| 1 |
#!/usr/bin/env python2
|
| 2 |
import sys, os, string, re
|
| 3 |
import urllib, urllister
|
| 4 |
import urllib2
|
| 5 |
import common
|
| 6 |
from RefDBInfo import RefDBInfo
|
| 7 |
from PubDBInfo import *
|
| 8 |
|
| 9 |
# ####################################
|
| 10 |
class PubDBError:
|
| 11 |
def __init__(self, url):
|
| 12 |
print '\nERROR accessing PubDB at '+url+'\n'
|
| 13 |
pass
|
| 14 |
|
| 15 |
# ####################################
|
| 16 |
class PubDBGetAnalysisError:
|
| 17 |
def __init__(self, url,Collections):
|
| 18 |
print '\nERROR extracting info for collections '+Collections+' from PubDB '+url+'.\n'
|
| 19 |
pass
|
| 20 |
|
| 21 |
# ####################################
|
| 22 |
class RefDBmapError:
|
| 23 |
def __init__(self, url):
|
| 24 |
print '\nERROR accessing RefDB-PubDBs map at '+url+'\n'
|
| 25 |
pass
|
| 26 |
|
| 27 |
# ####################################
|
| 28 |
class NoPHPError:
|
| 29 |
def __init__(self, url):
|
| 30 |
#print '\nERROR accessing PHP at '+url+' \n'
|
| 31 |
print 'ERROR accessing PHP: ',url,' \n'
|
| 32 |
pass
|
| 33 |
|
| 34 |
# ####################################
|
| 35 |
class pubDBResult:
|
| 36 |
def __init__(self,
|
| 37 |
contents):
|
| 38 |
self.contents=contents
|
| 39 |
|
| 40 |
|
| 41 |
def dump(self):
|
| 42 |
print 'Contents : ',self.contents
|
| 43 |
pass
|
| 44 |
|
| 45 |
# ####################################
|
| 46 |
# class to access PubDBs
|
| 47 |
class PubDB:
|
| 48 |
def __init__(self, owner, dataset, dataTiers, cfg_params):
|
| 49 |
|
| 50 |
# Attributes
|
| 51 |
self.owner = owner
|
| 52 |
self.dataset = dataset
|
| 53 |
self.dataTiers = dataTiers
|
| 54 |
self.NeededdataTiers=[]
|
| 55 |
self.cfg_params = cfg_params
|
| 56 |
|
| 57 |
self.RefDBurl_ = 'http://cmsdoc.cern.ch/cms/production/www/'
|
| 58 |
self.RefDBphp_ = 'PubDB/GetIdCollection.php'
|
| 59 |
self.RefDBMotherphp_ = 'cgi/SQL/CollectionTree.php'
|
| 60 |
|
| 61 |
self.PubDBCentralUrl_ = 'http://cmsdoc.cern.ch/cms/production/www/PubDB/'
|
| 62 |
self.PubDBCentralPhp_ = 'GetPublishedCollectionInfoFromRefDB.php'
|
| 63 |
|
| 64 |
self.PubDBAnalysisPhp_ = 'get-pubdb-analysisinfo.php'
|
| 65 |
self.PubDBAnalysisPhpOld_ = 'get-pubdb-analysisinfo.php'
|
| 66 |
|
| 67 |
## link to the modified RefDB-PubDBs map script that allow the display option
|
| 68 |
self.RefDBPubDBsmapPhp_ = 'GetPublishedCollectionInfoFromRefDB.mod.php?display=1'
|
| 69 |
|
| 70 |
# Costructor procedures
|
| 71 |
|
| 72 |
CEBlackList = []
|
| 73 |
try:
|
| 74 |
tmpBad = string.split(self.cfg_params['EDG.ce_black_list'],',')
|
| 75 |
#tmpBad = ['fnal']
|
| 76 |
for tmp in tmpBad:
|
| 77 |
tmp=string.strip(tmp)
|
| 78 |
#if (tmp == 'cnaf'): tmp = 'webserver' ########## warning: temp. patch
|
| 79 |
CEBlackList.append(tmp)
|
| 80 |
except KeyError:
|
| 81 |
pass
|
| 82 |
|
| 83 |
CEWhiteList = []
|
| 84 |
try:
|
| 85 |
tmpGood = string.split(self.cfg_params['EDG.ce_white_list'],',')
|
| 86 |
#tmpGood = ['cern']
|
| 87 |
for tmp in tmpGood:
|
| 88 |
tmp=string.strip(tmp)
|
| 89 |
#if (tmp == 'cnaf'): tmp = 'webserver' ########## warning: temp. patch
|
| 90 |
CEWhiteList.append(tmp)
|
| 91 |
except KeyError:
|
| 92 |
pass
|
| 93 |
|
| 94 |
#print 'CEWhiteList: ',CEWhiteList
|
| 95 |
self.reCEWhiteList=[]
|
| 96 |
for Good in CEWhiteList:
|
| 97 |
self.reCEWhiteList.append(re.compile( Good ))
|
| 98 |
#print 'ReGood: ',self.reCEWhiteList
|
| 99 |
|
| 100 |
common.logger.debug(5,'CEBlackList: '+str(CEBlackList))
|
| 101 |
common.logger.debug(5,'CEWhiteList: '+str(CEWhiteList))
