Merge pull request #25 from yutanagano/develop

Prepare patch 1.0.1

Author: yutanagano

.github/workflows/tests.yaml

@@ -5,7 +5,7 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest, windows-latest, macos-latest]
-        python_ver: ['3.9', '3.10', '3.11', '3.12']
+        python_ver: ['3.12']
 
     runs-on: ${{ matrix.os }}
     steps:

README.md

@@ -12,7 +12,7 @@
 </div>
 
 **SCEPTR** (**S**imple **C**ontrastive **E**mbedding of the **P**rimary sequence of **T** cell **R**eceptors) is a small, fast, and accurate TCR representation model that can be used for alignment-free TCR  analysis, including for TCR-pMHC interaction prediction and TCR clustering (metaclonotype discovery).
-Our [manuscript (coming soon)](about:blank) demonstrates that SCEPTR can be used for few-shot TCR specificity prediction with improved accuracy over previous methods.
+Our [preprint](https://arxiv.org/abs/2406.06397) demonstrates that SCEPTR can be used for few-shot TCR specificity prediction with improved accuracy over previous methods.
 
 SCEPTR is a BERT-like transformer-based neural network implemented in [Pytorch](https://pytorch.org).
 With the default model providing best-in-class performance with only 153,108 parameters (typical protein language models have tens or hundreds of millions), SCEPTR runs fast- even on a CPU!

docs/citing.rst

@@ -1,4 +1,18 @@
 Citing sceptr
 =============
 
-Coming soon...
+To cite sceptr, please refer to our `preprint <https://arxiv.org/abs/2406.06397>`_.
+
+BibTex
+------
+
+.. code-block:: bibtex
+
+	@misc{nagano2024contrastive,
+	      title={Contrastive learning of T cell receptor representations}, 
+	      author={Yuta Nagano and Andrew Pyo and Martina Milighetti and James Henderson and John Shawe-Taylor and Benny Chain and Andreas Tiffeau-Mayer},
+	      year={2024},
+	      eprint={2406.06397},
+	      archivePrefix={arXiv},
+	      primaryClass={q-bio.BM}
+	}

docs/index.rst

@@ -2,7 +2,7 @@ SCEPTR
 ======
 
 **SCEPTR** (\ **S**\ imple **C**\ ontrastive **E**\ mbedding of the **P**\ rimary sequence of **T** cell **R**\ eceptors) is a small, fast, and performant TCR representation model that can be used for alignment-free downstream TCR and TCR repertoire analysis such as TCR clustering or classification.
-Our `manuscript (coming soon) <about:blank>`_ demonstrates SCEPTR's state-of-the-art performance (as of 2024) on downstream TCR specificity prediction.
+Our `preprint <https://arxiv.org/abs/2406.06397>`_ demonstrates SCEPTR's state-of-the-art performance (as of 2024) on downstream TCR specificity prediction.
 
 SCEPTR is a BERT-like transformer-based neural network implemented in `Pytorch <https://pytorch.org>`_.
 With the default model providing best-in-class performance with only 153,108 parameters (typical protein language models have tens or hundreds of millions), SCEPTR runs fast- even on a CPU!
@@ -18,7 +18,7 @@ What's even better is that they are fully compliant with `pyrepseq <https://pyre
    A visual introduction to how SCEPTR works, taken from our SCEPTR preprint.
    SCEPTR is a TCR language model (a,b) pre-trained using masked-language modelling and autocontrastive learning (c,d).
    (a) The default model uses the ``<cls>`` pooling method, but there is also a variant that is trained to use average-pooling (see :py:func:`sceptr.variant.average_pooling`).
-   Please see the manuscript for more details.
+   Please see the `preprint <https://arxiv.org/abs/2406.06397>`_  for more details.
 
 .. toctree::
    :maxdepth: 2

src/sceptr/__init__.py

@@ -10,7 +10,7 @@
 from pandas import DataFrame
 
 
-__version__ = "1.0.0"
+__version__ = "1.0.1"
 
 
 def calc_cdist_matrix(anchors: DataFrame, comparisons: DataFrame) -> ndarray:

tox.ini

@@ -4,9 +4,6 @@ env_list =
 
 [gh-actions]
 python =
-    3.9: py39
-    3.10: py310
-    3.11: py311
     3.12: py312
 
 [testenv]