# Why is the change needed?
The GATE identifiers column was meant to be named `GATE-ID`, but the current header reports `ID`.
# What changes does the patch introduce?
We change the wrong column header into `GATE-ID`.
# How was this patch tested?
No test required.
# Why is the change needed?
As the reference paper for Neo-GATE is now part of the ACL Anthology, we want to update old references pointing to the arXiv pre-print.
# What changes does the patch introduce?
We update the bib citation and the old links to the paper.
As part of the maintenance, we also add one missing link to the curator's profile on Hugging Face.
# How was this patch tested?
No test required.
# Why is the change needed?
There is a typo in the `schwa_simple` configuration, which breaks Hugging Face's correct loading of it.
# What changes does the patch introduce?
We fix the typo in the path to `adapted/Neo-GATE_dev_schwa-simple.tsv`.
# How was this patch tested?
No test required.
# Why is the change needed?
We want to provide a dataset configuration which includes pre-adapted annotations and references.
We need this configuration for our CALAMITA 2024 submission.
# What changes does the patch introduce?
With this patch we introduce several files:
* `adapted/Neo-GATE_schwa-simple.tsv`: the `test` split of the adapted dataset.
* `adapted/Neo-GATE_dev_schwa-simple.tsv`: the `dev` split of the adapted dataset.
* `schwa-simple.json`: the tagset mapping used to create this adapted version.
We also make changes to existing files:
* `README.md` now includes the new configuration in the metadata and a brief description of the two available configurations.
* `schwa.json` was renamed into `schwa-complex.json`, and `README.md` was updated accordingly.
# How was this patch tested?
No test required.
# Why is the change needed?
Currently, the script that allows to adapt Neo-GATE to the given neomorpheme paradigm is named `neo-gate_adapt.py`.
The dash (or hyphen) in the name makes importing code from the script more complex than it needs to be, due to how Python handles dashes.
While there are ways to import scripts that include dashes in their name, removing the dash would make it easier to import code from the script and make the naming more coherent with the evaluation code published in [fbk-NEUTR-evAL](https://github.com/hlt-mt/fbk-NEUTR-evAL/blob/main/src/fbk_neutreval/cli/neogate_evaluation.py).
# What changes does the patch introduce?
This changes renames `neo-gate_adapt.py` into `neogate_adapt.py`.
It also updates the repository's README file to reflect this change and corrects a typo.
# How was this patch tested?
No test required.
# Why is the change needed?
As the evaluation script for Neo-GATE is now available on GitHub, we must provide a link to it so that users can actually use the corpus to evaluate.
The EAMT proceedings are also out, so the pre-print citation is now obsolete.
# What changes does the patch introduce?
It adds a link to the evaluation script reference document in GitHub and replaces the pre-print citation with the EAMT one.
# How was this patch tested?
No test required.
# Why is the change needed?
The guidelines for Neo-GATE's annotation were intended to be released alongside the corpus, we are including them with this change.
# What changes does the patch introduce?
The patch adds the `annotation_guidelines.pdf` file to the repository.
# How was this patch tested?
No test required.
Re-added the 'asterisk' neomorpheme paradigm tagset mapping for reproducibility
# Why is the change needed?
**This change was originally included in a different MR, but it was erroneously discarded.**
The original paper used two neomorpheme paradigms: 'schwa' and 'asterisk'.
Currently, the repository only includes a tagset mapping for the 'schwa' neomorpheme paradigm.
This change introduces `asterisk.json`, i.e. the tagset mapping for the 'asterisk' paradigm in JSON format, for reproducibility purposes.
With this addition, we make both versions of Neo-GATE used in the original paper reproducible.
`asterisk.json` also serves as a further exemplar tagset mapping.
# What changes does the patch introduce?
The patch adds the `asterisk.json` file to the repository.
# How was this patch tested?
No tests are required.
# Why is the change needed?
HuggingFace requires to have different splits of the corpus stored in different files.
Thus, we are splitting Neo-GATE in two files, one for the test split and one for the dev split.
# What changes does the patch introduce?
With this change, rather than having the whole corpus in a single TSV, we split it into two different files, one for the test split (`Neo-GATE.tsv`) and one for the dev split (`Neo-GATE-dev.tsv`).
The SPLIT column is also removed from the TSV since it is no longer needed to identify entries belonging to each split.
# How was this patch tested?
No tests are required.
# Why is the change needed?
This change is required to include proper licensing information about both Neo-GATE and GATE, as the latter is licensed under the MIT license, which requires derived work to include a copy of the license itself.
# What changes does the patch introduce?
This patch introduce the `LICENSE` file, which contains all licensing information required for Neo-GATE.
# How was this patch tested?
No tests are required.
# Why is the change needed?
This changes introduces a revised and updated README file.
# What changes does the patch introduce?
The new version of the `README.md` includes tags and metadata required by HuggingFace to properly visualize the corpus, as well as corrections and improvements.
# How was this patch tested?
No tests are required.
Added the 'asterisk' neomorpheme paradigm tagset mapping for reproducibility
# Why is the change needed?
The original paper used two neomorpheme paradigms: 'schwa' and 'asterisk'.
Currently, the repository only includes a tagset mapping for the 'schwa' neomorpheme paradigm.
This change introduces `asterisk.json`, i.e. the tagset mapping for the 'asterisk' paradigm in JSON format, for reproducibility purposes.
With this addition, we make both versions of Neo-GATE used in the original paper reproducible.
`asterisk.json` also serves as a further exemplar tagset mapping.
# What changes does the patch introduce?
The patch adds the `asterisk.json` file to the repository.
# How was this patch tested?
No tests are required.
# Why is the change needed?
This change introduces an exemplary tagset mapping for Neo-GATE in JSON format.
The file will be referenced in `README.md` to provide users with a starting point to create custom neomorpheme paradigm mappings to Neo-GATE's tagset.
# What changes does the patch introduce?
The patch adds the `schwa.json` file to the repository.
# How was this patch tested?
No tests are required.
# Why is the change needed?
This change introduces the adaptation script for Neo-GATE.
# What changes does the patch introduce?
The patch adds the script `neo-gate_adapt.py`, which formats annotations and references with the forms found in the provided tagset mapping. A tagset mapping is a dictionary which maps tags from Neo-GATE's tagset to the desired forms to be evaluated.
# How was this patch tested?
Manual runs.
