def getCommonDatesLocations(iterator,metadata,times,locations,dimensions,k):
vNInd = '0'
overallDateFormat = iterator.overallDateFormat if hasattr(iterator,'overallDateFormat') else ''
dateFormat = iterator.dateFormat if hasattr(iterator,'dateFormat') else ''
overallDate = iterator.overallDate if hasattr(iterator,'overallDate') else ''
if overallDateFormat or dateFormat:
DF = overallDateFormat + dateFormat
F = td.mongotimeformatter(DF)
T1 = [F(overallDate + x) for x in iterator.columnGroups['timeColNames'] if x in dimensions[k]]
if overallDateFormat:
reverseF = td.reverse(dateFormat)
T2 = [F(overallDate + y) for y in map(reverseF,times[k])]
else:
T2 = times[k]
mindate = min(T1 + T2)
maxdate = max(T1 + T2)
divisions = uniqify(ListUnion([td.getLowest(t) for t in T1 + T2]))
metadata[k]['beginDate'] = mindate
metadata[k]['endDate'] = maxdate
metadata[k]['dateDivisions'] = divisions
#locations
if locations[k]:
if hasattr(iterator,'overallLocation'):
locs = [loc.integrate(iterator.overallLocation,l) for l in locations[k]]
else:
locs = locations[k]
locs = locListUniqify(locs)
metadata[k]['spatialDivisions'] = uniqify(ListUnion([loc.divisions(x) for x in locs]))
metadata[k]['commonLocation'] = reduce(loc.intersect,locs)
def getSliceColTuples(collection):
sliceColList = collection.sliceCols
sliceColU = uniqify(ListUnion(sliceColList))
sliceColInds = getStrs(collection,sliceColU)
OK = dict([(x,x in collection.columnGroups.keys() or MoreThanOne(collection,y)) for (y,x) in zip(sliceColInds,sliceColU)])
sliceColList = [tuple([x for x in sliceColU if x in sc and OK[x]]) for sc in sliceColList]
sliceColTuples = uniqify(ListUnion([subTuples(sc) for sc in sliceColList]))
return sliceColTuples
def queryToSolr(timeQuery):
"""converts timeQuery for use in find ie solr"""
#TODO: query between march and december fails handle things that don't start with year
F = mongotimeformatter(timeQuery['format'])
for k in timeQuery.keys():
if k != 'format':
timeQuery[k] = F(timeQuery[k])
if timeQuery.keys() == ['format']:
divisions = [TIME_DIVISIONS[x] for x in uniqify(timeQuery['format'])]
fq = 'dateDivisions:' + (divisions[0] if len(divisions) == 1 else '(' + ' AND '.join(divisions) + ')')
else:
if 'on' in timeQuery.keys():
start = timeQuery['on']
end = timeQuery['on']
else:
start = timeQuery['start'] if 'start' in timeQuery.keys() else None
end = timeQuery['end'] if 'end' in timeQuery.keys() else None
start = convertToSolrDT(start,convertMode='High') if start else None
end = convertToSolrDT(end) if end else None
fq = []
if start:
fq.append('beginDate:[* TO ' + start + ']')
if end:
fq.append('endDate:[' + end + ' TO *]')
return fq
def regionsGuts(g,level_code):
D = {'method':'filter'}
if 'return' in g:
D['return'] = g['return']
D['field'] = 'geom'
if 'bounds' in g:
w,s,e,n = g['bounds'].split(',')
if 'type' in g:
D['type'] = g['type']
else:
D['type'] = 'intersects'
D['query'] = 'POLYGON((' + ', '.join([w + ' ' + n , w + ' ' + s, e + ' ' + s , e + ' ' + n, w + ' ' + n]) + ' ))'
elif 'radius' in g and 'center' in g:
D['radius'] = g['radius']
x,y = g['center'].split(',')
D['type'] = 'distance_lte'
D['query'] = 'POINT(' + x + ' ' + y + ')'
if 'units' in g:
D['units'] = str(g['units'])
R = uniqify([tuple(x.items()) for x in geodbGuts(D,level_code)])
R = [dict([(k,v) for (k,v) in r if v != 'None']) for r in R]
return R
def locListUniqify(D):
D = [tuple([(y,z) if y != 'f' else (y,tuple(z.items())) for (y,z) in x.items()]) for x in D]
D = uniqify(D)
D = [dict(d) for d in D]
for d in D:
if d.has_key('f'):
d['f'] = dict(d['f'])
return D
def sliceInsert(c,collection,sliceColTuples,VarMap,sliceDB,version):
dontcheck = []
for sct in sliceColTuples:
if all([VarMap[k] in c.keys() for k in sct]):
slice = pm.son.SON([(k,c[VarMap[k]]) for k in sct if VarMap[k] in c.keys()])
dc = sct in dontcheck
if dc or not sliceDB.find_one({'slice':slice,'version':version}):
if not dc:
SCT = set(sct)
dontcheck = uniqify(dontcheck + [ss for ss in sliceColTuples if SCT <= set(ss)])
sliceDB.update({'slice':slice},{'$set':{'version':version,'original':version}},upsert=True)
def generateQueries(DateFormat,timeQuery):
"""Converts nice DateFormat string and simple query format for time and generate mongo icky
DateFormat : String e.g. YYYYmmdd
timeQuery : Dict keys = format, begin, end, on
These are not necessary"""
timeQueryFormat = timeQuery['format'] if 'format' in timeQuery.keys() else DateFormat
tQFset = set(timeQueryFormat)
tFset = set(DateFormat)
if tQFset <= tFset:
tQHier = getHierarchy(tQFset)
Hier = getHierarchy(DateFormat)
mergedTimeFormat = ''.join(tFset.difference(tQFset)) + timeQueryFormat
timeFormatter = mongotimeformatter(mergedTimeFormat)
zeroLen = len(tFset.difference(tQFset))
tQHier0 = [x[0] for x in tQHier]
Hier0 = [x[0] for x in Hier]
basePathDict = dict([(m,getPathsTo(m,Hier)) for m in tQHier0])
belowPathDict = dict([(m,getPathsBelow(m,Hier)) for m in tQHier0])
Q = {}
for (k,op) in [('begin','$gte'),('end','$lt')]:
if k in timeQuery.keys():
timeObj = timeFormatter('X'*zeroLen + timeQuery[k])
for m in basePathDict.keys():
for p in basePathDict[m]:
if p in Q.keys():
Q[p][op] = rgetattr(timeObj,p)
else:
Q[p] = {op: rgetattr(timeObj,p)}
if 'on' in timeQuery.keys():
timeObj = timeFormatter('X'*zeroLen + timeQuery['on'])
paths = uniqify(ListUnion([getPathsTo(m,Hier) for m in tQFset]))
for p in paths:
p = p + ('',)
Q[p] = rgetattr(timeObj,p)
if not set(timeQuery.keys()).intersection(['begin','end','on']):
for m in set(tQHier0):
for p in basePathDict[m]:
p = p + ('',)
Q[p] = {'$exists':True}
for p in belowPathDict[m]:
p = p + ('',)
Q[p] = {'$exists':False}
return Q
def generateQueries(spaceQuery):
"""Generates queries for get in mongo"""
Q = {}
if isinstance(spaceQuery,list) or isinstance(spaceQuery,tuple):
for x in uniqify(spaceQuery):
Q[tuple(x.split('.'))] = {'$exists':True}
elif hasattr(spaceQuery,'keys'):
spaceQuery = convertSQ(spaceQuery)
for x in spaceQuery.keys():
Q[tuple(x.split('.'))] = spaceQuery[x]
return Q
def SpaceComplete(x):
"""FIPS -> names and upwards when possible"""
if 'f' in x.keys():
x = x.copy()
iFIPS = ListUnion([SPACE_HIERARCHY_R[c] for c in x['f'].keys()])
iFIPS = [c for c in iFIPS if c not in x.keys()]
Cset = [c + '=' + x['f'][c] for c in x['f'].keys() if uniqify(x['f'][c]) != ['0']]
if iFIPS and Cset:
X = eval(urllib2.urlopen('http://localhost:8000/geo/fips/?' + '&'.join(Cset)).read())