|
| 102 |
self.reCEBlackList=[]
|
| 103 |
for bad in CEBlackList:
|
| 104 |
self.reCEBlackList.append(re.compile( bad ))
|
| 105 |
#print 'ReBad: ',self.reCEBlackList
|
| 106 |
|
| 107 |
|
| 108 |
########################################################################
|
| 109 |
def findAllCollections(self):
|
| 110 |
"""
|
| 111 |
Contact RefDB and find the CollID of all the user required collections
|
| 112 |
"""
|
| 113 |
## download from RefDB all the info about the given dataset-owner
|
| 114 |
refdb=RefDBInfo(self.owner,self.dataset)
|
| 115 |
#print refdb.GetRefDBInfo()
|
| 116 |
try:
|
| 117 |
collInfos=refdb.GetRefDBInfo()
|
| 118 |
except :
|
| 119 |
sys.exit(10)
|
| 120 |
#print "collInfos=", collInfos
|
| 121 |
|
| 122 |
first=1
|
| 123 |
NeededCollID=[]
|
| 124 |
refdbdataTiers=[]
|
| 125 |
for coll in collInfos:
|
| 126 |
## select the primary collection
|
| 127 |
if first:
|
| 128 |
NeededCollID.append(coll[0])
|
| 129 |
self.NeededdataTiers.append(coll[2])
|
| 130 |
refdbdataTiers.append(coll[2])
|
| 131 |
common.logger.message("\n --> primary collection for owner "+self.owner+" is: ID="+coll[0]+" DataTier="+coll[2])
|
| 132 |
first=0
|
| 133 |
else:
|
| 134 |
## select only the parents collections corresponding to data-tiers requested by the user
|
| 135 |
if self.dataTiers.count(coll[2]):
|
| 136 |
NeededCollID.append(coll[0])
|
| 137 |
self.NeededdataTiers.append(coll[2])
|
| 138 |
common.logger.message(" --> further collection required: ID="+coll[0]+" DataTier="+coll[2])
|
| 139 |
refdbdataTiers.append(coll[2])
|
| 140 |
|
| 141 |
## check that the user asks for Data Tier really existing in RefDB, otherwise give a warning message
|
| 142 |
for dt in self.dataTiers:
|
| 143 |
if refdbdataTiers.count(dt)<=0:
|
| 144 |
msg = "ERROR: Data Tier ( =>",dt,"<= ) not existing for dataset/owner "+self.dataset+"/"+self.owner+"! "
|
| 145 |
msg = str(msg) + 'Owner Dataset not published with asked dataTiers! '+\
|
| 146 |
self.owner+' '+ self.dataset+' '+str(self.dataTiers)+'\n'
|
| 147 |
msg = str(msg) + ' Check the data_tier variable in crab.cfg !\n'
|
| 148 |
common.logger.message(msg)
|
| 149 |
return []
|
| 150 |
|
| 151 |
#print 'Needed Collections are ', NeededCollID
|
| 152 |
#return collInfos
|
| 153 |
#print "NeededCollID= ", NeededCollID
|
| 154 |
return NeededCollID
|
| 155 |
|
| 156 |
########################################################################
|
| 157 |
def findPubDBsbyCollID(self,CollID):
|
| 158 |
"""
|
| 159 |
Find the list of PubDB URLs having a given Collection
|
| 160 |
"""
|
| 161 |
### contact the RefDB-PubDBs map to discovery where the given CollID is
|
| 162 |
url = self.PubDBCentralUrl_+self.RefDBPubDBsmapPhp_+'&CollID=' + CollID
|
| 163 |
# print "%s"%(url)
|
| 164 |
try:
|
| 165 |
f = urllib.urlopen(url)
|
| 166 |
except IOError:
|
| 167 |
# print 'Cannot access URL: '+url
|
| 168 |
raise RefDBmapError(url)
|
| 169 |
|
| 170 |
### search for the PubDBURL string
|
| 171 |
reURLLine=re.compile( r'PubDBURL=(\S*)' )
|
| 172 |
|
| 173 |
PubDBURLs = []
|
| 174 |
for line in f.readlines():
|
| 175 |
#print '#',line,'#'
|
| 176 |
if reURLLine.search(line) :
|
| 177 |
URLLine=reURLLine.search(line).group()
|
| 178 |
#print string.split(URLLine,'=')[1]
|
| 179 |
PubDBURLs.append(string.split(URLLine,'=')[1])
|
| 180 |
|
| 181 |
### return the list of PubDBURL where the collection is present
|
| 182 |
#return PubDBURLs
|
| 183 |
return self.uniquelist(PubDBURLs)
|
| 184 |
|
| 185 |
################################################################
