classpath)
105 blast2go_classpath = "%s/*:%s/ext/*:" % (blast2go_classpath, blast2go_classpath) 100 blast2go_classpath = "%s/*:%s/ext/*:" % (blast2go_classpath, blast2go_classpath)
106 101
107 prepare_xml(xml_file, tmp_xml_file) 102 prepare_xml(xml_file, tmp_xml_file)
108 #print "XML file prepared for Blast2GO" 103 # print "XML file prepared for Blast2GO"
109 104
110 #We will have write access wherever the output should be, 105 # We will have write access wherever the output should be,
111 #so we'll ask Blast2GO to use that as the stem for its output 106 # so we'll ask Blast2GO to use that as the stem for its output
112 #(it will append .annot to the filename) 107 # (it will append .annot to the filename)
113 cmd = ["java", "-cp", blast2go_classpath, "es.blast2go.prog.B2GAnnotPipe", 108 cmd = ["java", "-cp", blast2go_classpath, "es.blast2go.prog.B2GAnnotPipe",
114 "-in", tmp_xml_file, 109 "-in", tmp_xml_file,
115 "-prop", prop_file, 110 "-prop", prop_file,
116 "-out", tabular_file, #Used as base name for output files 111 "-out", tabular_file, # Used as base name for output files
117 "-annot", # Generate *.annot tabular file 112 "-annot", # Generate *.annot tabular file
118 #NOTE: For v2.3.5 must use -a, for v2.5 must use -annot instead 113 # NOTE: For v2.3.5 must use -a, for v2.5 must use -annot instead
119 #"-img", # Generate images, feature not in v2.3.5 114 # "-img", # Generate images, feature not in v2.3.5
120 ] 115 ]
121 #print " ".join(cmd) 116 # print " ".join(cmd)
122 run(cmd) 117 run(cmd)
123 118
124 #Remove the temp XML file 119 # Remove the temp XML file
125 os.remove(tmp_xml_file) 120 os.remove(tmp_xml_file)
126 121
127 out_file = tabular_file + ".annot" 122 out_file = tabular_file + ".annot"
128 if not os.path.isfile(out_file): 123 if not os.path.isfile(out_file):
129 stop_err("ERROR - No output annotation file from Blast2GO") 124 sys.exit("ERROR - No output annotation file from Blast2GO")
130 125
131 #Move the output file where Galaxy expects it to be: 126 # Move the output file where Galaxy expects it to be:
132 os.rename(out_file, tabular_file) 127 os.rename(out_file, tabular_file)
133 128
134 print "Done" 129 print "Done"