Extract duplicated import-continuation while loops into collect_import_continuation

Both parse.file_contents and identify.imports contained an almost-identical
~37-line block of while loops for collecting multi-line import statements
(parenthesised and backslash-continued forms).  The only differences were:
  * parse.py accesses lines via in_lines[index] (list + mutable index)
  * identify.py accesses lines via next(indexed_input) (iterator)
  * parse.py additionally tracks comments, nested_comments, and raw_lines

New helper in _parse_utils.py:

  collect_import_continuation(line, import_string, get_next_line,
                              line_separator="\n")
    -> (final_line, complete_import_string, extra_lines)

Both callers provide a tiny get_next_line closure that abstracts over their
respective data source.  The extra_lines list carries (line, comment,
appended_with_separator) tuples so parse.py can reconstruct its per-line
side-effects; identify.py ignores it with _.

parse.py: 70-line while block -> 25-line call site (+ post-processing loop)
identify.py: 37-line while block -> 6-line call site

Co-authored-by: DanielNoord <13665637+DanielNoord@users.noreply.github.com>