if len(X) == 1:
X = convertToCodes(X[0])
x['f'] = X['f']
for c in X.keys():
if c not in x.keys():
x[c] = X[c]
return x
def getQueryList(collection,keys,atVersion,toVersion,slicesCorrespondToIndexes):
totalVariables = collection.columns
VarMap = dict(zip(totalVariables,[str(x) for x in range(len(totalVariables))]))
origInd = VarMap['__originalVersion__'] ; retInd = VarMap['__retained__'] ; vNInd = VarMap['__versionNumber__']
keys = [str(x) for x in keys]
existence = [(k,{'$exists':True,'$ne':''}) for k in keys]
if keys:
Q1 = processArg(dict([(origInd,{'$gt':atVersion}),(vNInd,toVersion)] + existence),collection)
Q2 = processArg(dict([(retInd,{'$exists':False}),(vNInd,{'$lt':toVersion,'$gte':atVersion})] + existence),collection)
Q3 = processArg(dict([(retInd,True),(vNInd,{'$lt':toVersion,'$gte':atVersion}),(origInd,{'$lte':atVersion})] + existence),collection)
colnames = [k for k in keys if k.split('.')[0] in collection.columns]
colgroups = [k for k in keys if k in collection.columnGroups]
T= ListUnion([collection.columnGroups[k] for k in colgroups])
kInds = getStrs(collection,colnames + T)
R = list(collection.find(Q1,fields = kInds)) + list(collection.find(Q2,fields = kInds)) + (list(collection.find(Q3,fields = kInds)) if not slicesCorrespondToIndexes else [])
R = [son.SON([(collection.columns[int(k)],r[k]) for k in r.keys() if k.isdigit() and r[k]]) for r in R]
R = [[(k,rgetattr(r,k.split('.'))) for k in keys if rhasattr(r,k.split('.')) if k not in T] + [(g,[r[k] for k in collection.columnGroups[g] if k in r.keys() and r[k]]) for g in colgroups ] for r in R]
return uniqify(ListUnion([expand(r) for r in R]))
else:
return [()]
def initialize_argdict(collection):
d = {} ; ArgDict = {}
sliceCols = uniqify(Flatten(collection.sliceCols))
sliceColList = ListUnion([[x] if x.split('.')[0] in collection.columns else collection.columnGroups.get(x,[]) for x in sliceCols])
if hasattr(collection,'contentCols'):
contentColList = ListUnion([[x] if x.split('.')[0] in collection.columns else collection.columnGroups.get(x,[]) for x in collection.contentCols])
contentCols = uniqify(contentColList + sliceColList)
else:
contentCols = sliceColList
contentColNums = getStrs(collection,contentCols)
ArgDict['contentColNums'] = contentColNums
if hasattr(collection,'dateFormat'):
dateFormat = collection.dateFormat
ArgDict['overallDateFormat'] = dateFormat
timeFormatter = td.mongotimeformatter(dateFormat)
ArgDict['timeFormatter'] = timeFormatter
else:
dateFormat = ''
if hasattr(collection,'overallDate'):
od = collection.overallDate['date']
odf = collection.overallDate['format']
ArgDict['overallDate'] = od
overallDateFormat = odf + dateFormat
ArgDict['overallDateFormat'] = overallDateFormat
timeFormatter = td.mongotimeformatter(overallDateFormat)
ArgDict['timeFormatter'] = timeFormatter
OD = timeFormatter(overallDate +'X'*len(dateFormat))
ArgDict['dateDivisions'] = td.getLowest(OD)
ArgDict['datePhrases'] = [td.phrase(OD)]
ArgDict['mindate'] = OD
ArgDict['maxdate'] = OD
if dateFormat:
reverseTimeFormatter = td.reverse(dateFormat)
ArgDict['reverseTimeFormatter'] = reverseTimeFormatter
else:
od = ''
if 'timeColNames' in collection.columnGroups.keys():
