Check a DB path for VIROGENESIS tasks

When one tries to build a database and specifies an output folder where another database or its draft exists, the workflow works incorrectly. It happens when you launch, for example, the same workflow twice.

See a comment in UGENE-5836 about the Kraken database. Here is the scenario for CLARK:

• Add the "Build CLARK database" element to the scene and set some appropriate input data.
• Run the workflow.
  Current state: you can see in the log that CLARK is executing.
• Cancel the task.
  Current state: the database is in an invalid state.
• Launch it again.
  Current state: the workflow ends successfully, but CLARK actually does nothing.