|
| 186 |
def findPubDBs(self,CollIDs):
|
| 187 |
"""
|
| 188 |
Find the list of PubDB URLs having ALL the required collections
|
| 189 |
"""
|
| 190 |
### loop over all the required collections
|
| 191 |
#pubdbmap={}
|
| 192 |
allurls=[]
|
| 193 |
countColl=0
|
| 194 |
for CollID in CollIDs :
|
| 195 |
countColl=countColl+1
|
| 196 |
### map the CollectionID with the list of PubDB URLs
|
| 197 |
#pubdbmap[CollID]=self.findPubDBsbyCollID(CollID)
|
| 198 |
### prepare a list all PubDB urls for all collections
|
| 199 |
allurls.extend(self.findPubDBsbyCollID(CollID))
|
| 200 |
#print pubdbmap.values()
|
| 201 |
|
| 202 |
### select only PubDB urls that contains all the collections
|
| 203 |
unique_urls=self.uniquelist(allurls)
|
| 204 |
SelectedPubDBURLs=[]
|
| 205 |
# loop on a unique list of PubDB urls
|
| 206 |
for url in unique_urls :
|
| 207 |
# check that PubDBurl occurrance is the same as the number of collections
|
| 208 |
if ( allurls.count(url)==countColl ) :
|
| 209 |
SelectedPubDBURLs.append(url)
|
| 210 |
common.logger.debug(5,'PubDBs '+str(SelectedPubDBURLs))
|
| 211 |
|
| 212 |
#print 'Required Collections',CollIDs,'are all present in PubDBURLs : ',SelectedPubDBURLs,'\n'
|
| 213 |
#### check based on CE black list: select only PubDB not in the CE black list
|
| 214 |
tmp=self.checkBlackList(SelectedPubDBURLs)
|
| 215 |
common.logger.debug(5,'PubDBs after black list '+str(tmp))
|
| 216 |
|
| 217 |
### check based on CE white list: select only PubDB defined by user
|
| 218 |
GoodPubDBURLs=self.checkWhiteList(tmp)
|
| 219 |
common.logger.debug(5,'PubDBs after white list '+str(GoodPubDBURLs))
|
| 220 |
return GoodPubDBURLs
|
| 221 |
|
| 222 |
#######################################################################
|
| 223 |
def uniquelist(self, old):
|
| 224 |
"""
|
| 225 |
remove duplicates from a list
|
| 226 |
"""
|
| 227 |
nd={}
|
| 228 |
for e in old:
|
| 229 |
nd[e]=0
|
| 230 |
return nd.keys()
|
| 231 |
|
| 232 |
#######################################################################
|
| 233 |
def checkWhiteList(self, pubDBUrls):
|
| 234 |
"""
|
| 235 |
select PubDB URLs that are at site defined by the user (via CE white list)
|
| 236 |
"""
|
| 237 |
if len(self.reCEWhiteList)==0: return pubDBUrls
|
| 238 |
goodurls = []
|
| 239 |
for url in pubDBUrls:
|
| 240 |
#print 'connecting to the URL ',url
|
| 241 |
good=0
|
| 242 |
for re in self.reCEWhiteList:
|
| 243 |
if re.search(url):
|
| 244 |
common.logger.debug(5,'CE in white list, adding PubDB URL '+url)
|
| 245 |
good=1
|
| 246 |
if not good: continue
|
| 247 |
goodurls.append(url)
|
| 248 |
if len(goodurls) == 0:
|
| 249 |
common.logger.message("No selected PubDB URLs \n")
|
| 250 |
else:
|
| 251 |
common.logger.debug(5,"Selected PubDB URLs are "+str(goodurls)+"\n")
|
| 252 |
return goodurls
|
| 253 |
|
| 254 |
#######################################################################
|
| 255 |
def checkBlackList(self, pubDBUrls):
|
| 256 |
"""
|
| 257 |
select PubDB URLs that are at site not exluded by the user (via CE black list)
|
| 258 |
"""
|
| 259 |
if len(self.reCEBlackList)==0: return pubDBUrls
|
| 260 |
goodurls = []
|
| 261 |
for url in pubDBUrls:
|
| 262 |
common.logger.debug(10,'connecting to the URL '+url)
|
| 263 |
good=1
|
| 264 |
for re in self.reCEBlackList:
|
| 265 |
if re.search(url):
|
| 266 |
common.logger.message('CE in black list, skipping PubDB URL '+url)
|
| 267 |
good=0
|
| 268 |
pass
|
| 269 |
if good: goodurls.append(url)
|
| 270 |
if len(goodurls) == 0:
