Select multiple hashfiles during job creation, combined into a new hashfile

Summary

During job creation, the "Use existing" hashfile picker only allows selecting a
single hashfile at a time. Allow selecting multiple existing hashfiles and
combine their hashes into one new auto-named hashfile for the job.

Motivation

Operators frequently accumulate hashes across several uploads (e.g. per-host
pwdumps, staged captures) and want to crack them together in one job. Today the
only workaround is manually merging files outside Hashview and re-uploading,
which loses the per-source hashfiles and duplicates data.

Current behavior

• UI: single-select radio per row
  hashview/templates/jobs_assigned_hashfiles.html.j2:238

  <input type="radio" name="hashfile_id" value="{{ hashfile.id }}" ...>
  

• Route: reads one hashfile_id and assigns it directly to the job
  hashview/jobs/routes.py:434-438

  elif request.method == 'POST' and request.form.get('hashfile_id'):
      job.hashfile_id = request.form['hashfile_id']
  

• A job references exactly one Hashfiles.id; there is no notion of combining files.

Proposed behavior

1. Change the "Use existing" picker from radio to checkboxes so multiple
   hashfiles can be selected.
2. On submit with 2+ selections, create a new Hashfiles record and populate it
   by unioning the HashfileHashes rows of the selected files, then assign it as
   the job's hashfile_id.
3. Name is auto-generated (e.g. combined-<timestamp> or derived from the
   source names) to avoid collisions — the user does not name it.
4. Single selection keeps today's behavior (assign existing hashfile_id
   directly, no new record).

Constraints

• Mixed-customer combining is blocked. The job-creation flow is already
  scoped to a single customer, so the "Use existing" list only shows that
  customer's hashfiles — combining across customers is not reachable in the
  current UI and must remain so. The combined Hashfiles record inherits that
  single customer_id.

Deduplication

Hash dedup infrastructure already exists and should be reused:

• import_hash_only() dedupes by (hash_type, MD5(ciphertext))
  (hashview/utils/utils.py:544-555).
• HashfileHashes is already many-to-many, so one hash can belong to the
  combined file and its sources.
  Combining should produce a distinct set of hash_ids (respecting username
  where present) — no duplicate HashfileHashes rows in the combined file.

No schema changes

The combined hashfile is a normal saved Hashfiles record — no new columns and
no Alembic migration needed. Jobs still references a single hashfile_id.

Affected files

• hashview/templates/jobs_assigned_hashfiles.html.j2
• hashview/jobs/routes.py (jobs_assigned_hashfile)
• hashview/utils/utils.py (new combine_hashfiles() helper)

Start in hashview/templates/jobs_assigned_hashfiles.html.j2 and hashview/jobs/routes.py at the picker and POST handling, then read import_hash_only() in hashview/utils/utils.py and the HashfileHashes relationships. Confirm how selected hash IDs, usernames, customer IDs, and auto-generated names should be represented; done means single selections still assign directly and multiple selections create one deduplicated combined Hashfiles record assigned to the job.