def streamingExtensionsLoader(modelXbrl, mappedUri, filepath, *args, **kwargs):
# check if big instance and has header with an initial incomplete tree walk (just 2 elements
if not _streamingExtensionsCheck:
return None
# track whether modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = False
def logSyntaxErrors(parsercontext):
for error in parsercontext.error_log:
modelXbrl.error(
"xmlSchema:syntax",
_("%(error)s, %(fileName)s, line %(line)s, column %(column)s, %(sourceAction)s source element"
),
modelObject=modelXbrl,
fileName=os.path.basename(filepath),
error=error.message,
line=error.line,
column=error.column,
sourceAction="streaming")
#### note: written for iterparse of lxml prior to version 3.3, otherwise rewrite to use XmlPullParser ###
#### note: iterparse wants a binary file, but file is text mode
_file, = modelXbrl.fileSource.file(filepath, binary=True)
startedAt = time.time()
modelXbrl.profileActivity()
''' this seems twice as slow as iterparse
class instInfoTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.streamingAspects = None
self.foundInstance = False
self.creationSoftwareComment = ''
self.currentEltTag = "(before xbrli:xbrl)"
self.numRootFacts = 0
def start(self, tag, attrib, nsmap=None):
if self.newTree:
if tag == "{http://www.xbrl.org/2003/instance}xbrl":
self.foundInstance = True
self.newTree = False
else: # break
raise NotInstanceDocumentException()
elif not tag.startswith("{http://www.xbrl.org/"):
self.numRootFacts += 1
if self.numRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
self.currentEltTag = tag
def end(self, tag):
pass
def data(self, data):
pass
def comment(self, text):
if not self.foundInstance: # accumulate comments before xbrli:xbrl
self.creationSoftwareComment += ('\n' if self.creationSoftwareComment else '') + text
elif not self.creationSoftwareComment:
self.creationSoftwareComment = text # or first comment after xbrli:xbrl
def pi(self, target, data):
if target == "xbrl-streamable-instance":
if self.currentEltTag == "{http://www.xbrl.org/2003/instance}xbrl":
self.streamingAspects = dict(etree.PI(target,data).attrib.copy()) # dereference target results
else:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(target)s, must follow xbrli:xbrl element but was found at %(element)s"),
modelObject=modelXbrl, target=target, element=self.currentEltTag)
def close(self):
if not self.creationSoftwareComment:
self.creationSoftwareComment = None
return True
instInfo = instInfoTarget()
infoParser = etree.XMLParser(recover=True, huge_tree=True, target=instInfo)
try:
etree.parse(_file, parser=infoParser, base_url=filepath)
except NotInstanceDocumentException:
pass
'''
foundErrors = False
foundInstance = False
streamingAspects = None
creationSoftwareComment = None
instInfoNumRootFacts = 0
numElts = 0
elt = None
instInfoContext = etree.iterparse(_file,
events=("start", "end"),
huge_tree=True)
try:
for event, elt in instInfoContext:
if event == "start":
if elt.getparent() is not None:
if elt.getparent(
).tag == "{http://www.xbrl.org/2003/instance}xbrl":
if not foundInstance:
foundInstance = True
pi = precedingProcessingInstruction(
elt, "xbrl-streamable-instance")
if pi is None:
break
else:
streamingAspects = dict(pi.attrib.copy())
if creationSoftwareComment is None:
creationSoftwareComment = precedingComment(
elt)
if not elt.tag.startswith("{http://www.xbrl.org/"):
instInfoNumRootFacts += 1
if instInfoNumRootFacts % 1000 == 0:
modelXbrl.profileActivity(
"... streaming tree check",
minTimeToShow=20.0)
elif not foundInstance:
break
elif elt.tag == "{http://www.xbrl.org/2003/instance}xbrl":
creationSoftwareComment = precedingComment(elt)
if precedingProcessingInstruction(
elt, "xbrl-streamable-instance") is not None:
modelXbrl.error(
"streamingExtensions:headerMisplaced",
_("Header is misplaced: %(error)s, must follow xbrli:xbrl element"
),
modelObject=elt)
elif event == "end":
elt.clear()
numElts += 1
if numElts % 1000 == 0 and elt.getparent() is not None:
while elt.getprevious() is not None and elt.getparent(
) is not None:
del elt.getparent()[0]
except etree.XMLSyntaxError as err:
modelXbrl.error("xmlSchema:syntax",
_("Unrecoverable error: %(error)s"),
error=err)
_file.close()
return err
_file.seek(0, io.SEEK_SET) # allow reparsing
if not foundInstance or streamingAspects is None:
del elt
_file.close()
return None
modelXbrl.profileStat(_("streaming tree check"), time.time() - startedAt)
startedAt = time.time()
try:
version = Decimal(streamingAspects.get("version"))
if int(version) != 1:
modelXbrl.error(
"streamingExtensions:unsupportedVersion",
_("Streaming version %(version)s, major version number must be 1"
),
modelObject=elt,
version=version)
foundErrors = True
except (InvalidOperation, OverflowError):
modelXbrl.error("streamingExtensions:versionError",
_("Version %(version)s, number must be 1.n"),
modelObject=elt,
version=streamingAspects.get("version", "(none)"))
foundErrors = True
for bufAspect in ("contextBuffer", "unitBuffer", "footnoteBuffer"):
try:
bufLimit = Decimal(streamingAspects.get(bufAspect, "INF"))
if bufLimit < 1 or (bufLimit.is_finite() and bufLimit % 1 != 0):
raise InvalidOperation
elif bufAspect == "contextBuffer":
contextBufferLimit = bufLimit
elif bufAspect == "unitBuffer":
unitBufferLimit = bufLimit
elif bufAspect == "footnoteBuffer":
footnoteBufferLimit = bufLimit
except InvalidOperation:
modelXbrl.error(
"streamingExtensions:valueError",
_("Streaming %(attrib)s %(value)s, number must be a positive integer or INF"
),
modelObject=elt,
attrib=bufAspect,
value=streamingAspects.get(bufAspect))
foundErrors = True
if _streamingExtensionsValidate:
incompatibleValidations = []
_validateDisclosureSystem = modelXbrl.modelManager.validateDisclosureSystem
_disclosureSystem = modelXbrl.modelManager.disclosureSystem
if _validateDisclosureSystem and _disclosureSystem.validationType == "EFM":
incompatibleValidations.append("EFM")
if _validateDisclosureSystem and _disclosureSystem.validationType == "GFM":
incompatibleValidations.append("GFM")
if _validateDisclosureSystem and _disclosureSystem.validationType == "HMRC":
incompatibleValidations.append("HMRC")
if modelXbrl.modelManager.validateCalcLB:
incompatibleValidations.append("calculation LB")
if incompatibleValidations:
modelXbrl.error(
"streamingExtensions:incompatibleValidation",
_("Streaming instance validation does not support %(incompatibleValidations)s validation"
),
modelObject=modelXbrl,
incompatibleValidations=', '.join(incompatibleValidations))
foundErrors = True
if instInfoContext.error_log:
foundErrors = True
logSyntaxErrors(instInfoContext)
del instInfoContext # dereference
for pluginMethod in pluginClassMethods("Streaming.BlockStreaming"):
_blockingPluginName = pluginMethod(modelXbrl)
if _blockingPluginName: # name of blocking plugin is returned
modelXbrl.error(
"streamingExtensions:incompatiblePlugIn",
_("Streaming instance not supported by plugin %(blockingPlugin)s"
),
modelObject=modelXbrl,
blockingPlugin=_blockingPluginName)
foundErrors = True
if foundErrors:
_file.close()
return None
_encoding = XmlUtil.encoding(_file.read(512))
_file.seek(0, io.SEEK_SET) # allow reparsing
if _streamingExtensionsValidate:
validator = Validate(modelXbrl)
instValidator = validator.instValidator
contextBuffer = []
contextsToDrop = []
unitBuffer = []
unitsToDrop = []
footnoteBuffer = []
footnoteLinksToDrop = []
_streamingFactsPlugin = any(
True for pluginMethod in pluginClassMethods("Streaming.Facts"))
_streamingValidateFactsPlugin = (_streamingExtensionsValidate and any(
True
for pluginMethod in pluginClassMethods("Streaming.ValidateFacts")))
''' this is very much slower than iterparse
class modelLoaderTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.currentMdlObj = None
self.beforeInstanceStream = True
self.beforeStartStreamingPlugin = True
self.numRootFacts = 1
modelXbrl.makeelementParentModelObject = None
modelXbrl.isStreamingMode = True
self.factsCheckVersion = None
self.factsCheckMd5s = Md5Sum()
def start(self, tag, attrib, nsmap=None):
modelXbrl.makeelementParentModelObject = self.currentMdlObj # pass parent to makeelement for ModelObjectFactory
mdlObj = _parser.makeelement(tag, attrib=attrib, nsmap=nsmap)
mdlObj.sourceline = 1
if self.newTree:
self.newTree = False
self.currentMdlObj = mdlObj
modelDocument = ModelDocument(modelXbrl, Type.INSTANCE, mappedUri, filepath, mdlObj.getroottree())
modelXbrl.modelDocument = modelDocument # needed for incremental validation
mdlObj.init(modelDocument)
modelDocument.parser = _parser # needed for XmlUtil addChild's makeelement
modelDocument.parserLookupName = _parserLookupName
modelDocument.parserLookupClass = _parserLookupClass
modelDocument.xmlRootElement = mdlObj
modelDocument.schemaLocationElements.add(mdlObj)
modelDocument.documentEncoding = _encoding
modelDocument._creationSoftwareComment = creationSoftwareComment
modelXbrl.info("streamingExtensions:streaming",
_("Stream processing this instance."),
modelObject = modelDocument)
else:
self.currentMdlObj.append(mdlObj)
self.currentMdlObj = mdlObj
mdlObj._init()
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if (self.beforeInstanceStream and (
(ns == XbrlConst.link and ln not in ("schemaRef", "linkbaseRef")) or
(ns == XbrlConst.xbrli and ln in ("context", "unit")) or
(ns not in (XbrlConst.link, XbrlConst.xbrli)))):
self.beforeInstanceStream = False
if _streamingExtensionsValidate:
instValidator.validate(modelXbrl, modelXbrl.modelManager.formulaOptions.typedParameters(modelXbrl.prefixedNamespaces))
else: # need default dimensions
ValidateXbrlDimensions.loadDimensionDefaults(modelXbrl)
elif not self.beforeInstanceStream and self.beforeStartStreamingPlugin:
for pluginMethod in pluginClassMethods("Streaming.Start"):
pluginMethod(modelXbrl)
self.beforeStartStreamingPlugin = False
return mdlObj
def end(self, tag):
modelDocument = modelXbrl.modelDocument
mdlObj = self.currentMdlObj
parentMdlObj = mdlObj.getparent()
self.currentMdlObj = parentMdlObj
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if ns == XbrlConst.xbrli:
if ln == "context":
if mdlObj.get("sticky"):
del mdlObj.attrib["sticky"]
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
else:
if _streamingExtensionsValidate and len(contextBuffer) >= contextBufferLimit:
# drop before adding as dropped may have same id as added
cntx = contextBuffer.pop(0)
if _streamingValidateFactsPlugin:
contextsToDrop.append(cntx)
else:
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
cntx = None
#>>XmlValidate.validate(modelXbrl, mdlObj)
#>>modelDocument.contextDiscover(mdlObj)
if contextBufferLimit.is_finite():
contextBuffer.append(mdlObj)
if _streamingExtensionsValidate:
contextsToCheck = (mdlObj,)
instValidator.checkContexts(contextsToCheck)
if modelXbrl.hasXDT:
instValidator.checkContextsDimensions(contextsToCheck)
del contextsToCheck # dereference
elif ln == "unit":
if _streamingExtensionsValidate and len(unitBuffer) >= unitBufferLimit:
# drop before adding as dropped may have same id as added
unit = unitBuffer.pop(0)
if _streamingValidateFactsPlugin:
unitsToDrop.append(unit)
else:
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
unit = None
#>>XmlValidate.validate(modelXbrl, mdlObj)
#>>modelDocument.unitDiscover(mdlObj)
if unitBufferLimit.is_finite():
unitBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkUnits( (mdlObj,) )
elif ln == "xbrl": # end of document
# check remaining batched facts if any
if _streamingValidateFactsPlugin:
# plugin attempts to process batch of all root facts not yet processed (not just current one)
# finish any final batch of facts
if len(modelXbrl.facts) > 0:
factsToCheck = modelXbrl.facts.copy()
factsHaveBeenProcessed = True
# can block facts deletion if required data not yet available, such as numeric unit for DpmDB
for pluginMethod in pluginClassMethods("Streaming.ValidateFacts"):
if not pluginMethod(modelXbrl, factsToCheck):
factsHaveBeenProcessed = False
if factsHaveBeenProcessed:
for fact in factsToCheck:
dropFact(modelXbrl, fact, modelXbrl.facts)
del parentMdlObj[parentMdlObj.index(fact)]
for cntx in contextsToDrop:
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
for unit in unitsToDrop:
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
for footnoteLink in footnoteLinksToDrop:
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
fact = cntx = unit = footnoteLink = None
del contextsToDrop[:]
del unitsToDrop[:]
del footnoteLinksToDrop[:]
del factsToCheck
# check remaining footnote refs
for footnoteLink in footnoteBuffer:
checkFootnoteHrefs(modelXbrl, footnoteLink)
for pluginMethod in pluginClassMethods("Streaming.Finish"):
pluginMethod(modelXbrl)
elif ns == XbrlConst.link:
if ln == "footnoteLink":
XmlValidate.validate(modelXbrl, mdlObj)
footnoteLinks = (mdlObj,)
modelDocument.linkbaseDiscover(footnoteLinks, inInstance=True)
if footnoteBufferLimit.is_finite():
footnoteBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkLinks(footnoteLinks)
if len(footnoteBuffer) > footnoteBufferLimit:
# check that hrefObjects for locators were all satisfied
# drop before addition as dropped may have same id as added
footnoteLink = footnoteBuffer.pop(0)
checkFootnoteHrefs(modelXbrl, footnoteLink)
if _streamingValidateFactsPlugin:
footnoteLinksToDrop.append(footnoteLink)
else:
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
footnoteLink = None
footnoteLinks = None
elif ln in ("schemaRef", "linkbaseRef"):
modelDocument.discoverHref(mdlObj)
elif not modelXbrl.skipDTS:
if ln in ("roleRef", "arcroleRef"):
modelDocument.linkbaseDiscover((mdlObj,), inInstance=True)
elif parentMdlObj.qname == XbrlConst.qnXbrliXbrl:
self.numRootFacts += 1
#>>XmlValidate.validate(modelXbrl, mdlObj)
#>>modelDocument.factDiscover(mdlObj, modelXbrl.facts)
if self.factsCheckVersion:
self.factCheckFact(mdlObj)
if _streamingExtensionsValidate or _streamingValidateFactsPlugin:
factsToCheck = (mdlObj,) # validate current fact by itself
if _streamingExtensionsValidate:
instValidator.checkFacts(factsToCheck)
if modelXbrl.hasXDT:
instValidator.checkFactsDimensions(factsToCheck)
if _streamingValidateFactsPlugin:
# plugin attempts to process batch of all root facts not yet processed (not just current one)
# use batches of 1000 facts
if len(modelXbrl.facts) > 1000:
factsToCheck = modelXbrl.facts.copy()
factsHaveBeenProcessed = True
# can block facts deletion if required data not yet available, such as numeric unit for DpmDB
for pluginMethod in pluginClassMethods("Streaming.ValidateFacts"):
if not pluginMethod(modelXbrl, factsToCheck):
factsHaveBeenProcessed = False
if factsHaveBeenProcessed:
for fact in factsToCheck:
dropFact(modelXbrl, fact, modelXbrl.facts)
del parentMdlObj[parentMdlObj.index(fact)]
for cntx in contextsToDrop:
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
for unit in unitsToDrop:
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
for footnoteLink in footnoteLinksToDrop:
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
fact = cntx = unit = footnoteLink = None
del contextsToDrop[:]
del unitsToDrop[:]
del footnoteLinksToDrop[:]
del factsToCheck # dereference fact or batch of facts
else:
dropFact(modelXbrl, mdlObj, modelXbrl.facts) # single fact has been processed
del parentMdlObj[parentMdlObj.index(mdlObj)]
if self.numRootFacts % 1000 == 0:
pass
#modelXbrl.profileActivity("... streaming fact {0} of {1} {2:.2f}%".format(self.numRootFacts, instInfoNumRootFacts,