|
| 271 |
common.logger.debug(3,"No selected PubDB URLs")
|
| 272 |
return goodurls
|
| 273 |
|
| 274 |
########################################################################
|
| 275 |
def checkPubDBNewVersion(self, baseurl):
|
| 276 |
"""
|
| 277 |
Check PubDB version to find out if it's new-style or old-style
|
| 278 |
"""
|
| 279 |
### check based on the existance of pubdb-get-version.php
|
| 280 |
urlversion=baseurl+'pubdb-get-version.php'
|
| 281 |
newversion=1;
|
| 282 |
try:
|
| 283 |
v = urllib2.urlopen(urlversion)
|
| 284 |
except urllib2.URLError, msg:
|
| 285 |
#print "WARNING: no URL to get PubDB version "
|
| 286 |
newversion=0;
|
| 287 |
|
| 288 |
if (newversion) :
|
| 289 |
schemaversion = v.read()
|
| 290 |
#print schemaversion;
|
| 291 |
|
| 292 |
return newversion
|
| 293 |
|
| 294 |
########################################################################
|
| 295 |
def getPubDBData(self, CollIDs, url , newversion):
|
| 296 |
"""
|
| 297 |
Contact a PubDB to collect all the relevant information
|
| 298 |
"""
|
| 299 |
result = []
|
| 300 |
|
| 301 |
### get the base PubDb url
|
| 302 |
end=string.rfind(url,'/')
|
| 303 |
lastEq=string.rfind(url,'=')
|
| 304 |
|
| 305 |
if (newversion) :
|
| 306 |
### from PubDB V4 : get info for all the collections in one shot and unserialize the content
|
| 307 |
Collections=string.join(CollIDs,'-')
|
| 308 |
## add the PU among the required Collections if the Digi are requested
|
| 309 |
# ( for the time being asking it directly to the PubDB so the RefDB
|
| 310 |
# level data discovery is bypassed..... in future when every site
|
| 311 |
# will have the new style it will be possible to ask for PU , at RefDB level, in method findAllCollections )
|
| 312 |
if ( self.NeededdataTiers.count('Digi') ):
|
| 313 |
PUCollID=self.getDatatierCollID(url[:end+1],Collections,"PU")
|
| 314 |
if (PUCollID) : CollIDs.append(PUCollID)
|
| 315 |
##
|
| 316 |
Collections=string.join(CollIDs,'-')
|
| 317 |
### download from PubDB all the info about the given collections
|
| 318 |
pubdb_analysis=PubDBInfo(url[:end+1],Collections)
|
| 319 |
#print pubdb_analysis.GetPubDBInfo()
|
| 320 |
ok=0
|
| 321 |
try:
|
| 322 |
catInfos=pubdb_analysis.GetPubDBInfo()
|
| 323 |
ok=1
|
| 324 |
except :
|
| 325 |
#print "WARNING: can't get PubDB content out of "+url[:end+1]+"\n"
|
| 326 |
print '\nERROR extracting info for collections '+Collections+' from PubDB '+url[:end+1]+'. The PU might not be published at that site.\n'
|
| 327 |
#raise PubDBGetAnalysisError(url[:end+1],Collections)
|
| 328 |
if (ok): result=catInfos;
|
| 329 |
|
| 330 |
else:
|
| 331 |
|
| 332 |
### before PubDB V4 : get info for each collection and read the key-value pair text
|
| 333 |
|
| 334 |
for CollID in CollIDs:
|
| 335 |
urlphp=url[:end+1]+self.PubDBAnalysisPhp_+'?CollID='+CollID
|
| 336 |
# print 'PHP URL: '+urlphp+' \n'
|
| 337 |
|
| 338 |
reOld=re.compile( r'V24' )
|
| 339 |
#print urlphp,'Old PubDB ',reOld.search(urlphp)
|
| 340 |
if reOld.search(urlphp):
|
| 341 |
raise NoPHPError(urlphp)
|
| 342 |
else:
|
| 343 |
try:
|
| 344 |
f = urllib2.urlopen(urlphp)
|
| 345 |
except urllib2.URLError, msg:
|
| 346 |
print "WARNING: ", msg
|
| 347 |
raise PubDBError(urlphp)
|
| 348 |
except urllib2.HTTPError, msg:
|
| 349 |
print "WARNING: ", msg
|
| 350 |
raise NoPHPError(urlphp)
|
| 351 |
content = f.read()
|
| 352 |
result.append(pubDBResult(content))
|
| 353 |
#print "Coll",CollID," content ",content
|
| 354 |
pass
|
| 355 |
pass
|
| 356 |
|
| 357 |
#print '.....'
