Added md5sum to Message (fixes #165) (#169)

* added stub _compute_md5

* added md5sum to Message

* don't treat md5 as a property

* outlined preprocessing

* added custom __str__ to Constant

* added __str__ to Field

* added without_package_name to Field

* implemented _compute_md5

* added test_md5sum

* bug fix: md5 -> md5sum

* bug fix: incorrect md5sum calculator, also moved to MsgFormat

* added extra cases to test_md5sum

* bug fix: corrected md5text calculation

* added additional test assertion

src/roswire/definitions/msg.py

@@ -5,6 +5,7 @@
 from io import BytesIO
 import logging
 import functools
+import hashlib
 import struct
 import re
 import os
@@ -31,7 +32,7 @@
 ConstantValue = Union[str, int, float]
 
 
-@attr.s(frozen=True)
+@attr.s(frozen=True, str=False)
 class Constant:
     typ = attr.ib(type=str)
     name = attr.ib(type=str)
@@ -46,8 +47,11 @@ def to_dict(self) -> Dict[str, Any]:
                 'name': self.name,
                 'value': self.value}
 
+    def __str__(self) -> str:
+        return f"{self.typ} {self.name}={str(self.value)}"
 
-@attr.s(frozen=True)
+
+@attr.s(frozen=True, str=False)
 class Field:
     typ: str = attr.ib()
     name: str = attr.ib()
@@ -85,6 +89,13 @@ def to_dict(self) -> Dict[str, str]:
         return {'type': self.typ,
                 'name': self.name}
 
+    def without_package_name(self) -> 'Field':
+        typ = self.typ.partition('/')[2] if '/' in self.typ else self.typ
+        return Field(typ, self.name)
+
+    def __str__(self) -> str:
+        return f"{self.typ} {self.name}"
+
 
 @attr.s(frozen=True)
 class MsgFormat:
@@ -209,6 +220,27 @@ def flatten(self,
                 fmt = name_to_format[field.typ]
                 yield from fmt.flatten(name_to_format, ctx + (field.name,))
 
+    def md5text(self, name_to_msg: Mapping[str, 'MsgFormat']) -> str:
+        """Computes the MD5 text for this format."""
+        lines: List[str] = []
+        lines += [str(c) for c in self.constants]
+        for f in self.fields:
+            if is_builtin(f.base_type):
+                lines += [str(f.without_package_name())]
+            else:
+                f_md5 = name_to_msg[f.base_type].md5sum(name_to_msg)
+                lines += [f'{f_md5} {f.name}']
+        return '\n'.join(lines)
+
+    def md5sum(self, name_to_msg: Mapping[str, 'MsgFormat']) -> str:
+        """Computes the MD5 sum for this format."""
+        logger.debug("generating md5sum: %s", self.fullname)
+        txt = self.md5text(name_to_msg)
+        logger.debug("generated md5 text [%s]:\n%s", self.fullname, txt)
+        md5sum = hashlib.md5(txt.encode('utf-8')).hexdigest()
+        logger.debug("generated md5sum [%s]: %s", self.fullname, md5sum)
+        return md5sum
+
 
 class Message:
     """Base class used by all messages."""
@@ -241,6 +273,11 @@ def to_dict(self) -> Dict[str, Any]:
             d[name] = self._to_dict_value(val)
         return d
 
+    @classmethod
+    def md5sum(cls) -> str:
+        """Returns the md5sum for this message type."""
+        raise NotImplementedError
+
     @classmethod
     def read(cls, b: BinaryIO) -> 'Message':
         raise NotImplementedError

src/roswire/definitions/type_db.py

@@ -1,7 +1,7 @@
 __all__ = ('TypeDatabase',)
 
 from typing import (Collection, Type, Mapping, Iterator, Dict, ClassVar, Any,
-                    Sequence, Callable, BinaryIO)
+                    Sequence, Callable, BinaryIO, List)
 from collections import OrderedDict
 
 import attr
@@ -42,6 +42,8 @@ def build(cls, db_format: FormatDatabase) -> 'TypeDatabase':
             ns['format'] = fmt
             ns['read'] = classmethod(cls._build_read(name_to_type, fmt))
             ns['write'] = cls._build_write(name_to_type, fmt)
+            md5 = fmt.md5sum(db_format.messages)
+            ns['md5sum'] = classmethod(lambda cls, md5=md5: md5)
             t: Type[Message] = type(fmt.name, (Message,), ns)
             t = attr.s(t, frozen=True, slots=True)
             name_to_type[fmt.fullname] = t

test/test_type_database.py

@@ -5,6 +5,7 @@
                                  MsgFormat, Time)
 
 from test_file import build_file_proxy
+from test_bag import load_mavros_type_db
 
 
 def test_build():
@@ -119,3 +120,23 @@ def test_to_and_from_dict():
          'frame_id': 'foo'}
     assert m.to_dict() == d
     assert db_type.from_dict(fmt, d) == m
+
+
+def test_md5sum():
+    db_type = load_mavros_type_db()
+
+    def check(name: str, md5sum_expected: str) -> None:
+        assert db_type[name].md5sum() == md5sum_expected
+
+    check('std_msgs/Header', '2176decaecbce78abc3b96ef049fabed')
+    check('std_msgs/Duration', '3e286caf4241d664e55f3ad380e2ae46')
+    check('mavros_msgs/CommandCode', 'f7e54ea3892a961cc44c9350fdb0855e')
+    check('geometry_msgs/Quaternion', 'a779879fadf0160734f906b8c19c7004')
+    check('geometry_msgs/Vector3', '4a842b65f413084dc2b10fb484ea7f17')
+    check('geometry_msgs/Wrench', '4f539cf138b23283b520fd271b567936')
+    check('geometry_msgs/Transform', 'ac9eff44abf714214112b05d54a3cf9b')
+    check('geometry_msgs/TransformStamped', 'b5764a33bfeb3588febc2682852579b0')
+    check('tf/tfMessage', '94810edda583a504dfda3829e70d7eec')
+    check('visualization_msgs/InteractiveMarkerPose', 'a6e6833209a196a38d798dadb02c81f8')
+    check('mavros_msgs/State', '9e3d873fae342c8f48a8bd64c53d991e')
+    check('mavros_msgs/Mavlink', '6dd71a38b8541fdc2de89a548c7dbc2f')