autoplot ftml files need the path and file stored

[JanEisermann]

Purpose of PR?:

Fixes #2714

Does this PR introduce a breaking change?

The configuration now stores the flight track with its path and --fpath overrides the
directory. This fixes the two cases the GUI still got wrong:

• a configuration selected in the dockwidget which names the flight track without a path
  was looked up in the working directory msui happens to be started in, not next to the
  configuration file

• a row whose file is not there killed the download through SystemExit, with a path the
  user never typed and no hint where it came from. For example:

The flight track file of 'flight1' does not exist:
/home/mss/example.ftml

The configuration /home/mss/mssautoplot.json names it 'example.ftml', without a
directory, so it is looked up next to that file.
Correct it there, or open the flight track in the MSUI, save it and add the row again.

or

The flight track file of 'flight1' does not exist:
/home/mss/flight1.ftml

The flight track was never saved, only its name is stored, so the file is looked up in
the working directory /home/mss.
Save the flight track and add the row again.

If the changes in this PR are manually verified, list down the scenarios covered::
tests succeeded, manually verified: Download Plots button works correctly and hints are shown when it fails (see above)

Additional information for reviewer? :
Mention if this PR is part of any design or a continuation of previous PRs

Does this PR results in some Documentation changes?
docs/autoplot_dock_widget.rst: how a selected configuration resolves a relative flight
track, and what the Download Plots button does with a file which is not there.

Checklist:

• Bug fix. Fixes autoplot ftml files need the path and file stored #2714
• New feature (Non-API breaking changes that adds functionality)
• PR Title follows the convention of <type>: <subject>
• Commit has unit tests

coded with help by Claude

[ReimarBauer]

1. mslib/utils/config.py:668 — tighten the string-match fix (correctness, most severe)

What's there now: compare_data() was changed to accept any string as a match for any string-typed default — this was meant to fix comparison for the automated_plotting_flights path field, but it's not scoped to that field.
Fix: Scope the relaxed check to the specific field/type this PR is about (e.g. keep using match_type()'s Url/Filepath/Str classification for other string keys like data_dir, wms_cache, and only relax it for the path-list field being changed).

2. mslib/msui/autoplot_dockwidget.py:384 — wrong error message for legacy configs (correctness)

What's there now: missing_flighttrack_message() assumes flighttrack_sources.get(entry) being None + no parent dir means "never saved." But flighttrack_sources is only populated inside configure_from_path() — if the config loaded as the default startup config, it's empty even for legitimately-saved legacy (bare-filename) entries.
Fix: Either populate flighttrack_sources on default-config load too, or don't conclude "never saved" purely from the dict being empty — check whether a config file is actually the source before choosing that message.

3. mslib/msui/autoplot_dockwidget.py:307 — path-collision overwrites (correctness)

What's there now: The dict comprehension building flighttrack_sources keys by resolved absolute path, so two rows that resolve to the same file (one written as a bare filename, one as an absolute path) collide and the later one silently wins.
Fix: Decide desired behavior on collision (keep first, warn, or merge) instead of silent last-write-wins — at minimum a comment explaining why last-write-wins is acceptable, if it is.

4. mslib/msui/autoplot_dockwidget.py:358 — redundant resolution (efficiency, minor)

What's there now: missing_flighttrack() calls resolve_ftml_path() on every row again, even though resolve_flights_paths() already resolved them at load time.
Fix: Reuse the already-resolved paths (e.g. from flighttrack_sources keys) instead of re-resolving, or factor resolution into one shared call site.

Items 1–3 are real correctness bugs worth fixing before merge; item 4 is a nice-to-have cleanup.

[ReimarBauer]

this also drops the cbValidTime.setCurrentText(vtime); valid_time_changed()

usually we csn only change these when a layer was choosen. So that should be ok.

[ReimarBauer]

on recent windows systems "~" is not working as before.

