def do_convert(self, filename=None):
if filename is None and exists(self.output + '.nofilename'):
output = self.output + '.nofilename'
else:
output = self.output
input = open(self.input)
orig = input.read()
input.close()
data = datastream(self.transform.name())
res_data = self.transform.convert(orig, data, filename=filename)
self.assert_(idatastream.isImplementedBy(res_data))
got = res_data.getData()
try:
output = open(output)
except IOError:
import sys
print >>sys.stderr, 'No output file found.'
print >>sys.stderr, 'File %s created, check it !' % self.output
output = open(output, 'w')
output.write(got)
output.close()
self.assert_(0)
expected = output.read()
print self.normalize
if self.normalize is not None:
expected = self.normalize(expected)
got = self.normalize(got)
output.close()
self.assertEquals(got, expected,
'[%s]\n\n!=\n\n[%s]\n\nIN %s(%s)' % (
got, expected, self.transform.name(), self.input))
self.assertEquals(self.subobjects, len(res_data.getSubObjects()),
'%s\n\n!=\n\n%s\n\nIN %s(%s)' % (
self.subobjects, len(res_data.getSubObjects()), self.transform.name(), self.input))
def transform(self, instance, mt, **kwargs):
"""Takes a mimetype so object.foo.transform('text/plain') should return
a plain text version of the raw content
return None if no data or if data is untranformable to desired output
mime type
"""
encoding = self.original_encoding
orig = self.getRaw(encoding, instance)
if not orig:
return None
#on ZODB Transaction commit there is by specification
#no acquisition context. If it is not present, take
#the untransformed getRaw, this is necessary for
#being used with APE
# Also don't break if transform was applied with a stale instance
# from the catalog while rebuilding the catalog
if not hasattr(instance, 'aq_parent'):
return orig
transformer = getToolByName(instance, 'portal_transforms')
data = transformer.convertTo(mt, orig, object=self, usedby=self.id,
context=instance,
mimetype=self.mimetype,
filename=self.filename)
if data:
assert idatastream.isImplementedBy(data)
_data = data.getData()
instance.addSubObjects(data.getSubObjects())
portal_encoding = kwargs.get('encoding',None) or \
self.portalEncoding(instance)
encoding = data.getMetadata().get("encoding") or encoding \
or portal_encoding
if portal_encoding != encoding:
_data = unicode(_data, encoding).encode(portal_encoding)
return _data
# we have not been able to transform data
# return the raw data if it's not binary data
# FIXME: is this really the behaviour we want ?
if not self.isBinary():
portal_encoding = kwargs.get('encoding',None) or \
self.portalEncoding(instance)
if portal_encoding != encoding:
orig = self.getRaw(portal_encoding)
return orig
return None
def convertTo(self, target_mimetype, orig, data=None, object=None,
usedby=None, context=None, **kwargs):
"""Convert orig to a given mimetype
* orig is an encoded string
* data an optional idatastream object. If None a new datastream will be
created and returned
* optional object argument is the object on which is bound the data.
If present that object will be used by the engine to bound cached data.
* additional arguments (kwargs) will be passed to the transformations.
Some usual arguments are : filename, mimetype, encoding
return an object implementing idatastream or None if no path has been
found.
"""
target_mimetype = str(target_mimetype)
if object is not None:
cache = Cache(object)
data = cache.getCache(target_mimetype)
if data is not None:
time, data = data
if self.max_sec_in_cache == 0 or time < self.max_sec_in_cache:
return data
if data is None:
data = self._wrap(target_mimetype)
registry = getToolByName(self, 'mimetypes_registry')
if not getattr(aq_base(registry), 'classify', None):
# avoid problems when importing a site with an old mimetype registry
# XXX return None or orig?
return None
orig_mt = registry.classify(orig,
mimetype=kwargs.get('mimetype'),
filename=kwargs.get('filename'))
orig_mt = str(orig_mt)
if not orig_mt:
log('Unable to guess input mime type (filename=%s, mimetype=%s)' %(
kwargs.get('mimetype'), kwargs.get('filename')), severity=DEBUG)
return None
target_mt = registry.lookup(target_mimetype)
if target_mt:
target_mt = target_mt[0]
else:
log('Unable to match target mime type %s'% str(target_mimetype),
severity=DEBUG)
return None
## fastpath
# If orig_mt and target_mt are the same, we only allow
# a one-hop transform, a.k.a. filter.
# XXX disabled filtering for now
filter_only = False
if orig_mt == str(target_mt):
filter_only = True
data.setData(orig)
md = data.getMetadata()
md['mimetype'] = str(orig_mt)
if object is not None:
cache.setCache(str(target_mimetype), data)
return data
## get a path to output mime type
requirements = self._policies.get(str(target_mt), [])
path = self._findPath(orig_mt, target_mt, list(requirements))
if not path and requirements:
log('Unable to satisfy requirements %s' % ', '.join(requirements),
severity=DEBUG)
path = self._findPath(orig_mt, target_mt)
if not path:
log('NO PATH FROM %s TO %s : %s' % (orig_mt, target_mimetype, path),
severity=DEBUG)
return None #XXX raise TransformError
if len(path) > 1:
## create a chain on the fly (sly)
transform = chain()
for t in path:
transform.registerTransform(t)
else:
transform = path[0]
result = transform.convert(orig, data, context=context, usedby=usedby, **kwargs)
assert(idatastream.isImplementedBy(result),
'result doesn\'t implemented idatastream')
self._setMetaData(result, transform)
# set cache if possible
if object is not None and result.isCacheable():
cache.setCache(str(target_mimetype), result)
# return idatastream object
return result
def _unwrap(self, data):
"""unwrap data from an icache"""
if idatastream.isImplementedBy(data):
data = data.getData()
return data