Metadata-Version: 2.4
Name: setupEM
Version: 0.18.3
Summary: Python tool for configuration of gds2palace workflow with GUI.
Author-email: Volker Muehlhaus <volker@muehlhaus.com>
License-Expression: GPL-3.0-or-later
Project-URL: Homepage, https://github.com/VolkerMuehlhaus/setupEM
Project-URL: Repository, https://github.com/VolkerMuehlhaus/setupEM
Project-URL: Changelog, https://github.com/VolkerMuehlhaus/setupEM/blob/main/doc/CHANGES.md
Requires-Python: >=3.9
Description-Content-Type: text/markdown
License-File: LICENSE
Requires-Dist: gds2palace>=0.10.7
Requires-Dist: gds_prepare_for_EM>=1.6.0
Requires-Dist: PySide6
Requires-Dist: shiboken6
Requires-Dist: scipy
Requires-Dist: requests
Requires-Dist: scikit-rf
Requires-Dist: matplotlib
Requires-Dist: numpy
Requires-Dist: gdspy
Requires-Dist: meshio
Requires-Dist: pyvista
Requires-Dist: pyvistaqt
Provides-Extra: test
Requires-Dist: pytest; extra == "test"
Dynamic: license-file

# Python GUI for gds2palace 

![Intro](https://raw.githubusercontent.com/VolkerMuehlhaus/setupEM/main/doc/png/setupEM_banner.png)

## Install

    pip install setupEM

**Full documentation, installation guide, and usage walkthrough:**
https://github.com/VolkerMuehlhaus/setupEM

## Recent changes

# What's New - November 15, 2026 (setupEM 0.18.3)

Needs gds2palace 0.10.7 or newer (`pip install` updates it).

**`combine_extend_snp.py` 1.5** (`combine_snp`, in `scripts/`), the same file as in gds2palace 0.10.7:
- The `_dc` Touchstone file keeps all simulated frequency points and adds only the 0 Hz point in front. Before, a result with the two low frequency points of a sweep from 0 Hz was resampled to a uniform frequency grid, so the simulated points, among them the two low ones, were replaced by interpolated values. The 0 Hz value itself is unchanged.
- Port walls (from gds2palace 0.10.6): the line resistance of the standards is taken from the model's lowest frequency, so R and Q of models with port walls are right also below fstop/20.

**Sweep from 0 Hz, the frequencies that replace DC** (a gds2palace 0.10.7 change, nothing to set in setupEM): with start frequency 0, gds2palace simulates two low frequency points instead of 0 Hz. These were always 10 MHz and 20 MHz. With a mesh finer than 0.5 µm (refined cellsize or a per-layer override), gds2palace 0.10.7 moves them up, to 5 MHz·µm divided by the smallest cell size and twice that (e.g. 25 and 50 MHz at 0.2 µm), because on such a fine mesh Palace could stop with "SuperLUSolver: Found a singular matrix" at 10 MHz. Models with cell sizes of 0.5 µm and more keep 10 and 20 MHz. The user's guide section on DC handling says this now.

**Citation:** `CITATION.cff` and a "How to cite" section in the README; GitHub's **Cite this repository** button gives APA and BibTeX entries.

**Internal:** the model script writer of setupEM and setupThermal is now a module without Qt (`script_writer.py`, issue #52). The written scripts are unchanged; `tests/test_script_writer.py` checks 8 recorded scripts against both the GUI and the writer.

# What's New - November 13, 2026 (setupEM 0.18.2)

**Ports and thermal objects:** Return in a text field of the form (source layer, Z0, voltage; in setupThermal layer, power, temperature) applies the settings to the selected row, like the Apply button. With no row selected, the status bar says so.

**Command line fixes** (from issue #52):
- A script from File > Export runs without a display: gds2palace 0.10.4 and newer builds the model without the gmsh window on Linux with no display (ssh, container, batch job). Before, the script stopped with "Can't open display". The exported script also has the switch as a comment, `# settings['no_gui'] = True`, for unattended runs where a display exists.
- `combine_extend_snp.py` (`combine_snp`, in `scripts/`) has `--help` and `--version`, and takes the folder to search as an optional argument. Before, any option was ignored and the conversion ran. Same file as in gds2palace 0.10.4.

