CRAB/python/JobDB.py

from WorkSpace import WorkSpace
from crab_exceptions import *
import common

import os, string

class dbEntry:
    def __init__(self):
        self.status = 'X'       # job status
        self.exitStatus = ''    # job exit status
        self.jid = ''           # scheduler job id
        self.bossid = ''        # BOSS job id
        self.collections = []   # EvCollection to be analyzed in this job
        self.inputSandbox = []  # InputSandbox
        self.outputSandbox = [] # OutputSandbox
        self.taskId = ''        # Task job belongs to
        self.arguments = []    ### Fabio: abstract job_type parameters
        return

    def __str__(self):
        txt  = 'Status <' + self.status + '>; '
        if self.exitStatus!='':
            txt += 'exitStatus <' + str(self.exitStatus) + '>\n'
        txt += 'Job Id <' + self.jid + '>\n'
        if self.arguments:
            txt += 'Job Type Arguments <' + str(self.arguments) + '>\n'
        if self.collections:
            txt += 'Collections <' + str(self.collections) + '>\n'

        return txt

class JobDB:
    def __init__(self):
        self._dir = common.work_space.shareDir() + 'db/'
        self._db_fname = 'jobs'
        self._jobs = []        # list of dbEntry's
        return

    def __str__(self):
        njobs = self.nJobs()
        if njobs == 1: plural = ''
        else:          plural = 's'
        txt = 'Total of %d job%s:\n' % (njobs, plural)
        for i in range(njobs):
            txt += ('Job %03d' % (i+1)) + ': '
            txt += str(self._jobs[i])
            pass
        return txt

    def dump(self, jobs):
        njobs = len(jobs)
        if njobs == 1: plural = ''
        else:          plural = 's'
        print 'Listing %d job%s:\n' % (njobs, plural)
        for job in jobs:
            print string.strip(('Job %03d' % (job)) + ': ' + str(self._jobs[job - 1]))
            pass

    def nJobs(self):
        return len(self._jobs)

    def create(self, njobs):

        if os.path.exists(self._dir):
            msg = 'Cannot create Job DB: already exists.'
            raise CrabException(msg)

        os.mkdir(self._dir)

        for i in range(njobs):
            self._jobs.append(dbEntry())
            pass

        common.logger.debug(5,'Created DB for '+str(njobs)+' jobs')

        self.save()
        return

    def save(self):
        db_file = open(self._dir+self._db_fname, 'w')
        for i in range(len(self._jobs)):
            db_file.write(`(i+1)`+';')
            db_file.write(self._jobs[i].status+';')
            db_file.write(self._jobs[i].exitStatus+';')
            db_file.write(self._jobs[i].jid+';')
            db_file.write(self._jobs[i].bossid+';')
            db_file.write(str(self._jobs[i].collections)+';')
            db_file.write(str(self._jobs[i].inputSandbox)+';')
            db_file.write(str(self._jobs[i].outputSandbox)+';')
            db_file.write(str(self._jobs[i].taskId)+';')
            db_file.write(str(self._jobs[i].arguments)+';')
            db_file.write('\n')
            pass
        db_file.close()
        return

    def load(self):
        self._jobs = []
        try:
            db_file = open(self._dir+self._db_fname, 'r')
        except IOError:
            raise DBException("Something really serious! no JobDB is present!!!")

        for line in db_file:
            db_entry = dbEntry()
            (n, db_entry.status, db_entry.exitStatus, db_entry.jid, db_entry.bossid, collectionsTMP,  inputSandboxTMP , outputSandboxTMP , db_entry.taskId, argumentsTMP, rest) = string.split(line, ';')
            db_entry.collections = self.strToList_(collectionsTMP)
            db_entry.inputSandbox = self.strToList_(inputSandboxTMP)
            db_entry.outputSandbox = self.strToList_(outputSandboxTMP)
            db_entry.arguments = self.strToList_(argumentsTMP)
            self._jobs.append(db_entry)
            pass
        db_file.close()
        return
    
    def strToList_(self, list):
        if list[0] == '[' and list[-1] == ']' :
            return string.split(string.replace(list[1:-1],"'",""),',')
        else :
            return list
        
    
    def setStatus(self, nj, status):
        self._jobs[int(nj)].status = status
        return
    
    def status(self, nj):
        return self._jobs[int(nj)].status
    
    def setExitStatus(self, nj, exitStatus):
        self._jobs[int(nj)].exitStatus = exitStatus
        return
    
    def exitStatus(self, nj):
        return self._jobs[int(nj)].exitStatus
    
    def setJobId(self, nj, jid):
        self._jobs[int(nj)].jid = jid
        return
    
    def jobId(self, nj):
        return self._jobs[int(nj)].jid

    def setBossId(self, nj, bossid):
        self._jobs[int(nj)].bossid = bossid
        return
    