timeColNamesInd = getNums(collection,collection.columnGroups['timeColNames'])
tcs = [timeFormatter(od + t) for t in collection.columnGroups['timeColNames']]
ArgDict['timeColNames'] = tcs
ArgDict['timeColNameInds'] = timeColNamesInd
ArgDict['timeColNameDivisions'] = [[td.TIME_DIVISIONS[x] for x in td.getLowest(tc)] for tc in tcs]
ArgDict['timeColNamePhrases'] = [td.phrase(t) for t in tcs]
if 'timeColumns' in collection.columnGroups.keys():
ArgDict['timeColInds'] = getNums(collection,collection.columnGroups['timeColumns'])
#overall location
if hasattr(collection,'overallLocation'):
ol = collection.overallLocation
ArgDict['overallLocation'] = ol
else:
ol = None
#get divisions and phrases from OverallLocation and SpaceColNames
if 'spaceColNames' in collection.columnGroups.keys():
spaceColNames = collection.columnGroups['spaceColNames']
ArgDict['spaceColNames'] = [loc.integrate(ol,x) for x in spaceColNames]
if 'spaceColumns' in collection.columnGroups.keys():
ArgDict['spaceColInds'] = getNums(collection,collection.columnGroups['spaceColumns'])
Source = collection.source
SourceNameDict = son.SON([(k,Source[k]['name'] if isinstance(Source[k],dict) else Source[k]) for k in Source.keys()])
SourceAbbrevDict = dict([(k,Source[k]['shortName']) for k in Source.keys() if isinstance(Source[k],dict) and 'shortName' in Source[k].keys() ])
d['sourceSpec'] = json.dumps(SourceNameDict,default=ju.default)
d['agency'] = SourceNameDict['agency']
d['subagency'] = SourceNameDict['subagency']
d['dataset'] = SourceNameDict['dataset']
for k in SourceNameDict.keys():
d['source_' + str(k).lower()] = SourceNameDict[k]
for k in SourceAbbrevDict.keys():
d['source_' + str(k).lower() + '_acronym'] = SourceAbbrevDict[k]
d['source'] = ' '.join(SourceNameDict.values() + SourceAbbrevDict.values())
if 'subcollections' in collection.columns:
ArgDict['subColInd'] = collection.columns.index('subcollections')
value_processor_instructions = stringifyDictElements(collection.valueProcessors)
vpcontext = commonjs.translatorContext(value_processor_instructions)
ArgDict['valueProcessors'],ArgDict['valueProcessorsKey'] = get_processors(value_processor_instructions,collection, vpcontext ,commonjs.js_call)
return d, ArgDict
def largeAdd(d,query,collection,contentColNums, timeColInds ,timeColNames , timeColNameInds ,timeColNameDivisions ,timeColNamePhrases ,overallDate , overallDateFormat, timeFormatter ,reverseTimeFormatter ,dateDivisions ,datePhrases ,mindate ,maxdate ,overallLocation , spaceColNames, spaceColInds, subColInd, valueProcessors):
print '0'
exists = []
check = range(len(collection.columns))
while check:
k = check.pop(0)
rec = collection.find_one( dict(query.items() + [(str(k),{'$exists':True})]))
if rec:
rec.pop('_id')
new = map(int,rec.keys())
check = list(set(check).difference(new))
check.sort()
exists += [(pos,collection.columns[pos]) for pos in new]
print '1'
exists = [e for e in exists if e[1] not in SPECIAL_KEYS]
(colnums,colnames) = zip(*exists)
d['columnNames'] = colnames
d['dimension'] = len(d['columnNames'])
if overallDateFormat:
d['dateFormat'] = overallDateFormat
if timeColInds:
dateColVals = ListUnion([collection.find(query).distinct(str(t)) for t in timeColInds if t in colnums])
