feat: allow set value as BigEndian Int64

**Is your feature request related to a problem? Please describe.**

> Starting with version 0.12.0, the cbt CLI can format certain complex types of data stored in table rows. When you use the cbt read or cbt lookup command, the cbt tool can "pretty print" values stored in the rows.
> https://cloud.google.com/bigtable/docs/cbt-formatting

from https://github.com/googleapis/cloud-bigtable-cbt-cli/pull/28/files, `cbt read` and `cbt lookup` is able to decode BigEndian Int64 with a format file.

BigEndian Int64 and UTF8 string are two popular value types. However, `cbt set` still only able to set value as UTF8 string.

**Describe the solution you'd like**

similar to `cbt read` or `cbt lookup`, we can improve `cbt set` to support format file and encode properly.

**Describe alternatives you've considered**

If it is easy to generalize, it would be nice if `cbt set` can also support full data types, e.g. Hexadecimal, JSON, Protobuf

**Additional context**
`cbt` was quite useful in setup instance/table/column_family in Bigtable emulator (used in development or test environment)
. I have seen a lot use cases where producer and consumer are different applications and mostly implement in different languages.

since `cbt` is not able to set BigEndian Int64, currently a workaround is to use client SDK to mock data in CI of Consumer applications. It would be more convenient if `cbt set` can also