def validate(filename, schema = None):
try:
folia.validate(filename, schema)
except Exception as e:
print >>sys.stderr, "VALIDATION ERROR in " + filename
print >>sys.stderr, str(e)
return False
print >>sys.stderr, "Validated succesfully: " + filename
return True
def validate(filepath):
global schema
print "\tValidating"
try:
folia.validate(filepath, schema)
print "\t\tSuccess"
except Exception, e:
errout("\t\tERROR: DOCUMENT DOES NOT VALIDATE! (" + filepath + "): " +
str(e))
def validate(filename, schema=None):
try:
folia.validate(filename, schema)
except Exception as e:
print >> sys.stderr, "VALIDATION ERROR in " + filename
print >> sys.stderr, str(e)
return False
print >> sys.stderr, "Validated succesfully: " + filename
return True
def validate(filename, schema = None, quick=False):
try:
folia.validate(filename, schema)
except Exception as e:
print("VALIDATION ERROR against RelaxNG schema (stage 1/2), in " + filename,file=sys.stderr)
print(str(e), file=sys.stderr)
return False
try:
folia.Document(file=filename)
except Exception as e:
print("VALIDATION ERROR on full parse by library (stage 2/2), in " + filename,file=sys.stderr)
print(e.__class__.__name__ + ": " + str(e),file=sys.stderr)
print("Full traceback follows:",file=sys.stderr)
ex_type, ex, tb = sys.exc_info()
traceback.print_tb(tb)
return False
print("Validated successfully: " + filename,file=sys.stderr)
return True
print "Generating Relax NG schema"
folia.relaxng('folia-min.rng')
#!/usr/bin/python
data = codecs.open("folia-min.rng",'r','utf-8').readlines()
data.insert(1,u"""<!--
RelaxNG schema for FoLiA XML v%s
by Maarten van Gompel
Induction of Linguistic Knowledge Research group
Tilburg University
http://ilk.uvt.nl/folia
http://github.com/proycon/folia
Schema version %s
(auto-generated by pynlpl.formats.folia)
Validation examples:
$ xmllint ‒‒relaxng folia.rng foliadocument.xml
$ jing folia.rng foliadocument.xml
Licensed under the GNU General Public License v3
-->
""" % (folia.FOLIAVERSION, folia.LIBVERSION))
codecs.open('folia.rng','w','utf-8').writelines(data)
print "Validating example document"
folia.validate('../test/example.xml')
data = io.open("folia-min.rng", 'r', encoding='utf-8').readlines()
data.insert(
1, """<!--
RelaxNG schema for FoLiA XML v%s
by Maarten van Gompel
Induction of Linguistic Knowledge Research group
Tilburg University
http://ilk.uvt.nl/folia
http://github.com/proycon/folia
Schema version %s
(auto-generated by pynlpl.formats.folia)
Validation examples:
$ xmllint ‒‒relaxng folia.rng foliadocument.xml
$ jing folia.rng foliadocument.xml
However, it is recommended to use the foliavalidator tool instead,
as it does extra validation that can not be captured by RelaxNG!
$ foliavalidator foliadocument.xml
Licensed under the GNU General Public License v3
-->
""" % (folia.FOLIAVERSION, folia.LIBVERSION))
io.open('folia.rng', 'w', encoding='utf-8').writelines(data)
print("Validating example document")
folia.validate('../test/example.xml')
def validate(filename,
schema=None,
quick=False,
deep=False,
stricttextvalidation=False,
warn=True,
debug=False):
try:
folia.validate(filename, schema)
except Exception as e:
print("VALIDATION ERROR against RelaxNG schema (stage 1/2), in " +
filename,
file=sys.stderr)
print(str(e), file=sys.stderr)
return False
try:
document = folia.Document(file=filename,
deepvalidation=deep,
textvalidation=True,
verbose=True,
debug=debug)
except folia.DeepValidationError as e:
print(
"DEEP VALIDATION ERROR on full parse by library (stage 2/2), in " +
filename,
file=sys.stderr)
print(e.__class__.__name__ + ": " + str(e), file=sys.stderr)
return False
except Exception as e:
print("VALIDATION ERROR on full parse by library (stage 2/2), in " +
filename,
file=sys.stderr)
print(e.__class__.__name__ + ": " + str(e), file=sys.stderr)
print("-- Full traceback follows -->", file=sys.stderr)
ex_type, ex, tb = sys.exc_info()
traceback.print_exception(ex_type, ex, tb)
return False
if not document.version:
print("VALIDATION ERROR: Document does not advertise FoLiA version (" +
filename + ")",
file=sys.stderr)
return False
elif folia.checkversion(document.version) == -1 and warn:
print(
"WARNING: Document (" + filename +
") uses an older FoLiA version (" + document.version +
") but is validated according to the newer specification (" +
folia.FOLIAVERSION +
"). You might want to increase the version attribute if this is a document you created and intend to publish.",
file=sys.stderr)
if document.textvalidationerrors:
if stricttextvalidation:
print("VALIDATION ERROR because of text validation errors, in " +
filename,
file=sys.stderr)
return False
elif warn:
print(
"WARNING: there were " + str(document.textvalidationerrors) +
" text validation errors but these are currently not counted toward the full validation result (use -t for strict text validation)",
file=sys.stderr)
print("Validated successfully: " + filename, file=sys.stderr)
return True