if overallDate:
dateColVals = [timeFormatter(overallDate + reverseTimeFormatter(time)) for time in dateColVals]
dateDivisions += uniqify(ListUnion(map(td.getLowest,dateColVals)))
datePhrases += uniqify(map(td.phrase, dateColVals))
mindate = td.makemin(mindate,min(dateColVals),)
maxdate = td.makemax(maxdate,max(dateColVals),)
if timeColNameInds:
K = [k for (k,j) in enumerate(timeColNameInds) if k in colnums]
dateDivisions += uniqify(ListUnion([timeColNameDivisions[k] for k in K]))
mindate = td.makemin(mindate,min([timeColNames[k] for k in K]),)
maxdate = td.makemax(maxdate,max([timeColNames[k] for k in K]),)
datePhrases += [timeColNamePhrases[k] for k in K]
dateDivisions = uniqify(dateDivisions)
datePhrases = uniqify(datePhrases)
d['beginDate'] = td.convertToDT(mindate)
d['endDate'] = td.convertToDT(maxdate,convertMode='High')
d['dateDivisions'] = uniqify(dateDivisions)
d['datePhrases'] = datePhrases
print '2'
if spaceColInds:
spaceColVals = ListUnion([collection.find(query).distinct(str(t)) for t in spaceColInds if t in colnums])
spaceColVals = [loc.integrate(overallLocation,scv) for scv in spaceColVals]
else:
spaceColVals = []
spaceVals = spaceColNames + spaceColVals
if spaceVals:
d['spatialDivisions'] = uniqify(ListUnion(map(loc.divisions,spaceVals)))
d['spatialDivisionsTight'] = uniqify(ListUnion(map(loc.divisions2,spaceVals)))
d['spatialPhrases'] = uniqify(map(loc.phrase,spaceVals))
d['spatialPhrasesTight'] = uniqify(map(loc.phrase2,spaceVals))
commonLocation = overallLocation
for sv in spaceVals:
commonLocation = loc.intersect(commonLocation,sv)
if not commonLocation:
break
if commonLocation:
d['commonLocation'] = loc.phrase(commonLocation)
print '3'
d['sliceContents'] = ' '.join(uniqify(ListUnion([translate_list(valueProcessors.get(x,None) ,map(decode_obj,collection.find(query).distinct(x))) for x in contentColNums])))
return d
def smallAdd(d,query,collection,contentColNums, timeColInds ,timeColNames , timeColNameInds ,timeColNameDivisions ,timeColNamePhrases ,overallDate, overallDateFormat, timeFormatter ,reverseTimeFormatter ,dateDivisions ,datePhrases ,mindate ,maxdate ,overallLocation , spaceColNames , spaceColInds ,subColInd, valueProcessors,slicecount):
R = collection.find(query,timeout=False)
colnames = []
d['sliceContents'] = []
Subcollections = []
spaceVals = spaceColNames
commonLocation = overallLocation
for sv in spaceColNames:
commonLocation = loc.intersect(commonLocation,sv)
if not commonLocation:
break
for (i,r) in enumerate(R):
d['sliceContents'].append(' '.join([translate(valueProcessors.get(x,None),decode_obj(rgetattr(r,x.split('.')))) if rhasattr(r,x.split('.')) else '' for x in contentColNums]))
colnames = uniqify(colnames + r.keys())
if subColInd:
Subcollections += r[str(subColInd)]
if timeColInds:
for x in timeColInds:
if str(x) in r.keys():
time = r[str(x)]
if overallDate:
time = timeFormatter(overallDate + reverseTimeFormatter(time))
dateDivisions += td.getLowest(time)
datePhrases.append(td.phrase(time))
mindate = td.makemin(mindate,time)
maxdate = td.makemax(maxdate,time)
if spaceColInds:
for x in spaceColInds:
if str(x) in r.keys():
location = loc.integrate(overallLocation,r[str(x)])
