robshakir
Support of bits type.
Hi,
I see pyangbind doesn't have support of type bits. I think this could be a good feature to extend the pyangbind plugin.
Hi,
Thanks for the suggestion. I would be happy to review a pull request if this were something that you'd like to contribute.
Kind regards, r.
Hi Rob,
Use case:
leaf latitude {
type bits {
bit N {
description
"Latitude bit.";
}
bit N2 {
description
"Latitude bit.";
}
}
description
"Bit that selects between North and South latitude.";
}
Representation:
latitude = ["N", "N2"] (Correct)
latitude = ["N2"] (Correct)
latitude = ["N3"] (Incorrect)
I was locating the issue and I had done the following:
- In build_elemtype() at line 1203 (plugin/pybind.py), Returned the elemtype for bits as follows:
if et.arg == "bits":
bits_dict = {}
for bit in et.search('bit'):
bits_dict[unicode(bit.arg)] = {}
position = bit.search_one('position')
if position is not None:
bits_dict[unicode(bit.arg)]["position"] = int(position.arg)
elemtype = {"native_type": """RestrictedClassType(base_type=unicode, \
restriction_type="sub_dict_keys", \
restriction_arg=%s,)""" %
(bits_dict),
"restriction_argument": bits_dict,
"restriction_type": "sub_dict_keys",
"parent_type": "string",
"base_type": True}
- In yangtypes: at line 286: (lib/yangtypes.py) 2.1: Filled the positions into empty bits in increasing order and added the restriction_tests as follows:
elif rtype == "sub_dict_keys":
new_rarg = copy.deepcopy(rarg)
for k in rarg:
if k.startswith("@"):
new_rarg.pop(k, None)
#populate bits positions
bits_positions = []
for k in new_rarg:
if "position" in new_rarg[k]:
bits_positions.append(int(new_rarg[k]["position"]))
p = 0
for k in new_rarg:
while p in bits_positions:
p += 1
if "position" not in new_rarg[k]:
new_rarg[k]["position"] = p
p += 1
print new_rarg.keys()
self._restriction_tests.append(sub_list_check(new_rarg.keys()))
self._bits_dict = new_rarg
- Added the sub_list_check function which checks if the input is present in the list, This function is being used at point 2 above.
def sub_list_check(lst):
def sub_lst_check(value):
for i in value:
if i not in lst:
return False
return True
return sub_lst_chec
Now, I am compiling my yangfiles with above template of bits. It is compiling fine with below command:
pyang --plugindir $PYBINDPLUGIN -f pybind -o binding.py ietf-lisp-address-types.yang
But in assigning a value to bit as follow is giving the error: obj.latitude=("N", "N2")
Traceback (most recent call last):
File "gen.py", line 5, in
File "/home/ashish/odl_test/bits/binding.py", line 333, in _set_latitude
'generated-type': """YANGDynClass(base=RestrictedClassType(base_type=unicode, restriction_type="sub_dict_keys", restriction_arg={u'N2': {}, u'N': {}},), is_leaf=True, yang_name="latitude", parent=self, path_helper=self._path_helper, extmethods=self._extmethods, register_paths=True, namespace='urn:ietf:params:xml:ns:yang:ietf-lisp-address-types', defining_module='ietf-lisp-address-types', yang_type='bits', is_config=True)""",
ValueError: {'error-string': 'latitude must be of a type compatible with bits', 'generated-type': 'YANGDynClass(base=RestrictedClassType(base_type=unicode, \t\t restriction_type="sub_dict_keys", \t\t\t\t restriction_arg={u\'N2\': {}, u\'N\': {}},), is_leaf=True, yang_name="latitude", parent=self, path_helper=self._path_helper, extmethods=self._extmethods, register_paths=True, namespace=\'urn:ietf:params:xml:ns:yang:ietf-lisp-address-types\', defining_module=\'ietf-lisp-address-types\', yang_type=\'bits\', is_config=True)', 'defined-type': 'ietf-lisp-address-types:bits'}
Can you help me with this, I think I am missing something.
Hi Rob,
Can you please tell me if I am right about my understanding (as explained above).
Thanks
Hi Ashish,
Apologies, I'd missed the update to this issue.
Is there a fork that I can pull to take a look at this by any chance?
The error that you supplied is thrown in the generated bindings when a new instance of the object is created. PyangBind generally treats member variables in the structure as immutable, so that we create a new version of the object rather than trying to mutate the existing one - the creation of the new object throws ValueError in the case that some error is returned during instantiation of the object with type(val, args).
So, to debug the class above, it seems like a good step would be to remove that ValueError being caught, and seeing what the underlying error that is returned is.
Happy to help if you can point me to the code.
Cheers, r.
Hello,
I am working with a few YANG models that use the YANG built-in type bits. I have found PyangBind to be a valuable tool. By chance is there an update to the bits type support enhancement?
Gayle
The ietf-netconf-acm.yang module also uses the bits built in type so a module importing this (I am trying to use ietf-system.yang) encounters an error "could not find a match for nacm:access-operations-type type -> [u'bits']" and "TypeError: could not resolve typedefs [u'nacm:access-operations-type']". Does anyone know a way to skip over this error or does the issue need to be resolved?
You can get around this if you are not planning to use the types in the yang files which use the bits built in types by commenting them out. Not ideal but it works as a short term fix.
I am using pyangbind with IETF models, and came up with this error. I can not ignore the yang parts that use this type: is there a plan to solve the issue ? Thanks a lot for your help!
You can get around this if you are not planning to use the types in the yang files which use the bits built in types by commenting them out. Not ideal but it works as a short term fix.
This worked for me, thank you.
Hi Ashish,
from RFC 6020, I think lexical representation should be:
<latitude> N N2 </latitude>
obj.latitude = "N N2" # space delimited
I modified your implementation as follow:
def sub_list_check(lst):
def sub_lst_check(value):
if isinstance(value, str) is False:
return False
bits = value.split()
for i in bits:
if i not in lst:
return False
return True
return sub_lst_check
Regards,
LL
Support for Bits type was added in https://github.com/robshakir/pyangbind/pull/315