# 100.0 * self.numRootFacts / instInfoNumRootFacts),
# minTimeToShow=20.0)
gc.collect()
sys.stdout.write ("\rAt fact {} of {} mem {}".format(self.numRootFacts, instInfoNumRootFacts, modelXbrl.modelManager.cntlr.memoryUsed))
return mdlObj
def data(self, data):
self.currentMdlObj.text = data
def comment(self, text):
pass
def pi(self, target, data):
if target == "xbrl-facts-check":
_match = re.search("([\\w-]+)=[\"']([^\"']+)[\"']", data)
if _match:
_matchGroups = _match.groups()
if len(_matchGroups) == 2:
if _matchGroups[0] == "version":
self.factsCheckVersion = _matchGroups[1]
elif _matchGroups[0] == "sum-of-fact-md5s":
try:
expectedMd5 = Md5Sum(_matchGroups[1])
if self.factsCheckMd5s != expectedMd5:
modelXbrl.warning("streamingExtensions:xbrlFactsCheckWarning",
_("XBRL facts sum of md5s expected %(expectedMd5)s not matched to actual sum %(actualMd5Sum)s"),
modelObject=modelXbrl, expectedMd5=expectedMd5, actualMd5Sum=self.factsCheckMd5s)
else:
modelXbrl.info("info",
_("Successful XBRL facts sum of md5s."),
modelObject=modelXbrl)
except ValueError:
modelXbrl.error("streamingExtensions:xbrlFactsCheckError",
_("Invalid sum-of-md5s %(sumOfMd5)s"),
modelObject=modelXbrl, sumOfMd5=_matchGroups[1])
def close(self):
del modelXbrl.makeelementParentModelObject
return None
def factCheckFact(self, fact):
self.factsCheckMd5s += fact.md5sum
for _tupleFact in fact.modelTupleFacts:
self.factCheckFact(_tupleFact)
_parser, _parserLookupName, _parserLookupClass = parser(modelXbrl, filepath, target=modelLoaderTarget())
etree.parse(_file, parser=_parser, base_url=filepath)
logSyntaxErrors(_parser)
'''
# replace modelLoaderTarget with iterparse (as it now supports CustomElementClassLookup)
streamingParserContext = etree.iterparse(_file,
events=("start", "end"),
huge_tree=True)
from arelle.ModelObjectFactory import setParserElementClassLookup
modelXbrl.isStreamingMode = True # must be set before setting element class lookup
(_parser, _parserLookupName,
_parserLookupClass) = setParserElementClassLookup(streamingParserContext,
modelXbrl)
foundInstance = False
beforeInstanceStream = beforeStartStreamingPlugin = True
numRootFacts = 0
factsCheckVersion = None
def factCheckFact(fact):
modelDocument._factsCheckMd5s += fact.md5sum
for _tupleFact in fact.modelTupleFacts:
factCheckFact(_tupleFact)
for event, mdlObj in streamingParserContext:
if event == "start":
if mdlObj.tag == "{http://www.xbrl.org/2003/instance}xbrl":
modelDocument = ModelDocument(modelXbrl, Type.INSTANCE,
mappedUri, filepath,
mdlObj.getroottree())
modelXbrl.modelDocument = modelDocument # needed for incremental validation
mdlObj.init(modelDocument)
modelDocument.parser = _parser # needed for XmlUtil addChild's makeelement
modelDocument.parserLookupName = _parserLookupName
modelDocument.parserLookupClass = _parserLookupClass
modelDocument.xmlRootElement = mdlObj
modelDocument.schemaLocationElements.add(mdlObj)
modelDocument.documentEncoding = _encoding
modelDocument._creationSoftwareComment = precedingComment(
mdlObj)
modelDocument._factsCheckMd5s = Md5Sum()
modelXbrl.info("streamingExtensions:streaming",
_("Stream processing this instance."),
modelObject=modelDocument)
elif mdlObj.getparent() is not None:
mdlObj._init() # requires discovery as part of start elements
if mdlObj.getparent(
).tag == "{http://www.xbrl.org/2003/instance}xbrl":
if not foundInstance:
foundInstance = True
pi = precedingProcessingInstruction(
mdlObj, "xbrl-facts-check")
if pi is not None:
factsCheckVersion = pi.attrib.get("version", None)
elif not foundInstance:
break
ns = mdlObj.qname.namespaceURI
ln = mdlObj.qname.localName
if beforeInstanceStream:
if ((ns == XbrlConst.link
and ln not in ("schemaRef", "linkbaseRef")) or
(ns == XbrlConst.xbrli and ln in ("context", "unit"))
or (ns not in (XbrlConst.link, XbrlConst.xbrli))):
beforeInstanceStream = False
if _streamingExtensionsValidate:
instValidator.validate(
modelXbrl,
modelXbrl.modelManager.formulaOptions.
typedParameters(modelXbrl.prefixedNamespaces))
else: # need default dimensions
ValidateXbrlDimensions.loadDimensionDefaults(
modelXbrl)
elif not beforeInstanceStream and beforeStartStreamingPlugin:
for pluginMethod in pluginClassMethods("Streaming.Start"):
pluginMethod(modelXbrl)
beforeStartStreamingPlugin = False
elif event == "end":
parentMdlObj = mdlObj.getparent()
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if ns == XbrlConst.xbrli:
if ln == "context":
if mdlObj.get("sticky"):
del mdlObj.attrib["sticky"]
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
else:
if len(contextBuffer) >= contextBufferLimit:
# drop before adding as dropped may have same id as added
cntx = contextBuffer.pop(0)
if _streamingFactsPlugin or _streamingValidateFactsPlugin:
contextsToDrop.append(cntx)
else:
dropContext(modelXbrl, cntx)
#>>del parentMdlObj[parentMdlObj.index(cntx)]
cntx = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
if contextBufferLimit.is_finite():
contextBuffer.append(mdlObj)
if _streamingExtensionsValidate:
contextsToCheck = (mdlObj, )
instValidator.checkContexts(contextsToCheck)
if modelXbrl.hasXDT:
instValidator.checkContextsDimensions(
contextsToCheck)
del contextsToCheck # dereference
elif ln == "unit":
if len(unitBuffer) >= unitBufferLimit:
# drop before additing as dropped may have same id as added
unit = unitBuffer.pop(0)
if _streamingFactsPlugin or _streamingValidateFactsPlugin:
unitsToDrop.append(unit)
else:
dropUnit(modelXbrl, unit)
#>>del parentMdlObj[parentMdlObj.index(unit)]
unit = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.unitDiscover(mdlObj)
if unitBufferLimit.is_finite():
unitBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkUnits((mdlObj, ))
elif ln == "xbrl": # end of document
# check remaining batched facts if any
if _streamingFactsPlugin or _streamingValidateFactsPlugin:
# plugin attempts to process batch of all root facts not yet processed (not just current one)
# finish any final batch of facts
if len(modelXbrl.facts) > 0:
factsToCheck = modelXbrl.facts.copy()
# can block facts deletion if required data not yet available, such as numeric unit for DpmDB
if _streamingValidateFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.ValidateFacts"):
pluginMethod(instValidator, factsToCheck)
if _streamingFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.Facts"):
pluginMethod(modelXbrl, factsToCheck)
for fact in factsToCheck:
dropFact(modelXbrl, fact, modelXbrl.facts)
#>>del parentMdlObj[parentMdlObj.index(fact)]
for cntx in contextsToDrop:
dropContext(modelXbrl, cntx)
#>>del parentMdlObj[parentMdlObj.index(cntx)]
for unit in unitsToDrop:
dropUnit(modelXbrl, unit)
#>>del parentMdlObj[parentMdlObj.index(unit)]
for footnoteLink in footnoteLinksToDrop:
dropFootnoteLink(modelXbrl, footnoteLink)
#>>del parentMdlObj[parentMdlObj.index(footnoteLink)]
fact = cntx = unit = footnoteLink = None
del contextsToDrop[:]
del unitsToDrop[:]
del footnoteLinksToDrop[:]
del factsToCheck
# check remaining footnote refs
for footnoteLink in footnoteBuffer:
checkFootnoteHrefs(modelXbrl, footnoteLink)
pi = childProcessingInstruction(mdlObj,
"xbrl-facts-check",
reversed=True)
if pi is not None: # attrib is in .text, not attrib, no idea why!!!
_match = re.search("([\\w-]+)=[\"']([^\"']+)[\"']",
pi.text)
if _match:
_matchGroups = _match.groups()
if len(_matchGroups) == 2:
if _matchGroups[0] == "sum-of-fact-md5s":
try:
expectedMd5 = Md5Sum(_matchGroups[1])
if modelDocument._factsCheckMd5s != expectedMd5:
modelXbrl.warning(
"streamingExtensions:xbrlFactsCheckWarning",
_("XBRL facts sum of md5s expected %(expectedMd5)s not matched to actual sum %(actualMd5Sum)s"
),
modelObject=modelXbrl,
expectedMd5=expectedMd5,
actualMd5Sum=modelDocument.