|
| 358 |
#for r in result:
|
| 359 |
# r.dump()
|
| 360 |
#print '.....'
|
| 361 |
return result
|
| 362 |
|
| 363 |
########################################################################
|
| 364 |
def getDatatierCollID(self,urlbase,CollIDString,datatier):
|
| 365 |
"""
|
| 366 |
Contact a script of PubDB to retrieve the collid a DataTier
|
| 367 |
"""
|
| 368 |
try:
|
| 369 |
f = urllib.urlopen(urlbase+'pubdb-get-collidbydatatier.php?collid='+CollIDString+"&datatier="+datatier)
|
| 370 |
except IOError:
|
| 371 |
raise PubDBGetAnalysisError(url[:end+1]+'pubdb-get-collidbydatatier.php',CollIDString)
|
| 372 |
data = f.read()
|
| 373 |
colldata=re.compile(r'collid=(\S*)').search(data);
|
| 374 |
if colldata:
|
| 375 |
datatier_CollID=colldata.group(1)
|
| 376 |
# print " --> asking to PubDB "+urlbase+" for an additional collection : ID= "+datatier_CollID+" DataTier= "+datatier
|
| 377 |
common.logger.message(" --> asking to PubDB "+urlbase+" for an additional collection : ID= "+datatier_CollID+" DataTier= "+datatier)
|
| 378 |
|
| 379 |
return datatier_CollID
|
| 380 |
|
| 381 |
########################################################################
|
| 382 |
def getAllPubDBData(self):
|
| 383 |
"""
|
| 384 |
Contact a list of PubDB to collect all the relevant information
|
| 385 |
"""
|
| 386 |
newPubDBResult=[]
|
| 387 |
oldPubDBResult=[]
|
| 388 |
Result={}
|
| 389 |
|
| 390 |
### find the user-required collection IDs
|
| 391 |
CollIDs = self.findAllCollections()
|
| 392 |
### find the PubDB URLs publishing the needed data
|
| 393 |
urllist = self.findPubDBs(CollIDs)
|
| 394 |
### collect information sparatelly from new-style PubDBs and old-style PubDBs
|
| 395 |
for pubdburl in urllist:
|
| 396 |
end=string.rfind(pubdburl,'/')
|
| 397 |
newversion=self.checkPubDBNewVersion(pubdburl[:end+1])
|
| 398 |
if (newversion):
|
| 399 |
res=self.getPubDBData(CollIDs,pubdburl,newversion)
|
| 400 |
if len(res)>0:
|
| 401 |
newPubDBResult.append(res)
|
| 402 |
else:
|
| 403 |
resold=self.getPubDBData(CollIDs,pubdburl,newversion)
|
| 404 |
if len(resold)>0:
|
| 405 |
oldPubDBResult.append(resold)
|
| 406 |
### fill a dictionary with all the PubBDs results both old-style and new-style
|
| 407 |
Result['newPubDB']=newPubDBResult
|
| 408 |
Result['oldPubDB']=oldPubDBResult
|
| 409 |
|
| 410 |
## print for debugging purpose
|
| 411 |
#
|
| 412 |
#for PubDBversion in Result.keys():
|
| 413 |
#print ("key %s, val %s" %(PubDBversion,Result[PubDBversion]))
|
| 414 |
# if len(Result[PubDBversion])>0 :
|
| 415 |
#print (" key %s"%(PubDBversion))
|
| 416 |
# for result in Result[PubDBversion]:
|
| 417 |
# for r in result:
|
| 418 |
#r.dump()
|
| 419 |
# common.log.write('----------------- \n')
|
| 420 |
#print '.....................................'
|
| 421 |
|
| 422 |
return Result
|
| 423 |
|
| 424 |
####################################################################
|
