invalid encoding starting at

I'm getting the following error when loading the proto file attached:

`** (UnicodeConversionError) invalid encoding starting at <<154, 179, 139, 210, 5, 18, 200, 5, 10, 5, 84, 49, 48, 48, 49, 16, 0, 26, 188, 5, 10, 32, 10, 3, 49, 50, 49, 18, 8, 49, 51, 58, 53, 56, 58, 48, 48, 26, 8, 50, 48, 49, 55, 49, 50, 50, 54, 32, 0, 42, ...>>
(elixir) lib/string.ex:1840: String.to_charlist/1
lib/exprotobuf/parser.ex:41: Protobuf.Parser.parse/2
lib/exprotobuf/parser.ex:55: Protobuf.Parser.parse!/2
lib/exprotobuf/parser.ex:10: anonymous fn/3 in Protobuf.Parser.parse_files!/2
(elixir) lib/enum.ex:1826: Enum."-reduce/3-lists^foldl/2-0-"/3
lib/exprotobuf/parser.ex:8: Protobuf.Parser.parse_files!/2
lib/exprotobuf.ex:80: Protobuf.parse/2
expanding macro: Protobuf.__using__/1...`

I was able to load the proto file in a sample java protobuf reader to ensure that the file is in the right format. Not sure if I'm missing a config parameter.

[GTFSDATAFILE_1514312552253.proto.zip](https://github.com/bitwalker/exprotobuf/files/1593746/GTFSDATAFILE_1514312552253.proto.zip)

Start by reproducing the failure with the attached GTFSDATAFILE_1514312552253.proto.zip, then read lib/exprotobuf/parser.ex and the Protobuf.parse entry point in lib/exprotobuf.ex. Trace why parsing reaches String.to_charlist/1 for this file and define done as successfully parsing the valid proto without UnicodeConversionError.

Written by the indexing model from the issue text.