commonLocation = loc.intersect(commonLocation,r[str(x)]) if commonLocation != None else None
spaceVals.append(location)
d['sliceContents'] = ' '.join(d['sliceContents'])
Subcollections = uniqify(Subcollections)
d['columnNames'] = [collection.columns[int(x)] for x in colnames if x.isdigit()]
d['dimension'] = len(d['columnNames'])
#time/date
if overallDateFormat:
d['dateFormat'] = overallDateFormat
if 'timeColNames' in collection.columnGroups.keys():
K = [k for (k,j) in enumerate(timeColNameInds) if str(j) in colnames]
dateDivisions += uniqify(ListUnion([timeColNameDivisions[k] for k in K]))
mindate = td.makemin(mindate,min([timeColNames[k] for k in K]))
maxdate = td.makemax(maxdate,max([timeColNames[k] for k in K]))
datePhrases += uniqify([timeColNamePhrases[k] for k in K])
d['beginDate'] = td.convertToDT(mindate)
d['endDate'] = td.convertToDT(maxdate,convertMode='High')
d['dateDivisions'] = uniqify(dateDivisions)
d['datePhrases'] = datePhrases if d['volume'] < 10000 else uniqify(datePhrases)
if spaceVals:
d['spatialDivisions'] = uniqify(ListUnion(map(loc.divisions,spaceVals)))
d['spatialDivisionsTight'] = uniqify(ListUnion(map(loc.divisions2,spaceVals)))
d['spatialPhrases'] = uniqify(map(loc.phrase,spaceVals))
d['spatialPhrasesTight'] = uniqify(map(loc.phrase2,spaceVals))
return d
def addToIndex(q,d,collection,solr_interface,slicecount,contentColNums = None, timeColInds=None,timeColNames=None, timeColNameInds = None,timeColNameDivisions = None,timeColNamePhrases=None,overallDate = '', overallDateFormat = '', timeFormatter = None,reverseTimeFormatter = None,dateDivisions=None,datePhrases=None,mindate = None,maxdate = None,overallLocation = None, spaceColNames = None, spaceColInds = None,subColInd = None,Return=False,valueProcessors=None,valueProcessorsKey=None):
q['__versionNumber__'] = collection.currentVersion
query = processArg(q,collection)
q.pop('__versionNumber__')
d['collectionName'] = collection.name
d['query'] = json.dumps(q,default=ju.default)
d['mongoID'] = mongoID(q,collection.name)
d['mongoText'] = queryToText(q,valueProcessorsKey)
d['sliceValues'] = queryValues(q,valueProcessorsKey)
d['sliceKeys'] = queryKeys(q,valueProcessorsKey)
d['versionNumber'] = collection.currentVersion
if dateDivisions == None:
dateDivisions = []
else:
dateDivisions = dateDivisions[:]
if datePhrases == None:
datePhrases = []
else:
datePhrases = datePhrases[:]
if spaceColNames == None:
spaceColNames = []
#stats
d['volume'] = collection.find(query).count()
contentColNums = [i for i in contentColNums if i not in query.keys()]
if d['volume'] > 0:
if d['volume'] < 5000:
smallAdd(d,query,collection,contentColNums, timeColInds ,timeColNames , timeColNameInds ,timeColNameDivisions ,timeColNamePhrases ,overallDate , overallDateFormat, timeFormatter ,reverseTimeFormatter ,dateDivisions ,datePhrases ,mindate ,maxdate ,overallLocation , spaceColNames , spaceColInds ,subColInd, valueProcessors,slicecount)
else:
largeAdd(d,query,collection,contentColNums, timeColInds ,timeColNames , timeColNameInds ,timeColNameDivisions ,timeColNamePhrases ,overallDate, overallDateFormat, timeFormatter ,reverseTimeFormatter ,dateDivisions ,datePhrases ,mindate ,maxdate ,overallLocation , spaceColNames , spaceColInds ,subColInd, valueProcessors)
