Add Metis Map by aburtovoi · Pull Request #8519 · sunpy/sunpy

aburtovoi · 2026-02-19T14:20:18Z

We have created Issue #8518, requesting integration on METISMap Class to the Sunpy core package.

In this respect, we have modified the following files of the package:

sunpy/sunpy/map/sources/solo.py
sunpy/visualization/colormaps/color_tables.py
sunpy/visualization/colormaps/cm.py
sunpy/sunpy/map/sources/tests/test_metis_source.py

Accomplishments: - Implemented 12 validation tests for METISMap. - Verified WCS, metadata fallback, and colormap logic. - Implemented mask_bad_pix and get_fov_rsun tests. - Cleaned code with Ruff (fixed F401 and PT011).

Metis map gio

Cadair · 2026-02-19T14:43:18Z

Hi @aburtovoi and everyone.

Thanks so much for opening this PR, I'm really excited to get more support for different Solar Orbiter data into sunpy.

The first thing we need to talk about is the METIS specific methods you have added to this class. The approach we take in core is that all the different map sources have the same API as GenericMap, i.e. no extra methods or properties. The motivation for this is that if you write some code for one EUV imager you should be able to drop a map from another one in etc. i.e. code written for one map should work for others.

We should think about where this METIS functionality would best live, could you talk me through the various things you have added so we can figure it out?

samaloney · 2026-02-20T15:21:48Z

This is really cool, I seen this in action at a recent meeting thanks @aburtovoi. I'd suggest separating this into two broad chunks: getting the Metis data into the map container and Metis-specific processing.

As these are Solar Orbiter fits files, they are pretty compliant with the FITS standard. All we need is the colour table and normalisation thresholds, which should stay in the map source but use the the meta data .meta['btype'] rather than the custom attributes?

The masking functions could be moved to be plain functions rather than methods taking and returning MetisMaps and either live in the solo map module or sunkit-instruments?

def mask_occulter(metis_map: MetisMap):
    # occulter masking stuff
    return metis_map  #or Map(masked_data, meta)

def mask_bad_qualtiy(metis_map, MetisMap, quality_array)
    # bad quality masking stuff
    return metis_map  #or Map(masked_data, meta)

samaloney

Thanks again for the PR all the pieces are there may just need to move some of them around.

samaloney · 2026-02-20T15:28:48Z

+        rsun_deg = self.rsun_obs.value / 3600.0  # Convert arcsec to deg
+        rmin_rsun = self.meta["inn_fov"] / rsun_deg  # Inner FOV in rsun
+        rmax_rsun = self.meta["out_fov"] / rsun_deg  # Outer FOV in rsun
+        board_deg = 2.9  # Detector board edge in deg
+        board_rsun = board_deg / rsun_deg  # Board radius in rsun


This maybe easier to read, less error prone by using the unit rather than dropping:

Suggested change

rsun_deg = self.rsun_obs.value / 3600.0 # Convert arcsec to deg

rmin_rsun = self.meta["inn_fov"] / rsun_deg # Inner FOV in rsun

rmax_rsun = self.meta["out_fov"] / rsun_deg # Outer FOV in rsun

board_deg = 2.9 # Detector board edge in deg

board_rsun = board_deg / rsun_deg # Board radius in rsun

rsun_deg = self.rsun_obs.to(u.deg) # Convert arcsec to deg

rmin_rsun = (self.meta["inn_fov"]*u.deg) = / rsun_deg # Inner FOV in rsun

rmax_rsun = (self.meta["out_fov"]*u.deg) / rsun_deg # Outer FOV in rsun

board_deg = 2.9 * u.deg # Detector board edge in deg

board_rsun = board_deg / rsun_deg # Board radius in rsun

but here it looks like board_rsun would be unit less so I've gotten something wrong?