_factsCheckMd5s)
else:
modelXbrl.info(
"info",
_("Successful XBRL facts sum of md5s."
),
modelObject=modelXbrl)
except ValueError:
modelXbrl.error(
"streamingExtensions:xbrlFactsCheckError",
_("Invalid sum-of-md5s %(sumOfMd5)s"
),
modelObject=modelXbrl,
sumOfMd5=_matchGroups[1])
if _streamingValidateFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.ValidateFinish"):
pluginMethod(instValidator)
if _streamingFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.Finish"):
pluginMethod(modelXbrl)
elif ns == XbrlConst.link:
if ln in ("schemaRef", "linkbaseRef"):
modelDocument.discoverHref(
mdlObj,
urlRewritePluginClass=
"ModelDocument.InstanceSchemaRefRewriter")
elif ln in ("roleRef", "arcroleRef"):
modelDocument.linkbaseDiscover((mdlObj, ), inInstance=True)
elif ln == "footnoteLink":
XmlValidate.validate(modelXbrl, mdlObj)
footnoteLinks = (mdlObj, )
modelDocument.linkbaseDiscover(footnoteLinks,
inInstance=True)
if footnoteBufferLimit.is_finite():
footnoteBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkLinks(footnoteLinks)
if len(footnoteBuffer) > footnoteBufferLimit:
# check that hrefObjects for locators were all satisfied
# drop before addition as dropped may have same id as added
footnoteLink = footnoteBuffer.pop(0)
checkFootnoteHrefs(modelXbrl, footnoteLink)
if _streamingValidateFactsPlugin:
footnoteLinksToDrop.append(footnoteLink)
else:
dropFootnoteLink(modelXbrl, footnoteLink)
#>>del parentMdlObj[parentMdlObj.index(footnoteLink)]
footnoteLink = None
footnoteLinks = None
elif parentMdlObj.qname == XbrlConst.qnXbrliXbrl and isinstance(
mdlObj, ModelFact):
numRootFacts += 1
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.factDiscover(mdlObj, modelXbrl.facts)
if factsCheckVersion:
factCheckFact(mdlObj)
if _streamingExtensionsValidate or _streamingFactsPlugin or _streamingValidateFactsPlugin:
factsToCheck = (mdlObj,
) # validate current fact by itself
if _streamingExtensionsValidate:
instValidator.checkFacts(factsToCheck)
if modelXbrl.hasXDT:
instValidator.checkFactsDimensions(factsToCheck)
if _streamingFactsPlugin or _streamingValidateFactsPlugin:
# plugin attempts to process batch of all root facts not yet processed (not just current one)
# use batches of 1000 facts
if len(modelXbrl.facts) > 1000:
factsToCheck = modelXbrl.facts.copy()
# can block facts deletion if required data not yet available, such as numeric unit for DpmDB
if _streamingValidateFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.ValidateFacts"):
pluginMethod(instValidator, factsToCheck)
if _streamingFactsPlugin:
for pluginMethod in pluginClassMethods(
"Streaming.Facts"):
pluginMethod(modelXbrl, factsToCheck)
for fact in factsToCheck:
dropFact(modelXbrl, fact, modelXbrl.facts)
#>>del parentMdlObj[parentMdlObj.index(fact)]
for cntx in contextsToDrop:
dropContext(modelXbrl, cntx)
#>>del parentMdlObj[parentMdlObj.index(cntx)]
for unit in unitsToDrop:
dropUnit(modelXbrl, unit)
#>>del parentMdlObj[parentMdlObj.index(unit)]
for footnoteLink in footnoteLinksToDrop:
dropFootnoteLink(modelXbrl, footnoteLink)
#>>del parentMdlObj[parentMdlObj.index(footnoteLink)]
fact = cntx = unit = footnoteLink = None
del contextsToDrop[:]
del unitsToDrop[:]
del footnoteLinksToDrop[:]
del factsToCheck # dereference fact or batch of facts
else:
dropFact(
modelXbrl, mdlObj,
modelXbrl.facts) # single fact has been processed
#>>del parentMdlObj[parentMdlObj.index(mdlObj)]
if numRootFacts % 1000 == 0:
pass
#modelXbrl.profileActivity("... streaming fact {0} of {1} {2:.2f}%".format(self.numRootFacts, instInfoNumRootFacts,
# 100.0 * self.numRootFacts / instInfoNumRootFacts),
# minTimeToShow=20.0)
#gc.collect()
#sys.stdout.write ("\rAt fact {} of {} mem {}".format(numRootFacts, instInfoNumRootFacts, modelXbrl.modelManager.cntlr.memoryUsed))
if mdlObj is not None:
mdlObj.clear()
del _parser, _parserLookupName, _parserLookupClass
if _streamingExtensionsValidate and validator is not None:
_file.close()
del instValidator
validator.close()
# track that modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = True
modelXbrl.profileStat(_("streaming complete"), time.time() - startedAt)
return modelXbrl.modelDocument
def streamingExtensionsLoader(modelXbrl, mappedUri, filepath):
# check if big instance and has header with an initial incomplete tree walk (just 2 elements
def logSyntaxErrors(parsercontext):
for error in parsercontext.error_log:
modelXbrl.error("xmlSchema:syntax",
_("%(error)s, %(fileName)s, line %(line)s, column %(column)s, %(sourceAction)s source element"),
modelObject=modelDocument, fileName=os.path.basename(filepath),
error=error.message, line=error.line, column=error.column, sourceAction="streaming")
#### note: written for iterparse of lxml prior to version 3.3, otherwise rewrite to use XmlPullParser ###
#### note: iterparse wants a binary file, but file is text mode
_file, = modelXbrl.fileSource.file(filepath, binary=True)
startedAt = time.time()
modelXbrl.profileActivity()
parsercontext = etree.iterparse(_file, events=("start","end"), huge_tree=True)
foundInstance = False
foundErrors = False
streamingAspects = None
numRootFacts1 = 0
numElts = 0
elt = None
for event, elt in parsercontext:
if event == "start":
if elt.getparent() is not None:
if elt.getparent().tag == "{http://www.xbrl.org/2003/instance}xbrl":
if not foundInstance:
foundInstance = True
pi = precedingProcessingInstruction(elt, "xbrl-streamable-instance")
if pi is None:
break
else:
streamingAspects = dict(pi.attrib.copy())
if not elt.tag.startswith("{http://www.xbrl.org/"):
numRootFacts1 += 1
if numRootFacts1 % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
elif not foundInstance:
break
elif elt.tag == "{http://www.xbrl.org/2003/instance}xbrl" and precedingProcessingInstruction(elt, "xbrl-streamable-instance") is not None:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(error)s, must follow xbrli:xbrl element"),
modelObject=elt)
elif event == "end":
elt.clear()
numElts += 1
if numElts % 1000 == 0 and elt.getparent() is not None:
while elt.getprevious() is not None and elt.getparent() is not None:
del elt.getparent()[0]
if elt is not None:
elt.clear()
_file.seek(0,io.SEEK_SET) # allow reparsing
if not foundInstance or streamingAspects is None:
del elt, parsercontext
_file.close()
return None
modelXbrl.profileStat(_("streaming tree check"), time.time() - startedAt)
startedAt = time.time()
try:
version = Decimal(streamingAspects.get("version"))
if int(version) != 1:
modelXbrl.error("streamingExtensions:unsupportedVersion",
_("Streaming version %(version)s, major version number must be 1"),
modelObject=elt, version=version)
foundErrors = True
except (InvalidOperation, OverflowError):
modelXbrl.error("streamingExtensions:versionError",
_("Version %(version)s, number must be 1.n"),
modelObject=elt, version=streamingAspects.get("version", "(none)"))
foundErrors = True
for bufAspect in ("contextBuffer", "unitBuffer", "footnoteBuffer"):
try:
bufLimit = Decimal(streamingAspects.get(bufAspect, "INF"))
if bufLimit < 1 or (bufLimit.is_finite() and bufLimit % 1 != 0):
raise InvalidOperation
elif bufAspect == "contextBuffer":
contextBufferLimit = bufLimit
elif bufAspect == "unitBuffer":
unitBufferLimit = bufLimit
elif bufAspect == "footnoteBuffer":
footnoteBufferLimit = bufLimit
except InvalidOperation:
modelXbrl.error("streamingExtensions:valueError",
_("Streaming %(attrib)s %(value)s, number must be a positive integer or INF"),
modelObject=elt, attrib=bufAspect, value=streamingAspects.get(bufAspect))
foundErrors = True
if parsercontext.error_log:
foundErrors = True
logSyntaxErrors(parsercontext)
if foundErrors:
_file.close()
return None
parsercontext = etree.iterparse(_file, events=("start","end"), huge_tree=True)