**Scripts used on two machines** (e.g. Windows and Linux):
- A model script edited in place whose GDS file was taken from its folder on opening (the script had the other machine's path, e.g. `/mnt/pub/...` for `P:/...`) now gets that path when it is saved or the model is created; the Code tab shows it. Before, the GUI showed the new path and the script kept the old one, and the model was made from the old one.
- On Linux, a path with a drive letter from Windows (`D:/...`) in a script or settings file is no longer taken as relative to the script's folder (`/mnt/.../D:/...`, "Load a GDSII file and XML stackup first"). A stackup that is not found, also not next to the script, is taken from the stackups that ship with gds2palace if one has that name; the log says so.

# What's New - November 11, 2026 (setupEM 0.18.1)

**Command line fixes** (from issue #52, use without a display):
- `--help`, `--version` and a wrong argument of `setupEM`, `setupThermal`, `stackupEditor`, `resultViewer` and `fieldViewer` no longer need a display: the arguments are read before Qt starts. All five have `--version` now.
- `setupEM -simcfg <file>` opens a settings file of an Elmer model in Elmer mode. It opened in Palace mode unless `--elmer` was given, also when started from the KLayout plug-in.
- A `-simcfg`, `-gdsfile` or `-xmlfile` that does not exist, or a settings file of the other application, ends `setupEM` / `setupThermal` with a message on stderr and exit code 2 instead of an empty window. In the GUI, a settings file that is gone is reported as "File not found", not as "Unknown data format".
- Loading a settings file no longer shows a "Loaded" box: the log says "Config loaded from ...", with a GDS or stackup file that was not found at its saved path and was taken from the settings file's folder instead.
- Opening a `.simcfg` continues with the model script only when the file is linked to it, and asks only when the two have other settings (the question lists them). Create Model links: besides `<model>.simcfg` it writes the script's name (`"model_script"`) into the `.simcfg` the model was opened from, e.g. the KLayout plug-in's export, which keeps it on the next export. A script that only reads the same GDSII file is no longer taken for the model; the log names it. New options `--continue-with-script` and `--start-from-config` answer the question on the command line.
- `fieldViewer --screenshot` needs no display and no `QT_QPA_PLATFORM=offscreen`.
- `python -m setupEM.join_pieces` and `python -m setupEM.thermal_results` have `--help` and `--version`, write their errors to stderr, and a missing file is a message, not a traceback.
- The modules without a GUI (`join_pieces`, `thermal_results`, `script_model`, ...) can be imported without loading Qt.
- User guide: Model Fit starts snp2le with the result file loaded (the text said it has to be loaded by hand).

**Split regions, fix:**a side of the region made a port wall in setupEM after the export (e.g. with via ports of the layout on the region's edge) was a plain PEC wall in every piece and in the full check model: the wall shorted those ports, and the joined result and the full model were both wrong, in the same way. The pieces' port walls on the region's edge are now found from the ports on the edge of each piece's layout; the full model has all of the region's port walls. Where the region has a port wall but a piece none of its ports, that stretch stays PEC and now moves outward with the region's port wall extension (a length; written as `settings['wall_extension']`, needs gds2palace 0.10.1), so that all pieces and the full model have the same box.

Full history: [CHANGES.md](https://github.com/VolkerMuehlhaus/setupEM/blob/main/doc/CHANGES.md)

---

**Note:** the `setupEM` PyPI package itself requires:

- gds2palace>=0.10.7
- gds_prepare_for_EM>=1.6.0
- PySide6
- shiboken6
- scipy
- requests
- scikit-rf
- matplotlib
- numpy
- gdspy
- meshio
- pyvista
- pyvistaqt

(Other Python modules mentioned above are only needed to run standalone helper scripts in this repository, not to use the installed package.)
