add geqdsk_sign_convention and geometry_type as None in the ctx dict.

[Antoinehoff]

related to #179.

We refactor the structure

ctx['a_field'] = None
...
if ctx['a_field'] is not None:

but simply do now

ctx = {}
...
if 'a_field' in ctx.keys():

Also, to print in a raw format all keys, present in the metadata, that were not printed yet, we store the key of all previously printed keys in a local list that will be the comparison point instead of having a manually written list, going from

    for key, val in self.ctx.items():
      if key not in ["time", "frame", "changeset", "builddate", "basis_type",
          "poly_order","is_modal", "lower", "upper", "cells", "num_comps",
          "grid_type", "num_cdim", "num_vdim","geometry_type","geqdsk_sign_convention"] and val is not None:
        output += f"\n├─ {key:s}: {val}"

to

    for key, val in self.ctx.items():
      if key not in printed_keys:
        output += f"\n├─ {key:s}: {val}"

We also remove the commented # end statements.

With this branch, the failing example presented in the issue is fixed:

pgkyl rt_gk_tcv_nt_iwl_adapt_src_3x2v_p1-ion_0.gkyl info
(default#0)
├─ Time: 0.000000e+00
├─ Frame: 0
├─ Number of components: 48
├─ Number of dimensions: 5
├─ Grid: (uniform)
│  ├─ Dim 0: Num. cells: 24; Lower: 0.000000e+00; Upper: 1.200000e-01
│  ├─ Dim 1: Num. cells: 16; Lower: -1.049232e-01; Upper: 1.049232e-01
│  ├─ Dim 2: Num. cells: 12; Lower: -3.141593e+00; Upper: 3.141593e+00
│  ├─ Dim 3: Num. cells: 12; Lower: -7.071068e-01; Upper: 7.071068e-01
│  └─ Dim 4: Num. cells: 8; Lower: 0.000000e+00; Upper: 1.000000e+00
├─ Maximum: 2.266065e-06 at (2, 6, 4, 6, 2) component 0
├─ Minimum: -3.585019e-07 at (3, 2, 4, 6, 3) component 5
├─ DG info:
│  ├─ Polynomial Order: 1
│  └─ Basis Type: gkhybrid (modal)
├─ Created with Gkeyll:
│  ├─ Changeset: 5506bbe9d39a
│  └─ Build Date: 2026-02-05 12:55:25 PST
├─ Geometry info:
│  ├─ Type: GKYL_GEOMETRY_MAPC2P

[manauref]

I considered this when I made changes to meta, but I don't think it's the right approach. We need to move away from calling out specific meta in various places in the postgkyl, and instead making the code work with any meta (it doesn't yet). This means that:

• All those specifications of None meta at the top of gdata.py should go away
• The rest of the code should work regardless of what meta is in the file or not.
  See my suggestion in the issue, that improves things a bit I think. But we need to improve it further eventually so that some of those if-statements are replaced with something more abstract.

[Antoinehoff]

ok well, I am not against more global changes in the structure but here it is just a mistake to not set the dict entry as None as all of the other field of the dict are already set to None. This should be pushed to fix the retro compatibility issue that was introduced. Larger structural code changes can be done but are less urgent.

[manauref]

have you tried the fix mentioned in the issue?

…

On Tue, 10 Feb 2026 at 07:35, Antoine Hoffmann ***@***.***> wrote: *Antoinehoff* left a comment (ammarhakim/postgkyl#180) <#180 (comment)> ok well, I am not against more global changes in the structure but here it is just a mistake to not set the dict entry as None as all of the other field of the dict are already set to None. This should be pushed to fix the retro compatibility issue that was introduced. Larger structural code changes can be done but are less urgent. — Reply to this email directly, view it on GitHub <#180 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAQ6DKHBBQSWS2NKZTT46ED4LHUC7AVCNFSM6AAAAACUTZTLW2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTQNZYGAZDKNRTGU> . You are receiving this because your review was requested.Message ID: ***@***.***>

[Antoinehoff]

It may work too yes but I think it is better to follow the existing structure, and change it globally later if we want to, so I am not favorable to the fix mentioned in the issue.

[manauref]

We need to treat postgkyl with the same rigor and development philosophy as gkeyll, which includes not reproducing bad code that was already there.

…

On Tue, 10 Feb 2026 at 08:31, Antoine Hoffmann ***@***.***> wrote: *Antoinehoff* left a comment (ammarhakim/postgkyl#180) <#180 (comment)> It may work too yes but I think it is better to follow the existing structure, and change it globally later if we want to, so I am not favorable to the fix mentioned in the issue. — Reply to this email directly, view it on GitHub <#180 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAQ6DKDSKBFSEXHFOVSTQB34LH2U3AVCNFSM6AAAAACUTZTLW2VHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZTQNZYGQ4TKOJZGA> . You are receiving this because your review was requested.Message ID: ***@***.***>

[manauref]

unit tests failing