Can’t make sense of CalculiX errors? I built an open-source tool to help diagnose them

Tool interfaces
1

One of the most frustrating parts of using CalculiX, at least for me, is not always the modeling itself, but what happens after you hit solve.

Sometimes the job crashes with a wall of warnings and errors.
Sometimes it technically finishes, but the results still look suspicious.
The hard part is that you often do not know where the real problem is, or what you should fix first.

I ran into this enough times that I started building a small open-source tool called cae-cli.

What it does is simple: it reads CalculiX outputs and error messages, then gives you a first-pass diagnosis of the model. It tells you:

  • what problems were detected

  • which one should be handled first

  • the likely root cause

  • possible fixes

  • and for a few low-risk cases, whether a safe automatic fix is possible

image
image765×816 14.9 KB

One thing I care about a lot is reliability.
This is not just an AI wrapper guessing at errors.

The current diagnosis flow is mainly rule-based. The rules are derived from CalculiX source-level error patterns and common failure cases, and the tool also includes reference data from 638 official test cases.
AI is used more for explanation and assistance, not as the only decision-maker.

So the goal is not to “do the simulation for you.”
The goal is to make the most painful step easier:

when something goes wrong, you can quickly see what failed, why it likely failed, and what to check first.

Right now the project is still under active development, but I think it is already useful for people who:

  • are new to CalculiX and struggle to interpret errors

  • get completed runs with questionable results and want a quick diagnostic pass

  • want a more structured troubleshooting workflow

Install:

pip install cae-cxx
cae install

GitHub:
https://github.com/yd5768365-hue/cae-cli

image
image1579×794 34.2 KB

image
image1652×765 32.6 KB

I’m a first-year mechanical engineering student, and this project is currently being built mostly by myself.
If you want to try it, or if you have a CalculiX error case that gave you a hard time, feel free to comment or open an issue. I would especially like to collect real-world failure cases and improve the rule set based on them.

2

IMO CalculiX should have more syntax checks and clear error messages for syntax error. So-called datacheck stage in Abaqus thoroughly scans the whole input deck and manages to clearly point out the majority of the mistakes. Pretty much only some rare cases of misplaced keywords don’t trigger it. In CalculiX, some keyword combinations don’t generate errors even though they most definitely should.

Perhaps we should create a GitHub issue and keep listing potential syntax checks that could be implemented to improve the user experience.