1717import threading
1818from collections import OrderedDict
1919from collections .abc import MutableMapping
20- from typing import MutableSequence , Optional , Sequence
20+ from typing import MutableSequence , Optional , Sequence , Union
2121
2222from opentelemetry .util import types
2323
24- _VALID_ATTR_VALUE_TYPES = (bool , str , int , float )
24+ # bytes are accepted as a user supplied value for attributes but
25+ # decoded to strings internally.
26+ _VALID_ATTR_VALUE_TYPES = (bool , str , bytes , int , float )
2527
2628
2729_logger = logging .getLogger (__name__ )
2830
2931
30- def _is_valid_attribute_value (value : types .AttributeValue ) -> bool :
31- """Checks if attribute value is valid.
32+ def _clean_attribute (
33+ key : str , value : types .AttributeValue , max_len : Optional [int ]
34+ ) -> Optional [types .AttributeValue ]:
35+ """Checks if attribute value is valid and cleans it if required.
36+
37+ The function returns the cleaned value or None if the value is not valid.
3238
3339 An attribute value is valid if it is either:
3440 - A primitive type: string, boolean, double precision floating
3541 point (IEEE 754-1985) or integer.
3642 - An array of primitive type values. The array MUST be homogeneous,
3743 i.e. it MUST NOT contain values of different types.
38- """
3944
40- if isinstance (value , _VALID_ATTR_VALUE_TYPES ):
41- return True
45+ An attribute needs cleansing if:
46+ - Its length is greater than the maximum allowed length.
47+ - It needs to be encoded/decoded e.g, bytes to strings.
48+ """
4249
43- if isinstance (value , Sequence ):
50+ if key is None or key is "" :
51+ _logger .warning ("invalid key `%s` (empty or null)" , key )
52+ return None
4453
54+ if isinstance (value , _VALID_ATTR_VALUE_TYPES ):
55+ return _clean_attribute_value (value , max_len )
56+ elif isinstance (value , Sequence ):
4557 sequence_first_valid_type = None
58+ cleaned_seq = []
59+
4660 for element in value :
61+ # None is considered valid in any sequence
62+ if element is None :
63+ cleaned_seq .append (element )
64+
65+ element = _clean_attribute_value (element , max_len )
4766 if element is None :
4867 continue
68+
4969 element_type = type (element )
70+ # Reject attribute value if sequence contains a value with an incompatible type.
5071 if element_type not in _VALID_ATTR_VALUE_TYPES :
5172 _logger .warning (
5273 "Invalid type %s in attribute value sequence. Expected one of "
@@ -57,56 +78,51 @@ def _is_valid_attribute_value(value: types.AttributeValue) -> bool:
5778 for valid_type in _VALID_ATTR_VALUE_TYPES
5879 ],
5980 )
60- return False
81+ return None
82+
6183 # The type of the sequence must be homogeneous. The first non-None
6284 # element determines the type of the sequence
6385 if sequence_first_valid_type is None :
6486 sequence_first_valid_type = element_type
65- elif not isinstance (element , sequence_first_valid_type ):
87+ # use equality instead of isinstance as isinstance(True, int) evaluates to True
88+ elif element_type != sequence_first_valid_type :
6689 _logger .warning (
6790 "Mixed types %s and %s in attribute value sequence" ,
6891 sequence_first_valid_type .__name__ ,
6992 type (element ).__name__ ,
7093 )
71- return False
72- return True
94+ return None
95+
96+ if element is not None :
97+ cleaned_seq .append (element )
98+
99+ # Freeze mutable sequences defensively
100+ return tuple (cleaned_seq )
73101
74102 _logger .warning (
75103 "Invalid type %s for attribute value. Expected one of %s or a "
76104 "sequence of those types" ,
77105 type (value ).__name__ ,
78106 [valid_type .__name__ for valid_type in _VALID_ATTR_VALUE_TYPES ],
79107 )
80- return False
81-
82108
83- def _filter_attributes (attributes : types .Attributes ) -> None :
84- """Applies attribute validation rules and drops (key, value) pairs
85- that doesn't adhere to attributes specification.
86109
87- https://github.com/open-telemetry/opentelemetry-specification/blob/main/specification/common/common.md#attributes.
88- """
89- if attributes :
90- for attr_key , attr_value in list (attributes .items ()):
91- if not attr_key :
92- _logger .warning ("invalid key `%s` (empty or null)" , attr_key )
93- attributes .pop (attr_key )
94- continue
110+ def _clean_attribute_value (
111+ value : types .AttributeValue , limit : Optional [int ]
112+ ) -> Union [types .AttributeValue , None ]:
113+ if value is None :
114+ return
95115
96- if _is_valid_attribute_value (attr_value ):
97- if isinstance (attr_value , MutableSequence ):
98- attributes [attr_key ] = tuple (attr_value )
99- if isinstance (attr_value , bytes ):
100- try :
101- attributes [attr_key ] = attr_value .decode ()
102- except ValueError :
103- attributes .pop (attr_key )
104- _logger .warning ("Byte attribute could not be decoded." )
105- else :
106- attributes .pop (attr_key )
116+ if isinstance (value , bytes ):
117+ try :
118+ value = value .decode ()
119+ except ValueError :
120+ _logger .warning ("Byte attribute could not be decoded." )
121+ return None
107122
108-
109- _DEFAULT_LIMIT = 128
123+ if limit is not None and isinstance (value , str ):
124+ value = value [:limit ]
125+ return value
110126
111127
112128class BoundedAttributes (MutableMapping ):
@@ -118,9 +134,10 @@ class BoundedAttributes(MutableMapping):
118134
119135 def __init__ (
120136 self ,
121- maxlen : Optional [int ] = _DEFAULT_LIMIT ,
137+ maxlen : Optional [int ] = None ,
122138 attributes : types .Attributes = None ,
123139 immutable : bool = True ,
140+ max_value_len : Optional [int ] = None ,
124141 ):
125142 if maxlen is not None :
126143 if not isinstance (maxlen , int ) or maxlen < 0 :
@@ -129,10 +146,10 @@ def __init__(
129146 )
130147 self .maxlen = maxlen
131148 self .dropped = 0
149+ self .max_value_len = max_value_len
132150 self ._dict = OrderedDict () # type: OrderedDict
133151 self ._lock = threading .Lock () # type: threading.Lock
134152 if attributes :
135- _filter_attributes (attributes )
136153 for key , value in attributes .items ():
137154 self [key ] = value
138155 self ._immutable = immutable
@@ -158,7 +175,10 @@ def __setitem__(self, key, value):
158175 elif self .maxlen is not None and len (self ._dict ) == self .maxlen :
159176 del self ._dict [next (iter (self ._dict .keys ()))]
160177 self .dropped += 1
161- self ._dict [key ] = value
178+
179+ value = _clean_attribute (key , value , self .max_value_len )
180+ if value is not None :
181+ self ._dict [key ] = value
162182
163183 def __delitem__ (self , key ):
164184 if getattr (self , "_immutable" , False ):
0 commit comments