Add Python 3.14 support, remove Python 3.10 support, update Poetry to 2.4.1

.devcontainer/devcontainer.json

@@ -42,6 +42,7 @@
 				"editor.formatOnSave": true,
 				"editor.formatOnType": true,
 				"isort.args": [
+					"--gitignore",
 					"--profile",
 					"black"
 				]

.devcontainer/dockerfile

@@ -1,7 +1,5 @@
-#compatibility with Tensorflow 2.6.0 as per https://www.tensorflow.org/install/source#gpu
-ARG PYTHON_VERSION=3.10
-ARG UBUNTU_VERSION=noble
-ARG POETRY_VERSION=1.8.2
+ARG PYTHON_VERSION=3.11
+ARG POETRY_VERSION=2.4.1
 ARG CUDA_VERSION=12.6.1-base-ubuntu24.04
 
 FROM nvidia/cuda:$CUDA_VERSION
@@ -19,7 +17,6 @@ RUN apt-get update && \
     apt-get update && \
     apt-get install --no-install-recommends -y \
     python$PYTHON_VERSION \
-    python$PYTHON_VERSION-distutils \
     python$PYTHON_VERSION-dev \
     git vim curl gdb ca-certificates gnupg2 tar make gcc libssl-dev zlib1g-dev libncurses5-dev \
     libbz2-dev libreadline-dev libreadline6-dev libxml2-dev xz-utils libgdbm-dev libgdbm-compat-dev tk-dev dirmngr \

.github/workflows/ci.yml

@@ -14,13 +14,13 @@ jobs:
     strategy:
       matrix:
         os: [ubuntu-latest, macos-latest, windows-latest]
-        python-version: ["3.10", "3.11", "3.12", "3.13"]
+        python-version: ["3.11", "3.12", "3.13", "3.14"]
     defaults:
       run:
         shell: bash
 
     steps:
-      - uses: actions/checkout@v6
+      - uses: actions/checkout@v7
       - name: Set up Python ${{ matrix.python-version }}
         id: setup-python
         uses: actions/setup-python@v6
@@ -29,6 +29,7 @@ jobs:
       - name: Install Poetry
         uses: snok/install-poetry@v1
         with:
+          version: 2.4.1
           virtualenvs-create: true
           virtualenvs-in-project: true
           installer-parallel: true
@@ -48,14 +49,14 @@ jobs:
           node-version: "24"
       - name: Lint with pyright
         run: |
-          npm install -g pyright@1.1.400
+          npm install -g pyright@1.1.411
           poetry run pyright
       - name: Test with pytest
         run: poetry run pytest --cov --cov-report=xml
         env:
           EFLOMAL_PATH: /home/runner/work/machine.py/machine.py/.venv/lib/python${{ matrix.python-version }}/site-packages/eflomal/bin
       - name: Upload coverage reports to Codecov
-        uses: codecov/codecov-action@v5
+        uses: codecov/codecov-action@v7
         env:
           CODECOV_TOKEN: ${{ secrets.CODECOV_TOKEN }}
 
@@ -65,22 +66,23 @@ jobs:
     runs-on: ubuntu-latest
 
     steps:
-      - uses: actions/checkout@v6
-      - name: Set up Python 3.10
+      - uses: actions/checkout@v7
+      - name: Set up Python 3.11
         id: setup-python
         uses: actions/setup-python@v6
         with:
-          python-version: "3.10"
+          python-version: "3.11"
       - name: Install Poetry
         uses: snok/install-poetry@v1
         with:
+          version: 2.4.1
           virtualenvs-create: true
           virtualenvs-in-project: true
           installer-parallel: true
       - name: Build
         run: poetry build
       - name: Upload package
-        uses: actions/upload-artifact@v6
+        uses: actions/upload-artifact@v7
         with:
           name: wheel
           path: dist/*.whl

.github/workflows/docker-build-push.yml

@@ -26,12 +26,12 @@ jobs:
         with:
           docker-images: false
       - name: Checkout
-        uses: actions/checkout@v6
+        uses: actions/checkout@v7
       - name: Set up QEMU
         uses: docker/setup-qemu-action@v4
       - name: Generate Docker metadata
         id: meta
-        uses: docker/metadata-action@v5
+        uses: docker/metadata-action@v6
         with:
           images: ghcr.io/sillsdev/machine.py
           tags: |