_parser, _parserLookupName, _parserLookupClass = parser(modelXbrl,filepath)
eltMdlObjs = {}
beforeInstanceStream = True
validator = None
contextBuffer = []
unitBuffer = []
footnoteBuffer = []
factBuffer = []
numFacts = numRootFacts2 = 1
for event, elt in parsercontext:
if event == "start":
mdlObj = _parser.makeelement(elt.tag, attrib=elt.attrib, nsmap=elt.nsmap)
mdlObj.sourceline = elt.sourceline
eltMdlObjs[elt] = mdlObj
if elt.getparent() is None:
modelDocument = ModelDocument(modelXbrl, Type.INSTANCE, mappedUri, filepath, etree.ElementTree(mdlObj))
modelDocument.xmlRootElement = mdlObj
modelXbrl.modelDocument = modelDocument # needed for incremental validation
mdlObj.init(modelDocument)
modelXbrl.info("streamingExtensions:streaming",
_("Stream processing this instance."),
modelObject = modelDocument)
else:
eltMdlObjs[elt.getparent()].append(mdlObj)
mdlObj._init()
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if (beforeInstanceStream and (
(ns == XbrlConst.link and ln not in ("schemaRef", "linkbaseRef")) or
(ns == XbrlConst.xbrli and ln in ("context", "unit")) or
(ns not in (XbrlConst.link, XbrlConst.xbrli)))):
beforeInstanceStream = False
if _streamingExtensionsValidate:
validator = Validate(modelXbrl)
validator.instValidator.validate(modelXbrl, modelXbrl.modelManager.formulaOptions.typedParameters())
else: # need default dimensions
ValidateXbrlDimensions.loadDimensionDefaults(modelXbrl)
mdlObj = None # deref
elif event == "end":
mdlObj = eltMdlObjs.pop(elt)
if elt.text: # text available after child nodes processed
mdlObj.text = elt.text
ns = mdlObj.namespaceURI
ln = mdlObj.localName
parentMdlObj = mdlObj.getparent()
if ns == XbrlConst.xbrli:
if ln == "context":
if mdlObj.get("sticky"):
del mdlObj.attrib["sticky"]
modelDocument.contextDiscover(mdlObj)
else:
if _streamingExtensionsValidate and len(contextBuffer) >= contextBufferLimit:
# drop before adding as dropped may have same id as added
cntx = contextBuffer.pop(0)
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
cntx = None
modelDocument.contextDiscover(mdlObj)
if contextBufferLimit.is_finite():
contextBuffer.append(mdlObj)
if _streamingExtensionsValidate:
contextsToCheck = (mdlObj,)
validator.instValidator.checkContexts(contextsToCheck)
if modelXbrl.hasXDT:
validator.instValidator.checkContextsDimensions(contextsToCheck)
del contextsToCheck # dereference
elif ln == "unit":
if _streamingExtensionsValidate and len(unitBuffer) >= unitBufferLimit:
# drop before additing as dropped may have same id as added
unit = unitBuffer.pop(0)
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
unit = None
modelDocument.unitDiscover(mdlObj)
if unitBufferLimit.is_finite():
unitBuffer.append(mdlObj)
if _streamingExtensionsValidate:
validator.instValidator.checkUnits( (mdlObj,) )
elif ln == "xbrl": # end of document
# check remaining footnote refs
for footnoteLink in footnoteBuffer:
checkFootnoteHrefs(modelXbrl, footnoteLink)
elt.clear()
elif ns == XbrlConst.link:
if ln in ("schemaRef", "linkbaseRef"):
modelDocument.discoverHref(mdlObj)
elif ln in ("roleRef", "arcroleRef"):
modelDocument.linkbaseDiscover((mdlObj,), inInstance=True)
elif ln == "footnoteLink":
footnoteLinks = (mdlObj,)
modelDocument.linkbaseDiscover(footnoteLinks, inInstance=True)
if footnoteBufferLimit.is_finite():
footnoteBuffer.append(mdlObj)
if _streamingExtensionsValidate:
validator.instValidator.checkLinks(footnoteLinks)
if len(footnoteBuffer) > footnoteBufferLimit:
# check that hrefObjects for locators were all satisfied
# drop before addition as dropped may have same id as added
footnoteLink = footnoteBuffer.pop(0)
checkFootnoteHrefs(modelXbrl, footnoteLink)
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
footnoteLink = None
footnoteLinks = None
elt.clear()
elif parentMdlObj.qname == XbrlConst.qnXbrliXbrl:
numRootFacts2 += 1
modelDocument.factDiscover(mdlObj, modelXbrl.facts)
XmlValidate.validate(modelXbrl, mdlObj)
if _streamingExtensionsValidate:
factsToCheck = (mdlObj,)
validator.instValidator.checkFacts(factsToCheck)
if modelXbrl.hasXDT:
validator.instValidator.checkFactsDimensions(factsToCheck)
del factsToCheck
dropFact(modelXbrl, mdlObj, modelXbrl.facts)
del parentMdlObj[parentMdlObj.index(mdlObj)]
if numRootFacts2 % 1000 == 0:
modelXbrl.profileActivity("... streaming fact {0} of {1} {2:.2f}%".format(numRootFacts2, numRootFacts1, 100.0 * numRootFacts2 / numRootFacts1),
minTimeToShow=20.0)
# get rid of root element from iterparse's tree
elt.clear()
while elt.getprevious() is not None: # cleans up any prior siblings
del elt.getparent()[0]
mdlObj = None # deref
logSyntaxErrors(parsercontext)
del parsercontext
if validator is not None:
validator.close()
_file.close()
modelXbrl.profileStat(_("streaming complete"), time.time() - startedAt)
return modelDocument
def streamingExtensionsLoader(modelXbrl, mappedUri, filepath, **kwargs):
# check if big instance and has header with an initial incomplete tree walk (just 2 elements
if not _streamingExtensionsCheck:
return None
# track whether modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = False
def logSyntaxErrors(parsercontext):
for error in parsercontext.error_log:
modelXbrl.error("xmlSchema:syntax",
_("%(error)s, %(fileName)s, line %(line)s, column %(column)s, %(sourceAction)s source element"),
modelObject=modelXbrl, fileName=os.path.basename(filepath),
error=error.message, line=error.line, column=error.column, sourceAction="streaming")
#### note: written for iterparse of lxml prior to version 3.3, otherwise rewrite to use XmlPullParser ###
#### note: iterparse wants a binary file, but file is text mode
_file, = modelXbrl.fileSource.file(filepath, binary=True)
startedAt = time.time()
modelXbrl.profileActivity()
''' this seems twice as slow as iterparse
class instInfoTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.streamingAspects = None
self.foundInstance = False
self.creationSoftwareComment = ''
self.currentEltTag = "(before xbrli:xbrl)"
self.numRootFacts = 0
def start(self, tag, attrib, nsmap=None):
if self.newTree:
if tag == "{http://www.xbrl.org/2003/instance}xbrl":
self.foundInstance = True
self.newTree = False
else: # break
raise NotInstanceDocumentException()
elif not tag.startswith("{http://www.xbrl.org/"):
self.numRootFacts += 1
if self.numRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
self.currentEltTag = tag
def end(self, tag):
pass
def data(self, data):
pass
def comment(self, text):
if not self.foundInstance: # accumulate comments before xbrli:xbrl
self.creationSoftwareComment += ('\n' if self.creationSoftwareComment else '') + text
elif not self.creationSoftwareComment:
self.creationSoftwareComment = text # or first comment after xbrli:xbrl
def pi(self, target, data):
if target == "xbrl-streamable-instance":
if self.currentEltTag == "{http://www.xbrl.org/2003/instance}xbrl":
self.streamingAspects = dict(etree.PI(target,data).attrib.copy()) # dereference target results
else:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(target)s, must follow xbrli:xbrl element but was found at %(element)s"),
modelObject=modelXbrl, target=target, element=self.currentEltTag)
def close(self):
if not self.creationSoftwareComment:
self.creationSoftwareComment = None
return True
instInfo = instInfoTarget()
infoParser = etree.XMLParser(recover=True, huge_tree=True, target=instInfo)
try:
etree.parse(_file, parser=infoParser, base_url=filepath)
except NotInstanceDocumentException:
pass
'''
foundErrors = False
foundInstance = False
streamingAspects = None
creationSoftwareComment = None
instInfoNumRootFacts = 0
numElts = 0
elt = None
instInfoContext = etree.iterparse(_file, events=("start","end"), huge_tree=True)
for event, elt in instInfoContext:
if event == "start":
if elt.getparent() is not None:
