fix(SME): degrade gracefully when manual SME has no visiumDir

[atuldeshpande]

Summary

Reported by @dimalvovs:

```

sme_undirected <- SpaceMarkers(sme, directed=FALSE)
Preparing data from SpaceMarkersExperiment...
Computing optimal parameters...
Error in .apply_sme_filters(sme, ...) :
spatial_params(sme) is NULL and no visiumDir is stored in
sme@spacemarkers$params$visiumDir.
```

A user who builds an SME by hand — no `load10X()`, so no `visiumDir` and no scalefactors JSON to read — hits a hard `stop()` even though the underlying data.frame pipeline functions (`find_pattern_hotspots`, `calculate_influence`) already handle `params = NULL` by falling back to `sigma = 10`, `threshold = 4`.

Fix

In `.apply_sme_filters`, when both `spatial_params(sme)` and `visiumDir` are missing, build a default `optParams` matrix from the user's `sigma` argument (or 10) and `threshold` argument, populate it for every pattern column, attach it to `spatial_params(sme)`, and emit an informative `message()` rather than a fatal error. Users who want tuned values per pattern can still set `spatial_params(sme)` explicitly before calling.

Test

Regression case in `tests/testthat/test-SME-default-params.R` that:

• Builds an SME by hand with no visiumDir
• Calls `.apply_sme_filters`
• Asserts a `message()` (not a `stop()`) and verifies the resulting `spatial_params` slot has `sigmaOpt = 10`, `threshOpt = 4` for every pattern

5 PASS, 0 FAIL.

🤖 Generated with Claude Code

[Copilot]

Pull request overview

This PR updates the SME-based workflow to avoid a hard failure when spatial_params(sme) is missing and the SME was created manually (so sme@spacemarkers$params$visiumDir is not available to read Visium scalefactors). Instead, it populates default spatial parameters and continues.

Changes:

• In .apply_sme_filters(), when both spatial_params(sme) and visiumDir are missing, create a default spatial_params matrix using sigma (or 10) and threshold, and emit a message() instead of stop().
• Keep the existing “compute optimal parameters” path when visiumDir is available.
• Add a regression test ensuring manual SMEs without visiumDir get default params and only emit a message.

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.

File	Description
R/SpaceMarkers.R	Implements the graceful fallback to default spatial params for manual SMEs with no visiumDir.
tests/testthat/test-SME-default-params.R	Adds a regression test validating the new fallback behavior and message emission.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[dimalvovs]

this should work with no error, as sigma is to set to 10, but how important that is to be able to find an optimal sigma? also, can't the spot diameter (looks like the only thing needed for the calc here) be read from the SpatialExperiment object?

[atuldeshpande]

Looks good.