Subcollections = uniqify(ListUnion(collection.find(query).distinct(str(subColInd))))
metadata = collection.metadata['']
for sc in Subcollections:
metadata.update(collection.metadata.get(sc,{}))
for k in metadata.keys():
if k in STANDARD_META:
if k in STANDARD_META_FORMATS.keys():
val = coerceToFormat(metadata[k],STANDARD_META_FORMATS[k])
if val:
d[str(k)] = val
else:
d[str(k)] = str(metadata[k])
if Return:
return d
else:
solr_interface.add(**d)
def getSliceColTuples(sliceCols):
sliceColList = sliceCols
sliceColTuples = uniqify(ListUnion([subTuples(tuple(sc)) for sc in sliceColList]))
return sliceColTuples
def updateCollection(download_dir,collectionName,parserClass,checkpath,certpath,parserArgs=None,parserKwargs=None,incremental=False):
connection = pm.Connection(document_class=pm.son.SON)
source_metadata = get_source_data(collectionName)
db = connection['govdata']
assert not '__' in collectionName, 'collectionName must not contain consecutive underscores'
metaCollectionName = '__' + collectionName + '__'
versionName = '__' + collectionName + '__VERSIONS__'
sliceDBName = '__' + collectionName + '__SLICES__'
collection = db[collectionName]
metacollection = db[metaCollectionName]
versions = db[versionName]
sliceDB = db[sliceDBName]
if incremental:
if versionName not in db.collection_names():
startInc = 0
else:
startInc = get_max_increment_fromDB(versions) + 1
endInc = get_max_increment(download_dir)
sources = [increment_format(download_dir,i) for i in range(startInc,endInc + 1)]
else:
sources = [download_dir]
startInc = endInc = None
if parserArgs == None:
parserArgs = ()
if parserKwargs == None:
parserKwargs = {}
if sources:
iterator = parserClass(sources[0],*parserArgs,**parserKwargs)
iterator.set_source_metadata(source_metadata)
uniqueIndexes = iterator.uniqueIndexes
ColumnGroups = iterator.columnGroups
sliceColTuples = getSliceColTuples(iterator.sliceCols)
sliceColTuplesFlat = uniqify([tuple(sorted(uniqify(Flatten(sct)))) for sct in sliceColTuples])
sliceColList = uniqify(Flatten(ListUnion(sliceColTuples)))
ContentCols = set(sliceColList + getContentCols(iterator))
if hasattr(iterator,'dateFormat'):
TimeFormatter = td.mongotimeformatter(iterator.dateFormat)
if collectionName in db.collection_names():
versionNumber = max(versions.distinct('versionNumber')) + 1
storedAllMetadata = metacollection.find_one({'name':'','versionNumber':versionNumber-1})
totalVariables = storedAllMetadata['columns']
VarMap = dict(zip(totalVariables,[str(x) for x in range(len(totalVariables))]))
#check things are the same
#and check consistent do so for all soruces
else:
versionNumber = 0
IndexCols = uniqify([x for x in ['subcollections'] + sliceColList + ListUnion([ColGroupsFlatten(ColumnGroups,k) for k in ['indexColumns','labelColumns','timeColumns','spaceColumns']]) if x not in uniqueIndexes])
totalVariables = SPECIAL_KEYS + uniqueIndexes + IndexCols
assert not any(['.' in x or ('__' in x and x not in SPECIAL_KEYS) or x in ColumnGroups.keys() for x in totalVariables])
VarMap = dict(zip(totalVariables,map(str,range(len(totalVariables)))))
