Currently `Structure` has 4 identifiers which may or may not be set
depending on the source, file format, and parsing method. These should
be revisited and harmonized in the future, but for now I just document
them in this test.

Summary of results:

| Format | With ID | Parse Method    | PdbId | Identifier | Name | StructureIdentifier |
|--------|---------|-----------------|-------|------------|------|---------------------|
| cif    | No      | fromInputStream | null  | ""         | ""   | null                |
| pdb    | No      | fromInputStream | null  | ""         | ""   | null                |
| cif    | No      | fromUrl         | null  | ""         | ""   | null                |
| cif    | No      | StructureIO     | null  | file:      | 5PTI | file:               |
| pdb    | No      | StructureIO     | null  | file:      | ""   | file:               |
| cif    | Yes     | fromInputStream | 5PTI  | ""         | ""   | null                |
| pdb    | Yes     | fromInputStream | 5PTI  | ""         | ""   | null                |
| cif    | Yes     | fromUrl         | 5PTI  | ""         | ""   | null                |
| cif    | Yes     | StructureIO     | 5PTI  | file:      | 5PTI | file:               |
| pdb    | Yes     | StructureIO     | 5PTI  | file:      | 5PTI | file:               |

- `getPdbId` reflects the ID parsed from the file
- Only StructureIO is setting the StructureIdentifier (by design, but
  indicates StructureIO should be used wherever possible)
- StructureIO does not set the name properly for PDB files. This should
  be fixed.

The core URLIdentifier bug was fixed in biojava#1020, but this improves the
test.

The parser changes are mostly for logging, and are there to indicate that
null PdbId is an expected situation rather than an error.