if elt.getparent().tag == "{http://www.xbrl.org/2003/instance}xbrl":
if not foundInstance:
foundInstance = True
pi = precedingProcessingInstruction(elt, "xbrl-streamable-instance")
if pi is None:
break
else:
streamingAspects = dict(pi.attrib.copy())
if creationSoftwareComment is None:
creationSoftwareComment = precedingComment(elt)
if not elt.tag.startswith("{http://www.xbrl.org/"):
instInfoNumRootFacts += 1
if instInfoNumRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
elif not foundInstance:
break
elif elt.tag == "{http://www.xbrl.org/2003/instance}xbrl":
creationSoftwareComment = precedingComment(elt)
if precedingProcessingInstruction(elt, "xbrl-streamable-instance") is not None:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(error)s, must follow xbrli:xbrl element"),
modelObject=elt)
elif event == "end":
elt.clear()
numElts += 1
if numElts % 1000 == 0 and elt.getparent() is not None:
while elt.getprevious() is not None and elt.getparent() is not None:
del elt.getparent()[0]
if elt is not None:
elt.clear()
_file.seek(0,io.SEEK_SET) # allow reparsing
if not foundInstance or streamingAspects is None:
del elt
_file.close()
return None
modelXbrl.profileStat(_("streaming tree check"), time.time() - startedAt)
startedAt = time.time()
try:
version = Decimal(streamingAspects.get("version"))
if int(version) != 1:
modelXbrl.error("streamingExtensions:unsupportedVersion",
_("Streaming version %(version)s, major version number must be 1"),
modelObject=elt, version=version)
foundErrors = True
except (InvalidOperation, OverflowError):
modelXbrl.error("streamingExtensions:versionError",
_("Version %(version)s, number must be 1.n"),
modelObject=elt, version=streamingAspects.get("version", "(none)"))
foundErrors = True
for bufAspect in ("contextBuffer", "unitBuffer", "footnoteBuffer"):
try:
bufLimit = Decimal(streamingAspects.get(bufAspect, "INF"))
if bufLimit < 1 or (bufLimit.is_finite() and bufLimit % 1 != 0):
raise InvalidOperation
elif bufAspect == "contextBuffer":
contextBufferLimit = bufLimit
elif bufAspect == "unitBuffer":
unitBufferLimit = bufLimit
elif bufAspect == "footnoteBuffer":
footnoteBufferLimit = bufLimit
except InvalidOperation:
modelXbrl.error("streamingExtensions:valueError",
_("Streaming %(attrib)s %(value)s, number must be a positive integer or INF"),
modelObject=elt, attrib=bufAspect, value=streamingAspects.get(bufAspect))
foundErrors = True
if _streamingExtensionsValidate:
incompatibleValidations = []
_validateDisclosureSystem = modelXbrl.modelManager.validateDisclosureSystem
_disclosureSystem = modelXbrl.modelManager.disclosureSystem
if _validateDisclosureSystem and _disclosureSystem.EFM:
incompatibleValidations.append("EFM")
if _validateDisclosureSystem and _disclosureSystem.GFM:
incompatibleValidations.append("GFM")
if _validateDisclosureSystem and _disclosureSystem.EBA:
incompatibleValidations.append("EBA")
if _validateDisclosureSystem and _disclosureSystem.HMRC:
incompatibleValidations.append("EBA")
if modelXbrl.modelManager.validateCalcLB:
incompatibleValidations.append("calculation LB")
if incompatibleValidations:
modelXbrl.error("streamingExtensions:incompatibleValidation",
_("Streaming instance validation does not support %(incompatibleValidations)s validation"),
modelObject=modelXbrl, incompatibleValidations=', '.join(incompatibleValidations))
foundErrors = True
if instInfoContext.error_log:
foundErrors = True
logSyntaxErrors(instInfoContext)
del instInfoContext # dereference
for pluginMethod in pluginClassMethods("Streaming.BlockStreaming"):
_blockingPluginName = pluginMethod(modelXbrl)
if _blockingPluginName: # name of blocking plugin is returned
modelXbrl.error("streamingExtensions:incompatiblePlugIn",
_("Streaming instance not supported by plugin %(blockingPlugin)s"),
modelObject=modelXbrl, blockingPlugin=_blockingPluginName)
foundErrors = True
if foundErrors:
_file.close()
return None
_encoding = XmlUtil.encoding(_file.read(512))
_file.seek(0,io.SEEK_SET) # allow reparsing
if _streamingExtensionsValidate:
validator = Validate(modelXbrl)
instValidator = validator.instValidator
eltMdlObjs = {}
contextBuffer = []
unitBuffer = []
footnoteBuffer = []
factBuffer = []
numFacts = 1
_streamingValidateFactsPlugin = any(True for pluginMethod in pluginClassMethods("Streaming.ValidateFacts"))
class modelLoaderTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.currentMdlObj = None
self.beforeInstanceStream = True
self.beforeStartStreamingPlugin = True
self.numRootFacts = 1
modelXbrl.streamingParentModelObject = None
modelXbrl.isStreamingMode = True
def start(self, tag, attrib, nsmap=None):
modelXbrl.streamingParentModelObject = self.currentMdlObj # pass parent to makeelement for ModelObjectFactory
mdlObj = _parser.makeelement(tag, attrib=attrib, nsmap=nsmap)
mdlObj.sourceline = 1
if self.newTree:
self.newTree = False
self.currentMdlObj = mdlObj
modelDocument = ModelDocument(modelXbrl, Type.INSTANCE, mappedUri, filepath, mdlObj.getroottree())
modelXbrl.modelDocument = modelDocument # needed for incremental validation
mdlObj.init(modelDocument)
modelDocument.parser = _parser # needed for XmlUtil addChild's makeelement
modelDocument.parserLookupName = _parserLookupName
modelDocument.parserLookupClass = _parserLookupClass
modelDocument.xmlRootElement = mdlObj
modelDocument.schemaLocationElements.add(mdlObj)
modelDocument.documentEncoding = _encoding
modelDocument._creationSoftwareComment = creationSoftwareComment
modelXbrl.info("streamingExtensions:streaming",
_("Stream processing this instance."),
modelObject = modelDocument)
else:
self.currentMdlObj.append(mdlObj)
self.currentMdlObj = mdlObj
mdlObj._init()
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if (self.beforeInstanceStream and (
(ns == XbrlConst.link and ln not in ("schemaRef", "linkbaseRef")) or
(ns == XbrlConst.xbrli and ln in ("context", "unit")) or
(ns not in (XbrlConst.link, XbrlConst.xbrli)))):
self.beforeInstanceStream = False
if _streamingExtensionsValidate:
instValidator.validate(modelXbrl, modelXbrl.modelManager.formulaOptions.typedParameters())
else: # need default dimensions
ValidateXbrlDimensions.loadDimensionDefaults(modelXbrl)
elif not self.beforeInstanceStream and self.beforeStartStreamingPlugin:
for pluginMethod in pluginClassMethods("Streaming.Start"):
pluginMethod(modelXbrl)
self.beforeStartStreamingPlugin = False
return mdlObj
def end(self, tag):
modelDocument = modelXbrl.modelDocument
mdlObj = self.currentMdlObj
parentMdlObj = mdlObj.getparent()
self.currentMdlObj = parentMdlObj
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if ns == XbrlConst.xbrli:
if ln == "context":
if mdlObj.get("sticky"):
del mdlObj.attrib["sticky"]
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
else:
if _streamingExtensionsValidate and len(contextBuffer) >= contextBufferLimit:
# drop before adding as dropped may have same id as added
cntx = contextBuffer.pop(0)
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
cntx = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
if contextBufferLimit.is_finite():
contextBuffer.append(mdlObj)
if _streamingExtensionsValidate:
contextsToCheck = (mdlObj,)
instValidator.checkContexts(contextsToCheck)
if modelXbrl.hasXDT:
instValidator.checkContextsDimensions(contextsToCheck)
del contextsToCheck # dereference
elif ln == "unit":
if _streamingExtensionsValidate and len(unitBuffer) >= unitBufferLimit:
# drop before additing as dropped may have same id as added
unit = unitBuffer.pop(0)
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
unit = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.unitDiscover(mdlObj)
if unitBufferLimit.is_finite():
unitBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkUnits( (mdlObj,) )
