Niklas Kappel

Preserve table of contents when editing PDF

1

I use the following simple combination of `PdfReader` and `PdfWriter` to edit PDF files. ``` import pdfrw def main(): pdf = pdfrw.PdfReader(open("input.pdf", "rb")) writer = pdfrw.PdfWriter() for pageNumber in range(pdf.numPages):...

Enable completion for custom macro

1

Say I define a macro like this: ``` \newcommand{\mySuperGraphic}[1]{ ... \includegraphics{#1} ... } ``` Then I type (with `|` being the cursor position): ``` \mySuperGraphic{| ``` vimtex does not trigger...

This fixes misaligned annotations in a particular ebook I have as well as in the `tests/annotation_test` testcase. Optionally the fix could be implemented in the class `PDFPageLayout` instead.

Incompatible types between `Dataset.__getitem__` and `DataLoader.__init__`

1

### 🐛 Describe the bug For the following code (extracted from [here](https://colab.research.google.com/drive/1I8a0DfQ3fI7Njc62__mVXUlcAleUclnb?usp=sharing)) ```Python from torch_geometric.datasets import TUDataset from torch_geometric.loader import DataLoader dataset = TUDataset(root="data/TUDataset", name="MUTAG") train_dataset = dataset[:150] train_loader =...

ENG-654: Neptune reports "step must be strictly increasing" error if lightning logs in training and validation step

6

### Describe the bug Using the Neptune logger in lightning, I get multiple of the following errors: ``` [neptune] [error ] Error occurred during asynchronous operation processing: X-coordinates (step) must...

add typing support for fields

1

Node constructors already work with the addition of 'dataclass_transform' in ZnInit. For the field specifier functions we could lie about the return type. If they don't take arguments (e.g. `outs`)...

Ruff and Pyright duplicate some warnings

1

### Problem description LunarVim sets up both Ruff and Pyright as lsp servers. I would appreciate if someone could explain the rationale behind this choice, but I like the idea...

Can dtypes of later arguments be annotated to follow the dtypes of earlier arguments?

1

Consider the function ```python def fill_array(array: Shaped[Array, "N"], fill_value: Any) -> Shaped[Array, "N"]: return jnp.full_like(array, fill_value) ``` In principle, the dtypes of `fill_value` and the return value should be constrained...