I corrected this part as suggested.
Correct, rmin_rsun, rmax_rsun, board_rsun are in a unit of Rsun (i.e., unitless).

samaloney · 2026-02-20T16:15:26Z

+        inn_fov = self.meta["inn_fov"] * 3600 / self.meta["cdelt1"]  # in pix
+        out_fov = self.meta["out_fov"] * 3600 / self.meta["cdelt2"]  # in pix
+
+        # Create coordinate grids
+        x = np.arange(0, self.meta["naxis1"], 1)
+        y = np.arange(0, self.meta["naxis2"], 1)
+        xx, yy = np.meshgrid(x, y, sparse=True)
+
+        # Calculate distance from internal occulter center
+        in_xcen = self.meta["io_xcen"]
+        in_ycen = self.meta["io_ycen"]
+        dist_inncen = np.sqrt((xx - in_xcen) ** 2 + (yy - in_ycen) ** 2)


Usually prefer to try and use the map attributes, units and wcs methods so along the lines of

inn_fov = (self.meta["inn_fov"] * u.deg / self.scale[0]).decompose() .... in_xcen = self.meta["io_xcen"] * u.pix ... dist_inncen = np.sqrt((xx - in_xcen) ** 2 + (yy - in_ycen) ** 2)

Corrected, thanks!

samaloney · 2026-02-20T16:18:10Z

+# ============================================================================
+
+
+@pytest.fixture


Could parametrise this test with different header value to test the other code paths

samaloney · 2026-02-20T16:35:57Z

+        instrume = header.get("INSTRUME", "").strip().upper()
+        obsrvtry = header.get("OBSRVTRY", "").strip().upper()
+
+        return ("METIS" in instrume) and ("SOLAR ORBITER" in obsrvtry)


Should this source work or load only L2 files if so should add to the check here.

Yes, for the moment METISMap works only with L2 Metis data. Added the corresponding check

aburtovoi

Reviewed solo.py code (class METISMap) as requested

aburtovoi · 2026-02-25T10:11:10Z

+        instrume = header.get("INSTRUME", "").strip().upper()
+        obsrvtry = header.get("OBSRVTRY", "").strip().upper()
+
+        return ("METIS" in instrume) and ("SOLAR ORBITER" in obsrvtry)


Yes, for the moment METISMap works only with L2 Metis data. Added the corresponding check

aburtovoi · 2026-02-25T10:35:56Z

+        rsun_deg = self.rsun_obs.value / 3600.0  # Convert arcsec to deg
+        rmin_rsun = self.meta["inn_fov"] / rsun_deg  # Inner FOV in rsun
+        rmax_rsun = self.meta["out_fov"] / rsun_deg  # Outer FOV in rsun
+        board_deg = 2.9  # Detector board edge in deg
+        board_rsun = board_deg / rsun_deg  # Board radius in rsun


I corrected this part as suggested.
Correct, rmin_rsun, rmax_rsun, board_rsun are in a unit of Rsun (i.e., unitless).

aburtovoi · 2026-02-25T11:04:37Z

+        inn_fov = self.meta["inn_fov"] * 3600 / self.meta["cdelt1"]  # in pix
+        out_fov = self.meta["out_fov"] * 3600 / self.meta["cdelt2"]  # in pix
+
+        # Create coordinate grids
+        x = np.arange(0, self.meta["naxis1"], 1)
+        y = np.arange(0, self.meta["naxis2"], 1)
+        xx, yy = np.meshgrid(x, y, sparse=True)
+
+        # Calculate distance from internal occulter center
+        in_xcen = self.meta["io_xcen"]
+        in_ycen = self.meta["io_ycen"]
+        dist_inncen = np.sqrt((xx - in_xcen) ** 2 + (yy - in_ycen) ** 2)


Corrected, thanks!

samaloney · 2026-02-25T12:44:48Z

I don't think cmcrameri is a default matplotlib colormap which is causing some errors in the doc builds?

ayshih · 2026-02-25T19:23:28Z

+    # cmap_name, cmap, cmap.name
+    cmap_dict = {
+        'solometisvl-tb': (
+            cmcrameri.cm.batlow.copy(), 'SolO Metis VL Total Brightness'


Can we use matplotlib's (somewhat similar) "cividis" instead? Or is it important to use specifically Crameri's "batlow"?

wtbarnes · 2026-02-26T03:47:13Z

It's great to see more map sources, especially for Solar Orbiter data so thanks very much for this contribution! There was some discussion about this at the dev meeting today which I'll try my best to summarize. @Cadair or @samaloney can jump in if I've misrepresented something.