elif ln == "xbrl": # end of document
# check remaining footnote refs
for footnoteLink in footnoteBuffer:
checkFootnoteHrefs(modelXbrl, footnoteLink)
for pluginMethod in pluginClassMethods("Streaming.Finish"):
pluginMethod(modelXbrl)
elif ns == XbrlConst.link:
if ln == "footnoteLink":
XmlValidate.validate(modelXbrl, mdlObj)
footnoteLinks = (mdlObj,)
modelDocument.linkbaseDiscover(footnoteLinks, inInstance=True)
if footnoteBufferLimit.is_finite():
footnoteBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkLinks(footnoteLinks)
if len(footnoteBuffer) > footnoteBufferLimit:
# check that hrefObjects for locators were all satisfied
# drop before addition as dropped may have same id as added
footnoteLink = footnoteBuffer.pop(0)
checkFootnoteHrefs(modelXbrl, footnoteLink)
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
footnoteLink = None
footnoteLinks = None
elif ln in ("schemaRef", "linkbaseRef"):
modelDocument.discoverHref(mdlObj)
elif not modelXbrl.skipDTS:
if ln in ("roleRef", "arcroleRef"):
modelDocument.linkbaseDiscover((mdlObj,), inInstance=True)
elif parentMdlObj.qname == XbrlConst.qnXbrliXbrl:
self.numRootFacts += 1
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.factDiscover(mdlObj, modelXbrl.facts)
if _streamingExtensionsValidate or _streamingValidateFactsPlugin:
factsToCheck = (mdlObj,) # validate current fact by itself
if _streamingExtensionsValidate:
instValidator.checkFacts(factsToCheck)
if modelXbrl.hasXDT:
instValidator.checkFactsDimensions(factsToCheck)
if _streamingValidateFactsPlugin:
# plugin attempts to process batch of all root facts not yet processed (not just current one)
factsToCheck = modelXbrl.facts.copy()
factsHaveBeenProcessed = True
# can block facts deletion if required data not yet available, such as numeric unit for DpmDB
for pluginMethod in pluginClassMethods("Streaming.ValidateFacts"):
if not pluginMethod(modelXbrl, factsToCheck):
factsHaveBeenProcessed = False
if factsHaveBeenProcessed:
for fact in factsToCheck:
dropFact(modelXbrl, fact, modelXbrl.facts)
del parentMdlObj[parentMdlObj.index(fact)]
else:
dropFact(modelXbrl, mdlObj, modelXbrl.facts) # single fact has been processed
del parentMdlObj[parentMdlObj.index(mdlObj)]
del factsToCheck # dereference fact or batch of facts
if self.numRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming fact {0} of {1} {2:.2f}%".format(self.numRootFacts, instInfoNumRootFacts,
100.0 * self.numRootFacts / instInfoNumRootFacts),
minTimeToShow=20.0)
return mdlObj
def data(self, data):
self.currentMdlObj.text = data
def comment(self, text):
pass
def pi(self, target, data):
pass
def close(self):
del modelXbrl.streamingParentModelObject
return None
_parser, _parserLookupName, _parserLookupClass = parser(modelXbrl, filepath, target=modelLoaderTarget())
etree.parse(_file, parser=_parser, base_url=filepath)
logSyntaxErrors(_parser)
if _streamingExtensionsValidate and validator is not None:
_file.close()
del instValidator
validator.close()
# track that modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = True
modelXbrl.profileStat(_("streaming complete"), time.time() - startedAt)
return modelXbrl.modelDocument
def streamingExtensionsLoader(modelXbrl, mappedUri, filepath, **kwargs):
# check if big instance and has header with an initial incomplete tree walk (just 2 elements
if not _streamingExtensionsCheck:
return None
# track whether modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = False
def logSyntaxErrors(parsercontext):
for error in parsercontext.error_log:
modelXbrl.error("xmlSchema:syntax",
_("%(error)s, %(fileName)s, line %(line)s, column %(column)s, %(sourceAction)s source element"),
modelObject=modelXbrl, fileName=os.path.basename(filepath),
error=error.message, line=error.line, column=error.column, sourceAction="streaming")
#### note: written for iterparse of lxml prior to version 3.3, otherwise rewrite to use XmlPullParser ###
#### note: iterparse wants a binary file, but file is text mode
_file, = modelXbrl.fileSource.file(filepath, binary=True)
startedAt = time.time()
modelXbrl.profileActivity()
''' this seems twice as slow as iterparse
class instInfoTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.streamingAspects = None
self.foundInstance = False
self.creationSoftwareComment = ''
self.currentEltTag = "(before xbrli:xbrl)"
self.numRootFacts = 0
def start(self, tag, attrib, nsmap=None):
if self.newTree:
if tag == "{http://www.xbrl.org/2003/instance}xbrl":
self.foundInstance = True
self.newTree = False
else: # break
raise NotInstanceDocumentException()
elif not tag.startswith("{http://www.xbrl.org/"):
self.numRootFacts += 1
if self.numRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
self.currentEltTag = tag
def end(self, tag):
pass
def data(self, data):
pass
def comment(self, text):
if not self.foundInstance: # accumulate comments before xbrli:xbrl
self.creationSoftwareComment += ('\n' if self.creationSoftwareComment else '') + text
elif not self.creationSoftwareComment:
self.creationSoftwareComment = text # or first comment after xbrli:xbrl
def pi(self, target, data):
if target == "xbrl-streamable-instance":
if self.currentEltTag == "{http://www.xbrl.org/2003/instance}xbrl":
self.streamingAspects = dict(etree.PI(target,data).attrib.copy()) # dereference target results
else:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(target)s, must follow xbrli:xbrl element but was found at %(element)s"),
modelObject=modelXbrl, target=target, element=self.currentEltTag)
def close(self):
if not self.creationSoftwareComment:
self.creationSoftwareComment = None
return True
instInfo = instInfoTarget()
infoParser = etree.XMLParser(recover=True, huge_tree=True, target=instInfo)
try:
etree.parse(_file, parser=infoParser, base_url=filepath)
except NotInstanceDocumentException:
pass
'''
foundErrors = False
foundInstance = False
streamingAspects = None
creationSoftwareComment = None
instInfoNumRootFacts = 0
numElts = 0
elt = None
instInfoContext = etree.iterparse(_file, events=("start","end"), huge_tree=True)
for event, elt in instInfoContext:
if event == "start":
if elt.getparent() is not None:
if elt.getparent().tag == "{http://www.xbrl.org/2003/instance}xbrl":
if not foundInstance:
foundInstance = True
pi = precedingProcessingInstruction(elt, "xbrl-streamable-instance")
if pi is None:
break
else:
streamingAspects = dict(pi.attrib.copy())
if creationSoftwareComment is None:
creationSoftwareComment = precedingComment(elt)
if not elt.tag.startswith("{http://www.xbrl.org/"):
instInfoNumRootFacts += 1
if instInfoNumRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming tree check", minTimeToShow=20.0)
elif not foundInstance:
break
elif elt.tag == "{http://www.xbrl.org/2003/instance}xbrl":
creationSoftwareComment = precedingComment(elt)
if precedingProcessingInstruction(elt, "xbrl-streamable-instance") is not None:
modelXbrl.error("streamingExtensions:headerMisplaced",
_("Header is misplaced: %(error)s, must follow xbrli:xbrl element"),
modelObject=elt)
elif event == "end":
elt.clear()
numElts += 1
if numElts % 1000 == 0 and elt.getparent() is not None:
while elt.getprevious() is not None and elt.getparent() is not None:
del elt.getparent()[0]
if elt is not None:
elt.clear()
_file.seek(0,io.SEEK_SET) # allow reparsing
if not foundInstance or streamingAspects is None:
del elt
_file.close()
return None
modelXbrl.profileStat(_("streaming tree check"), time.time() - startedAt)
startedAt = time.time()
try:
version = Decimal(streamingAspects.get("version"))
if int(version) != 1:
modelXbrl.error("streamingExtensions:unsupportedVersion",
_("Streaming version %(version)s, major version number must be 1"),
modelObject=elt, version=version)
foundErrors = True
except (InvalidOperation, OverflowError):
