Fix log decode Python example

examples/decode_sensor_binary_log.py

@@ -80,8 +80,10 @@ class PingViewerLogReader:
     # timestamp format for recovery hh:mm:ss.xxx
     # includes optional \x00 (null byte) before every character because Windows
     TIMESTAMP_FORMAT = re.compile(
-        b'(\x00?\d){2}(\x00?:\x00?[0-5]\x00?\d){2}\x00?\.(\x00?\d){3}')
+        b'(\x00?\\d){2}(\x00?:\x00?[0-5]\x00?\\d){2}\x00?\\.(\x00?\\d){3}')
     MAX_TIMESTAMP_LENGTH = 12 * 2
+    # byte encoding assumes posix until proven otherwise
+    ENCODING = 'UTF-8'
 
     def __init__(self, filename: str):
         self.filename = filename
@@ -107,20 +109,17 @@ def unpack_array(cls, file: IO[Any]):
         if array_size <= cls.MAX_ARRAY_LENGTH:
             return file.read(array_size)
 
-    @classmethod
-    def unpack_string(cls, file: IO[Any]):
-        return cls.unpack_array(file).decode('UTF-8')
+    def unpack_string(self, file: IO[Any]):
+        return self.unpack_array(file).decode(self.ENCODING)
 
-    @classmethod
-    def unpack_message(cls, file: IO[Any]):
-        timestamp = cls.unpack_string(file)
-        message = cls.unpack_array(file)
+    def unpack_message(self, file: IO[Any]):
+        timestamp = self.unpack_string(file)
+        message = self.unpack_array(file)
         if message is None:
-            return cls.recover(file)
+            return self.recover(file)
         return (timestamp, message)
 
-    @classmethod
-    def recover(cls, file: IO[Any]):
+    def recover(self, file: IO[Any]):
         """ Attempt to recover from a failed read.
 
         Assumed that a bad number has been read from the last cls.UINT.size
@@ -129,32 +128,32 @@ def recover(cls, file: IO[Any]):
 
         """
         # TODO: log when recovery attempts occur, and bytes lost when they succeed
-        file.seek(current_pos := (file.tell() - cls.UINT.size))
+        file.seek(current_pos := (file.tell() - self.UINT.size))
         prev_ = next_ = b''
         start = amount_read = 0
-        while not (match := cls.TIMESTAMP_FORMAT.search(
+        while not (match := self.TIMESTAMP_FORMAT.search(
                 roi := (prev_ + next_), start)):
             prev_ = next_
-            next_ = file.read(cls.MAX_ARRAY_LENGTH)
+            next_ = file.read(self.MAX_ARRAY_LENGTH)
             if not next_:
                 break # run out of file
-            amount_read += cls.MAX_ARRAY_LENGTH
+            amount_read += self.MAX_ARRAY_LENGTH
             if start == 0 and prev_:
                 # onto the second read
                 # -> match on potential overlap + new region, not the
                 #    already-checked (impossible) region
-                start = cls.MAX_ARRAY_LENGTH - cls.MAX_TIMESTAMP_LENGTH
+                start = self.MAX_ARRAY_LENGTH - self.MAX_TIMESTAMP_LENGTH
         else:
             # match was found
             end = match.end()
-            timestamp = roi[match.start():end].decode('UTF-8')
+            timestamp = roi[match.start():end].decode(self.ENCODING)
             amount_read -= (len(roi) - end)
             self.failed_bytes += amount_read
             # return the file pointer to the end of this timestamp
             file.seek(current_pos + amount_read)
             # attempt to extract the corresponding message, or recover anew
-            if (message := cls.unpack_array(file)) is None:
-                return cls.recover(file)
+            if (message := self.unpack_array(file)) is None:
+                return self.recover(file)
             return (timestamp, message)
         # Calculate bytes from start of recovery attempt to end of the file
         file_size = file.tell()
@@ -165,6 +164,10 @@ def recover(cls, file: IO[Any]):
 
     def unpack_header(self, file: IO[Any]):
         self.header.string = self.unpack_string(file)
+        if '\x00' in self.header.string:
+            # Windows uses big-endian wide characters
+            self.ENCODING = 'UTF-16-be'
+            self.header.string = self.header.string.encode('UTF-8').decode(self.ENCODING)
         self.header.version = self.unpack_int(file)
 
         for info in ('hash_commit', 'date', 'tag', 'os_name', 'os_version'):