    def bossId(self, nj):
        return self._jobs[int(nj)].bossid

    def setArguments(self, nj, args):
        self._jobs[int(nj)].arguments = args
        return
    
    def arguments(self, nj):
        return self._jobs[int(nj)].arguments
    

    def setCollections(self, nj, Collections):
        self._jobs[int(nj)].Collections = Collections
        return
    
    def collections(self, nj):
        return self._jobs[int(nj)].collections

    def setInputSandbox(self, nj, InputSandbox):
        self._jobs[int(nj)].inputSandbox = InputSandbox
        return
    
    def inputSandbox(self, nj):
        return self._jobs[int(nj)].inputSandbox

    def setOutputSandbox(self, nj, OutputSandbox):
        self._jobs[int(nj)].outputSandbox = OutputSandbox
        return
    
    def outputSandbox(self, nj):
        return self._jobs[int(nj)].outputSandbox
        
    def setTaskId(self, nj, taskId):
        self._jobs[int(nj)].taskId = taskId
Revision:	1.14
Committed:	Wed Jun 28 00:36:52 2006 UTC (18 years, 10 months ago) by gutsche
Content type:	text/x-python
Branch:	MAIN
Changes since 1.13:	+5 -2 lines
Log Message:	fixed JobDB arguments extraction for CMSSW jobs, no doubling of backslashes
#	User	Rev	Content
1	nsmirnov	1.1	from WorkSpace import WorkSpace
2			from crab_exceptions import *
3			import common
4
5	nsmirnov	1.2	import os, string
6	nsmirnov	1.1
7			class dbEntry:
8			def __init__(self):
9	slacapra	1.5	self.status = 'X' # job status
10	slacapra	1.8	self.exitStatus = '' # job exit status
11	slacapra	1.5	self.jid = '' # scheduler job id
12	corvo	1.10	self.bossid = '' # BOSS job id
13	slacapra	1.5	self.collections = [] # EvCollection to be analyzed in this job
14			self.inputSandbox = [] # InputSandbox
15			self.outputSandbox = [] # OutputSandbox
16	corvo	1.10	self.taskId = '' # Task job belongs to
17	gutsche	1.13	self.arguments = [] ### Fabio: abstract job_type parameters
18	nsmirnov	1.1	return
19
20	nsmirnov	1.2	def __str__(self):
21			txt = 'Status <' + self.status + '>; '
22	slacapra	1.8	if self.exitStatus!='':
23			txt += 'exitStatus <' + str(self.exitStatus) + '>\n'
24	nsmirnov	1.2	txt += 'Job Id <' + self.jid + '>\n'
25	gutsche	1.13	if self.arguments:
26			txt += 'Job Type Arguments <' + str(self.arguments) + '>\n'
27	slacapra	1.8	if self.collections:
28	slacapra	1.5	txt += 'Collections <' + str(self.collections) + '>\n'
29	slacapra	1.4
30	nsmirnov	1.1	return txt
31
32			class JobDB:
33			def __init__(self):
34			self._dir = common.work_space.shareDir() + 'db/'
35			self._db_fname = 'jobs'
36			self._jobs = [] # list of dbEntry's
37			return
38
39	nsmirnov	1.2	def __str__(self):
40	nsmirnov	1.1	njobs = self.nJobs()
41			if njobs == 1: plural = ''
42			else: plural = 's'
43			txt = 'Total of %d job%s:\n' % (njobs, plural)
44			for i in range(njobs):
45	nsmirnov	1.3	txt += ('Job %03d' % (i+1)) + ': '
46	nsmirnov	1.1	txt += str(self._jobs[i])
47			pass
48			return txt
49
50	slacapra	1.4	def dump(self, jobs):
51			njobs = len(jobs)
52			if njobs == 1: plural = ''
53			else: plural = 's'
54			print 'Listing %d job%s:\n' % (njobs, plural)
55			for job in jobs:
56	spiga	1.12	print string.strip(('Job %03d' % (job)) + ': ' + str(self._jobs[job - 1]))
57	slacapra	1.4	pass
58
59	nsmirnov	1.1	def nJobs(self):
60			return len(self._jobs)
61
62			def create(self, njobs):
63
64			if os.path.exists(self._dir):
65			msg = 'Cannot create Job DB: already exists.'
66			raise CrabException(msg)
67
68			os.mkdir(self._dir)
69
70			for i in range(njobs):
71			self._jobs.append(dbEntry())
72			pass
73
74	slacapra	1.4	common.logger.debug(5,'Created DB for '+str(njobs)+' jobs')
75
76	nsmirnov	1.1	self.save()
77			return
78
79			def save(self):
80			db_file = open(self._dir+self._db_fname, 'w')