[ReimarBauer]

Impact: Allows ANY string through if default is string (simpler but may be too permissive)

develop version (commit eff4ce837) — More targeted fix:

if isinstance(match_type(default), UrlType) and isinstance(match_type(user_data), StrType):
    return user_data, True

Impact: Only allows string→string when the typed match marks it as a URL (more restrictive, closer to reviewer's intent)

Reviewer's Intent (1): Type checks should only relax for StrType defaults, not FilepathType or UrlType. The develop version appears to implement this more precisely by checking the typed result.

[ReimarBauer]

see comments

[ReimarBauer]

1. One missing flight-track file now stops the whole command-line run (medium, mslib/autoplot/init.py:285)

• What changed: read_ftml() now goes through ftml_file(), which calls SystemExit when the file is missing. draw() catches errors with except Exception, which doesn't catch SystemExit, so the whole run ends.
• Before the PR: a missing file later in the list raised FileNotFoundError, which draw() caught. It printed the error and went on to the next file.
• Example: automated_plotting_flights lists [A.ftml, B.ftml, C.ftml] and B.ftml has been moved. Plots for C.ftml are never made, and a GUI run gets no SUCCESS/FAILURE summary.
• Where it shows up: MSUI checks for missing files up front, which hides this there. It still affects the standalone mssautoplot command, and files deleted between that check and the download.
• Fix: raise a normal exception in ftml_file() instead of calling SystemExit, or catch it per file in draw().

2. The missing-file dialog can give the wrong reason (low, mslib/msui/autoplot_dockwidget.py:379)

• What happens: missing_flighttrack_message tells the user the configuration named the file "without a directory, so it is looked up next to that file" whenever the stored name differs from the resolved path.
• Why that's wrong: resolve_flights_paths also changes names that do include a directory. It expands ~, joins relative paths such as sub/x.ftml onto the config's directory, and follows symlinks and ...
• Example: the config names /data/x.ftml, where /data is a symlink to /mnt/data. The dialog says the file was named without a directory, which is false and points the user to the wrong place.
• Fix: only use that wording when the stored name has no directory part.

[ReimarBauer]

Only automated_plotting_flights needs to accept a path, but the PR relaxes compare_data for every option whose default is a string. The fix is to make that exception explicit, using the same pattern the file already uses for fixed_list_options.

1. List the options that accept any string (mslib/utils/config.py, near line 308):
   # List options with fixed length
   fixed_list_options = ["MSCOLAB_timeout", ]
   # List options whose string entries take any string, e.g. a flight track path
   free_string_list_options = ["automated_plotting_flights", ]
   Also add "free_string_list_options" to the del default_options[key] list (around line 414). Otherwise it would show up as a user option.

2. Pass a flag only from the list-structure loop in merge_dict (around line 657):
   data, match = compare_data(
   los_key_item, new_dict[key][i],
   any_string=key in MSUIDefaultConfig.free_string_list_options)

3. Replace the broad StrType rule in compare_data with the flag:
   def compare_data(default, user_data, any_string=False):
   ...
   if not isinstance(default, dict) and not isinstance(default, list):
   if isinstance(default, float) and isinstance(user_data, int):
   user_data = float(default)
   if any_string and isinstance(default, str) and isinstance(user_data, str):
   return user_data, True
   # ... existing UrlType / type comparison unchanged
   The flag also has to be passed on in both recursive calls (list and dict branches), because the flight row is a list nested inside a list.

Why this is better:

• Options like filepicker_default, MSCOLAB_category, new_flighttradule/function entries and CRS go back to their old checks. Nothingoutside autoplot changes.
• Relative names such as example.ftml still work in the flight rowe PR. They count as StrType, not FilepathType.
• The long comment explaining which defaults are paths or URLs can go, because the rule no longer reaches them.
• The two tests the PR added (test_user_option_with_path, test_patuld still pass. Add one test showing that a string option outsidethe list, e.g. filepicker_default: "/tmp/x", still falls back to its default.

[ReimarBauer]

see comment