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 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 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.')