81			for i in range(len(self._jobs)):
82			db_file.write(`(i+1)`+';')
83			db_file.write(self._jobs[i].status+';')
84	slacapra	1.8	db_file.write(self._jobs[i].exitStatus+';')
85	nsmirnov	1.1	db_file.write(self._jobs[i].jid+';')
86	fanzago	1.9	db_file.write(self._jobs[i].bossid+';')
87	slacapra	1.4	db_file.write(str(self._jobs[i].collections)+';')
88	slacapra	1.5	db_file.write(str(self._jobs[i].inputSandbox)+';')
89			db_file.write(str(self._jobs[i].outputSandbox)+';')
90	corvo	1.10	db_file.write(str(self._jobs[i].taskId)+';')
91	gutsche	1.13	db_file.write(str(self._jobs[i].arguments)+';')
92	nsmirnov	1.1	db_file.write('\n')
93			pass
94			db_file.close()
95			return
96
97			def load(self):
98			self._jobs = []
99	slacapra	1.6	try:
100			db_file = open(self._dir+self._db_fname, 'r')
101			except IOError:
102	slacapra	1.7	raise DBException("Something really serious! no JobDB is present!!!")
103	slacapra	1.6
104	nsmirnov	1.1	for line in db_file:
105	nsmirnov	1.2	db_entry = dbEntry()
106	gutsche	1.13	(n, db_entry.status, db_entry.exitStatus, db_entry.jid, db_entry.bossid, collectionsTMP, inputSandboxTMP , outputSandboxTMP , db_entry.taskId, argumentsTMP, rest) = string.split(line, ';')
107	slacapra	1.5	db_entry.collections = self.strToList_(collectionsTMP)
108			db_entry.inputSandbox = self.strToList_(inputSandboxTMP)
109			db_entry.outputSandbox = self.strToList_(outputSandboxTMP)
110	gutsche	1.13	db_entry.arguments = self.strToList_(argumentsTMP)
111	nsmirnov	1.1	self._jobs.append(db_entry)
112			pass
113			db_file.close()
114			return
115
116	slacapra	1.5	def strToList_(self, list):
117	gutsche	1.14	if list[0] == '[' and list[-1] == ']' :
118			return string.split(string.replace(list[1:-1],"'",""),',')
119			else :
120			return list
121
122	slacapra	1.5
123	nsmirnov	1.1	def setStatus(self, nj, status):
124	gutsche	1.11	self._jobs[int(nj)].status = status
125	nsmirnov	1.1	return
126
127			def status(self, nj):
128	gutsche	1.11	return self._jobs[int(nj)].status
129	nsmirnov	1.1
130	slacapra	1.8	def setExitStatus(self, nj, exitStatus):
131	gutsche	1.11	self._jobs[int(nj)].exitStatus = exitStatus
132	slacapra	1.8	return
133
134			def exitStatus(self, nj):
135	gutsche	1.11	return self._jobs[int(nj)].exitStatus
136	slacapra	1.8
137	nsmirnov	1.1	def setJobId(self, nj, jid):
138	gutsche	1.11	self._jobs[int(nj)].jid = jid
139	nsmirnov	1.1	return
140
141			def jobId(self, nj):
142	gutsche	1.11	return self._jobs[int(nj)].jid
143	slacapra	1.4
144	fanzago	1.9	def setBossId(self, nj, bossid):
145	gutsche	1.11	self._jobs[int(nj)].bossid = bossid
146	fanzago	1.9	return
147
148			def bossId(self, nj):
149	gutsche	1.11	return self._jobs[int(nj)].bossid
150	fanzago	1.9
151	gutsche	1.13	def setArguments(self, nj, args):
152			self._jobs[int(nj)].arguments = args
153	slacapra	1.4	return
154
155	gutsche	1.13	def arguments(self, nj):
156			return self._jobs[int(nj)].arguments
157	slacapra	1.4
158
159			def setCollections(self, nj, Collections):
160	gutsche	1.11	self._jobs[int(nj)].Collections = Collections
161	slacapra	1.4	return
162
163			def collections(self, nj):
164	gutsche	1.11	return self._jobs[int(nj)].collections
165	slacapra	1.5
166			def setInputSandbox(self, nj, InputSandbox):
167	gutsche	1.11	self._jobs[int(nj)].inputSandbox = InputSandbox
168	slacapra	1.5	return
169
170			def inputSandbox(self, nj):
171	gutsche	1.11	return self._jobs[int(nj)].inputSandbox
172	slacapra	1.5
173			def setOutputSandbox(self, nj, OutputSandbox):
174	gutsche	1.11	self._jobs[int(nj)].outputSandbox = OutputSandbox
175	slacapra	1.5	return
176
177			def outputSandbox(self, nj):
178	gutsche	1.11	return self._jobs[int(nj)].outputSandbox
179	corvo	1.10
180			def setTaskId(self, nj, taskId):
181	gutsche	1.11	self._jobs[int(nj)].taskId = taskId