Initialise an `IAM_OnDB_Dataset`. The data of the dataset needs to be stored on disk as follows to be readable by this present class:

Initialise a `PageDataset`. TODO.

Axis-aligned bounding box. ``(x_min, y_min)`` is the top-left corner, ``(x_max, y_max)`` the bottom-right corner. ``label`` is an opt

Collate ``IAM_Words_Element`` batches for the ``DataLoader``.

Collate ``IAM_Dataset_Element`` batches for the ``DataLoader``.

Start the demo container with dummy env vars; stop it after the test.

Build the Docker image once per test session and return its tag.

Processes a document to generate a PDF with handwritten text recognition (HTR) predictions and extracts the image of the first page. Arg

Converts the first page of the input document to an image. This function exports the input document to a PDF format using Xournal++, ext

TODO. Sources: - https://stackoverflow.com/a/40795835 and - https://code.tutsplus.com/collision-detection-using-t

Forward pass. Args: x: (batch, 1, height, width) grayscale image tensor. Returns: (seq_len, batch, num_class

Fixture to retrieve the path to the IAM-OnDB dataset. Downloads and caches the dataset from HuggingFace Hub if not already present locally. R

Fixture to retrieve the path to the minimal test data file. Downloads and caches the file from the HuggingFace Hub dataset ``PellelNitram/xou

Load Xournal document content.

Load Xournal document content.

``image`` is (H, W, 3) uint8 RGB; the return has the same shape.

Saves the Handwritten Text Recognition (HTR) document for download. This function logs the interaction of saving an HTR document for downloa

Tests `XournalppDocument` and thereby its `load_data` function.

All HF-extra Python packages are importable inside the container.

The Docker image builds without error.

The Gradio HTTP server inside the container responds with HTTP 200.

The xournalpp binary is installed and on PATH inside the container.

Test that `get_document` does not crash when processing an empty Xournal++ file.

Tests `get_document` function for an unsupported file extension.

Tests `get_document` function for a `xopp` file.

Regression test of `Document.get_min_max_coordintes_per_page` function.

Test if code from `htr_pipeline` package can be imported.

Tests the `main` function using minimal test data. Note: This function is currently not executed in Github Actions due to the requirement of

Test `parse_arguments` with no command-line arguments. Ensures that `parse_arguments` raises a `SystemExit` exception when called without arg

Test the `parse_arguments` function with a full set of input arguments. This test verifies that the `parse_arguments` function correctly