modelXbrl.error("streamingExtensions:versionError",
_("Version %(version)s, number must be 1.n"),
modelObject=elt, version=streamingAspects.get("version", "(none)"))
foundErrors = True
for bufAspect in ("contextBuffer", "unitBuffer", "footnoteBuffer"):
try:
bufLimit = Decimal(streamingAspects.get(bufAspect, "INF"))
if bufLimit < 1 or (bufLimit.is_finite() and bufLimit % 1 != 0):
raise InvalidOperation
elif bufAspect == "contextBuffer":
contextBufferLimit = bufLimit
elif bufAspect == "unitBuffer":
unitBufferLimit = bufLimit
elif bufAspect == "footnoteBuffer":
footnoteBufferLimit = bufLimit
except InvalidOperation:
modelXbrl.error("streamingExtensions:valueError",
_("Streaming %(attrib)s %(value)s, number must be a positive integer or INF"),
modelObject=elt, attrib=bufAspect, value=streamingAspects.get(bufAspect))
foundErrors = True
if _streamingExtensionsValidate:
incompatibleValidations = []
_validateDisclosureSystem = modelXbrl.modelManager.validateDisclosureSystem
_disclosureSystem = modelXbrl.modelManager.disclosureSystem
if _validateDisclosureSystem and _disclosureSystem.EFM:
incompatibleValidations.append("EFM")
if _validateDisclosureSystem and _disclosureSystem.GFM:
incompatibleValidations.append("GFM")
if _validateDisclosureSystem and _disclosureSystem.EBA:
incompatibleValidations.append("EBA")
if _validateDisclosureSystem and _disclosureSystem.HMRC:
incompatibleValidations.append("EBA")
if modelXbrl.modelManager.validateCalcLB:
incompatibleValidations.append("calculation LB")
if incompatibleValidations:
modelXbrl.error("streamingExtensions:incompatibleValidation",
_("Streaming instance validation does not support %(incompatibleValidations)s validation"),
modelObject=modelXbrl, incompatibleValidations=', '.join(incompatibleValidations))
foundErrors = True
if instInfoContext.error_log:
foundErrors = True
logSyntaxErrors(instInfoContext)
del instInfoContext # dereference
if foundErrors:
_file.close()
return None
_encoding = XmlUtil.encoding(_file.read(512))
_file.seek(0,io.SEEK_SET) # allow reparsing
if _streamingExtensionsValidate:
validator = Validate(modelXbrl)
instValidator = validator.instValidator
eltMdlObjs = {}
contextBuffer = []
unitBuffer = []
footnoteBuffer = []
factBuffer = []
numFacts = 1
class modelLoaderTarget():
def __init__(self, element_factory=None, parser=None):
self.newTree = True
self.currentMdlObj = None
self.beforeInstanceStream = True
self.numRootFacts = 1
def start(self, tag, attrib, nsmap=None):
mdlObj = _parser.makeelement(tag, attrib=attrib, nsmap=nsmap)
mdlObj.sourceline = 1
if self.newTree:
self.newTree = False
self.currentMdlObj = mdlObj
modelDocument = ModelDocument(modelXbrl, Type.INSTANCE, mappedUri, filepath, mdlObj.getroottree())
modelXbrl.modelDocument = modelDocument # needed for incremental validation
mdlObj.init(modelDocument)
modelDocument.parser = _parser # needed for XmlUtil addChild's makeelement
modelDocument.parserLookupName = _parserLookupName
modelDocument.parserLookupClass = _parserLookupClass
modelDocument.xmlRootElement = mdlObj
modelDocument.schemaLocationElements.add(mdlObj)
modelDocument.documentEncoding = _encoding
modelDocument._creationSoftwareComment = creationSoftwareComment
modelXbrl.info("streamingExtensions:streaming",
_("Stream processing this instance."),
modelObject = modelDocument)
else:
self.currentMdlObj.append(mdlObj)
self.currentMdlObj = mdlObj
mdlObj._init()
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if (self.beforeInstanceStream and (
(ns == XbrlConst.link and ln not in ("schemaRef", "linkbaseRef")) or
(ns == XbrlConst.xbrli and ln in ("context", "unit")) or
(ns not in (XbrlConst.link, XbrlConst.xbrli)))):
self.beforeInstanceStream = False
if _streamingExtensionsValidate:
instValidator.validate(modelXbrl, modelXbrl.modelManager.formulaOptions.typedParameters())
else: # need default dimensions
ValidateXbrlDimensions.loadDimensionDefaults(modelXbrl)
return mdlObj
def end(self, tag):
modelDocument = modelXbrl.modelDocument
mdlObj = self.currentMdlObj
parentMdlObj = mdlObj.getparent()
self.currentMdlObj = parentMdlObj
ns = mdlObj.namespaceURI
ln = mdlObj.localName
if ns == XbrlConst.xbrli:
if ln == "context":
if mdlObj.get("sticky"):
del mdlObj.attrib["sticky"]
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
else:
if _streamingExtensionsValidate and len(contextBuffer) >= contextBufferLimit:
# drop before adding as dropped may have same id as added
cntx = contextBuffer.pop(0)
dropContext(modelXbrl, cntx)
del parentMdlObj[parentMdlObj.index(cntx)]
cntx = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.contextDiscover(mdlObj)
if contextBufferLimit.is_finite():
contextBuffer.append(mdlObj)
if _streamingExtensionsValidate:
contextsToCheck = (mdlObj,)
instValidator.checkContexts(contextsToCheck)
if modelXbrl.hasXDT:
instValidator.checkContextsDimensions(contextsToCheck)
del contextsToCheck # dereference
elif ln == "unit":
if _streamingExtensionsValidate and len(unitBuffer) >= unitBufferLimit:
# drop before additing as dropped may have same id as added
unit = unitBuffer.pop(0)
dropUnit(modelXbrl, unit)
del parentMdlObj[parentMdlObj.index(unit)]
unit = None
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.unitDiscover(mdlObj)
if unitBufferLimit.is_finite():
unitBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkUnits( (mdlObj,) )
elif ln == "xbrl": # end of document
# check remaining footnote refs
for footnoteLink in footnoteBuffer:
checkFootnoteHrefs(modelXbrl, footnoteLink)
elif ns == XbrlConst.link:
if ln == "footnoteLink":
XmlValidate.validate(modelXbrl, mdlObj)
footnoteLinks = (mdlObj,)
modelDocument.linkbaseDiscover(footnoteLinks, inInstance=True)
if footnoteBufferLimit.is_finite():
footnoteBuffer.append(mdlObj)
if _streamingExtensionsValidate:
instValidator.checkLinks(footnoteLinks)
if len(footnoteBuffer) > footnoteBufferLimit:
# check that hrefObjects for locators were all satisfied
# drop before addition as dropped may have same id as added
footnoteLink = footnoteBuffer.pop(0)
checkFootnoteHrefs(modelXbrl, footnoteLink)
dropFootnoteLink(modelXbrl, footnoteLink)
del parentMdlObj[parentMdlObj.index(footnoteLink)]
footnoteLink = None
footnoteLinks = None
elif ln in ("schemaRef", "linkbaseRef"):
modelDocument.discoverHref(mdlObj)
elif not modelXbrl.skipDTS:
if ln in ("roleRef", "arcroleRef"):
modelDocument.linkbaseDiscover((mdlObj,), inInstance=True)
elif parentMdlObj.qname == XbrlConst.qnXbrliXbrl:
self.numRootFacts += 1
XmlValidate.validate(modelXbrl, mdlObj)
modelDocument.factDiscover(mdlObj, modelXbrl.facts)
if _streamingExtensionsValidate:
factsToCheck = (mdlObj,)
instValidator.checkFacts(factsToCheck)
if modelXbrl.hasXDT:
instValidator.checkFactsDimensions(factsToCheck)
del factsToCheck
dropFact(modelXbrl, mdlObj, modelXbrl.facts)
del parentMdlObj[parentMdlObj.index(mdlObj)]
if self.numRootFacts % 1000 == 0:
modelXbrl.profileActivity("... streaming fact {0} of {1} {2:.2f}%".format(self.numRootFacts, instInfoNumRootFacts,
100.0 * self.numRootFacts / instInfoNumRootFacts),
minTimeToShow=20.0)
return mdlObj
def data(self, data):
self.currentMdlObj.text = data
def comment(self, text):
pass
def pi(self, target, data):
pass
def close(self):
return None
_parser, _parserLookupName, _parserLookupClass = parser(modelXbrl, filepath, target=modelLoaderTarget())
etree.parse(_file, parser=_parser, base_url=filepath)
logSyntaxErrors(_parser)
_file.close()
if _streamingExtensionsValidate and validator is not None:
del instValidator
validator.close()
# track that modelXbrl has been validated by this streaming extension
modelXbrl._streamingExtensionValidated = True
modelXbrl.profileStat(_("streaming complete"), time.time() - startedAt)
return modelXbrl.modelDocument