cols = zip([VarMap[c] for c in uniqueIndexes + ['__versionNumber__']],[pm.DESCENDING]*(len(uniqueIndexes) + 1))
collection.ensure_index(cols,unique=True,dropDups=True)
for col in IndexCols:
collection.ensure_index(VarMap[col])
sliceDB.ensure_index('slice',unique=True,dropDups=True)
vNInd = VarMap['__versionNumber__']
retInd = VarMap['__retained__']
specialKeyInds = [VarMap[k] for k in SPECIAL_KEYS]
if 'timeColumns' in iterator.columnGroups.keys():
tcs = iterator.columnGroups['timeColumns']
else:
tcs = []
if 'spaceColumns' in iterator.columnGroups.keys():
spcs = iterator.columnGroups['spaceColumns']
else:
spcs = []
toParse = ListUnion([RecursiveFileList(source + '__PARSE__') for source in sources])
oldc = None
SpaceCache = {}
volumes = {'':0}
dimensions = {'':[]}
times = {'':[]}
locations = {'':[]}
varFormats = {}
for file in toParse:
iterator.refresh(file)
checkMetadata(iterator)
tcs = iterator.columnGroups.get('timeColumns',[])
spcs = iterator.columnGroups.get('spaceColumns',[])
index = 0
for c in iterator:
newVars = [x for x in c.keys() if not x in totalVariables]
assert not any (['__' in x or '.' in x or x in ColumnGroups.keys() for x in newVars]) , '__ and . must not appear in key names.'
totalVariables += newVars
VarMap.update(dict(zip(newVars,map(str,range(len(totalVariables) - len(newVars),len(totalVariables))))))
for tc in tcs: #time handling
if tc in c.keys():
c[tc] = TimeFormatter(c[tc])
if COMPLETE_SPACE:
for spc in spcs:
if spc in c.keys(): #space
t = getT(c[spc])
if t in SpaceCache.keys():
c[spc] = SpaceCache[t].copy()
else:
c[spc] = loc.SpaceComplete(c[spc])
SpaceCache[t] = c[spc].copy()
if index % 100 == 0:
print 'At', index
index += 1
sctf = processSct(sliceColTuplesFlat,oldc,c)
processRecord(c,collection,VarMap,totalVariables,uniqueIndexes,versionNumber,specialKeyInds,incremental,sliceDB,sctf,ContentCols)
incrementThings(c,volumes,dimensions,times,locations,varFormats,tcs,spcs)
oldc = c
any_deleted = False
if incremental:
collection.update({vNInd:{'$lte': versionNumber - 1}}, {'$set':{vNInd:versionNumber}})
sliceDB.update({},{'$set':{'version':versionNumber}})
else:
deleted = collection.find({vNInd:versionNumber - 1, retInd : {'$exists':False}})
for d in deleted:
any_deleted = True
sliceDelete(d,collection,sliceColTuples,VarMap,sliceDB,version)
if any_deleted:
subColInd = str(totalVariables.index('Subcollections'))
subcols = [''] + uniqify(ListUnion(collection.distinct(subColInd)))
for sc in subcols:
volumes[sc] = collection.find({subColInd:sc}).count()
dimensions[sc] = [k for k in totalVariables if collection.find_one({subColInd:sc,str(totalVariables.index(k)):{'$exists':True}})]
times[sc] = ListUnion([collection.find({subColInd:sc}).distinct(t) for t in tcs])
locations[sc] = ListUnion([collection.find({subColInd:sc}).distinct(t) for t in spcs])
updateMetacollection(iterator,metacollection,incremental,versionNumber,totalVariables,tcs,spcs,volumes,dimensions,times,locations,varFormats)
updateAssociatedFiles(sources,collection)
updateVersionHistory(versionNumber,versions,startInc,endInc)
updateSourceDBFromCollections(collectionNames = [collectionName])
connection.disconnect()
createCertificate(certpath,'Collection ' + collectionName + ' written to DB.')