As @Cadair noted in his initial comment, GenericMap subclasses should all have the same API, i.e. subclasses should not introduce additional (public) methods. From what I can tell, almost all of the additional methods here are related to either building the mask or setting the appropriate plot settings given the masked array. Given that, I would suggest doing the following:

Make all methods added here that do not already exist on GenericMap private
In the constructor, use the aforementioned private methods to set the mask property on the map. This will then propagate through to all of the map visualization methods (e.g. https://docs.sunpy.org/en/stable/generated/gallery/computer_vision_techniques/mask_disk.html) while leaving the original data untouched.
Also in the constructor, but following the step above, adjust plot settings as needed, using the mask that you've just constructed.

wtbarnes · 2026-02-26T03:50:15Z

+    @property
+    def prodtype(self):
+        """Product type identifier for this METIS data."""
+        return self._prodtype
+
+    @prodtype.setter
+    def prodtype(self, value):
+        raise AttributeError("Cannot manually set prodtype for METISMap")


Re: previous comments, prodtype is not a property exposed on other subclasses. I would suggest instead overriding measurement and just moving the logic of get_prodtype here (or just calling it as a private method _get_prodtype).

- adopting pr 8519 - changed prodtype to measurement from sunpy#8519 comments - changed cmaps to cividis from sunpy#8519 comments - changed pytest to use metis_map where tests are agnostic to specific header values from sunpy#8519 comments

aburtovoi and others added 9 commits September 15, 2025 15:38

METISMap integration

ab17197

updating sunpy/map/sources/solo.py

3a37ecb

updating metis input in solo.py, cm.py and color_tables.py

5923079

updating for DR2

2d95df3

updating solo.py, resolving ImageNormalize bug

2933f10

Integrated METISMap improvements, quality matrix masking, and WCS tests

a58aa6e

Add METISMap test suite and finalize solo.py integration

c315541

Accomplishments: - Implemented 12 validation tests for METISMap. - Verified WCS, metadata fallback, and colormap logic. - Implemented mask_bad_pix and get_fov_rsun tests. - Cleaned code with Ruff (fixed F401 and PT011).

Merge pull request #1 from gjerse/metis_map_gio

0abcdf1

Metis map gio

Update solo.py

793295f

samaloney requested changes Feb 20, 2026

View reviewed changes

samaloney reviewed Feb 20, 2026

View reviewed changes

Comment thread sunpy/map/sources/solo.py Outdated

samaloney added the map Affects the map submodule label Feb 20, 2026

Reviewing Metis Map code

728f590

aburtovoi commented Feb 25, 2026

View reviewed changes

aburtovoi closed this Feb 25, 2026

aburtovoi reopened this Feb 25, 2026

samaloney reviewed Feb 25, 2026

View reviewed changes

ayshih reviewed Feb 25, 2026

View reviewed changes

wtbarnes reviewed Feb 26, 2026

View reviewed changes

wtbarnes mentioned this pull request Mar 26, 2026

Should you be able to set mask inplace on a Map? #3166

Open

Cadair added this to the 8.0.0 milestone Apr 22, 2026

Cadair added the Needs Adoption PRs that went stale but should be picked up again and completed. label May 13, 2026

Hermanlrx mentioned this pull request May 20, 2026

Adds metis map #8641

Open

5 tasks

		# ============================================================================


		@pytest.fixture

Uh oh!

Conversation

aburtovoi commented Feb 19, 2026

Uh oh!

Cadair commented Feb 19, 2026

Uh oh!

samaloney commented Feb 20, 2026

Uh oh!

samaloney left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

aburtovoi left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

wtbarnes commented Feb 26, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

6 participants