Fix kgenpids string vs bytes usage for python3 for calibre 5.1.

In order to properly get pids etc. we need to pass bytes to MD5 and SHA1
instead of unicode strings. Also ord() is no longer needed since
data is bytes value gets int and we need chr() to get characters from
the mapping bytearrays.

DeDRM_plugin/kgenpids.py

@@ -52,11 +52,11 @@ def SHA1(message):
 def encode(data, map):
     result = ''
     for char in data:
-        value = ord(char)
+        value = char
         Q = (value ^ 0x80) // len(map)
         R = value % len(map)
-        result += map[Q]
-        result += map[R]
+        result += chr(map[Q])
+        result += chr(map[R])
     return result
 
 # Hash the bytes in data and then encode the digest with the characters in map
@@ -238,18 +238,18 @@ def getK4Pids(rec209, token, kindleDatabase):
                 # Get the UserName we added
                 UserName = bytearray.fromhex((kindleDatabase[1])['UserName']).decode()
                 # encode it
-                encodedUsername = encodeHash(UserName,charMap1)
+                encodedUsername = encodeHash(UserName.encode(),charMap1)
                 #print "encodedUsername",encodedUsername.encode('hex')
         except KeyError:
             print("Keys not found in the database {0}.".format(kindleDatabase[0]))
             return pids
 
         # Get the ID string used
-        encodedIDString = encodeHash(IDString,charMap1)
+        encodedIDString = encodeHash(IDString.encode(),charMap1)
         #print "encodedIDString",encodedIDString.encode('hex')
 
         # concat, hash and encode to calculate the DSN
-        DSN = encode(SHA1(MazamaRandomNumber+encodedIDString+encodedUsername),charMap1)
+        DSN = encode(SHA1((MazamaRandomNumber+encodedIDString+encodedUsername).encode()),charMap1)
         #print "DSN",DSN.encode('hex')
         pass