Fix the handling of modulemd files with unicode characters.

This commit introduces new to_text_type helper method and calls it
for return value of all mmd.dumps() calls. That way, we always
end up with proper unicode string represntation on both python
major versions.

This commit also adds unicode character to description of all
the yaml files we use in the tests so we can be sure MBS can
handle unicode characters properly.

This might be temporary fix, depending on the result of discussion
at https://github.com/fedora-modularity/libmodulemd/issues/184.

tests/test_models/test_models.py

@@ -21,6 +21,7 @@
 # Written by Ralph Bean <rbean@redhat.com>
 
 import os
+from module_build_service.utils import to_text_type
 
 from tests.test_models import init_data, module_build_from_modulemd
 from tests import (init_data as init_data_contexts, clean_database)
@@ -66,7 +67,7 @@ class TestModels:
             os.path.dirname(__file__), '..', 'staged_data', 'testmodule_dependencies.yaml')
         mmd = Modulemd.Module.new_from_file(yaml_path)
         mmd.upgrade()
-        build.modulemd = mmd.dumps()
+        build.modulemd = to_text_type(mmd.dumps())
         (build.ref_build_context, build.build_context, build.runtime_context,
          build.context) = ModuleBuild.contexts_from_mmd(build.modulemd)
         assert build.ref_build_context == 'f6e2aeec7576196241b9afa0b6b22acf2b6873d7'
@@ -85,7 +86,7 @@ class TestModels:
         mmd.upgrade()
         with make_session(conf) as session:
             for i in range(3):
-                build = module_build_from_modulemd(mmd.dumps())
+                build = module_build_from_modulemd(to_text_type(mmd.dumps()))
                 build.build_context = 'f6e2aeec7576196241b9afa0b6b22acf2b6873d' + str(i)
                 build.runtime_context = 'bbc84c7b817ab3dd54916c0bcd6c6bdf512f7f9c' + str(i)
                 session.add(build)