Refactor the logic in import_issues for trac
Currently when calling import issues the program does several things:
- Clone the git repo or rather lists all cloned repo and asks which one
should be used
- Queries all the tickets in trac
- Convert them to JSON blob
- Import them into the git repo selected in the first step
- Push said git repo and delete it from the filesystem
With this commits, the logic is moved into the CLI itself, which now
does select where is the git repo to use, then does the querying of trac
and converting to JSON, does the import into git in one step at the end
speeding up the process and finally push and clean as before.
This means that import_issues() can now be used outside of any git repo
and it will just output the JSON blob in the specified directory.