.gitignore

@@ -139,3 +139,6 @@ out/
 
 # omit IDE data
 .idea/
+
+# Ignore Poetry plugins
+.poetry/

dockerfile

@@ -1,7 +1,6 @@
-# syntax=docker/dockerfile:1.7-labs
+# syntax=docker/dockerfile:1
 ARG PYTHON_VERSION=3.12
-ARG UBUNTU_VERSION=noble
-ARG POETRY_VERSION=1.6.1
+ARG POETRY_VERSION=2.4.1
 
 FROM python:$PYTHON_VERSION-slim-bookworm AS builder
 ARG POETRY_VERSION
@@ -19,7 +18,8 @@ RUN python3 -m venv $POETRY_VENV \
 ENV PATH="${PATH}:${POETRY_VENV}/bin"
 
 WORKDIR /src
-COPY poetry.lock pyproject.toml /src
+COPY poetry.lock pyproject.toml /src/
+RUN poetry self add poetry-plugin-export
 RUN poetry export --with=gpu --without-hashes -f requirements.txt > requirements.txt
 
 

dockerfile.cpu_only

@@ -1,7 +1,6 @@
-# syntax=docker/dockerfile:1.7-labs
+# syntax=docker/dockerfile:1
 ARG PYTHON_VERSION=3.12
-ARG UBUNTU_VERSION=noble
-ARG POETRY_VERSION=1.6.1
+ARG POETRY_VERSION=2.4.1
 
 FROM python:$PYTHON_VERSION-slim-bookworm AS builder
 ARG POETRY_VERSION
@@ -19,7 +18,8 @@ RUN python3 -m venv $POETRY_VENV \
 ENV PATH="${PATH}:${POETRY_VENV}/bin"
 
 WORKDIR /src
-COPY poetry.lock pyproject.toml /src
+COPY poetry.lock pyproject.toml /src/
+RUN poetry self add poetry-plugin-export
 RUN poetry export --extras="thot jobs" --without-hashes -f requirements.txt > requirements.txt
 
 

machine/corpora/n_parallel_text_corpus.py

@@ -402,9 +402,9 @@ def default_row_ref_comparer(x: Any, y: Any) -> int:
     # Do not use the default comparer for ScriptureRef, since we want to ignore segments
     if isinstance(x, ScriptureRef) and isinstance(y, ScriptureRef):
         return x.compare_to(y, False)
-    if x is None and y is not None:
-        return 1
-    if x is not None and y is None:
+    if x is None:
+        return 0 if y is None else 1
+    if y is None:
         return -1
     if x == y:
         return 0

machine/corpora/parallel_text_corpus.py

@@ -415,9 +415,11 @@ def to_hf_dataset(
         if text_id_column is not None:
             features_dict[text_id_column] = Value("string")
         if ref_column is not None:
-            features_dict[ref_column] = Sequence(Value("string"))
+            features_dict[ref_column] = cast(FeatureType, Sequence(Value("string")))
         if alignment_column is not None:
-            features_dict[alignment_column] = Sequence({source_lang: Value("int32"), target_lang: Value("int32")})
+            features_dict[alignment_column] = cast(
+                FeatureType, Sequence({source_lang: Value("int32"), target_lang: Value("int32")})
+            )
         if content_type_column is not None:
             features_dict[content_type_column] = ClassLabel(names=[e.name for e in TextRowContentType])
         features = Features(features_dict)
@@ -468,9 +470,11 @@ def to_hf_iterable_dataset(
         if text_id_column is not None:
             features_dict[text_id_column] = Value("string")
         if ref_column is not None:
-            features_dict[ref_column] = Sequence(Value("string"))
+            features_dict[ref_column] = cast(FeatureType, Sequence(Value("string")))
         if alignment_column is not None:
-            features_dict[alignment_column] = Sequence({source_lang: Value("int32"), target_lang: Value("int32")})
+            features_dict[alignment_column] = cast(
+                FeatureType, Sequence({source_lang: Value("int32"), target_lang: Value("int32")})
+            )
         features = Features(features_dict)
 
         if info is not None and info.features is not None and info.features != features:

machine/translation/huggingface/hugging_face_nmt_engine.py

@@ -74,7 +74,7 @@ def __init__(
         ):
             self._pipeline_kwargs["prefix"] = f"translate {src_lang} to {tgt_lang}: "
         else:
-            additional_special_tokens = self._tokenizer.additional_special_tokens
+            additional_special_tokens = cast(list[str], self._tokenizer.additional_special_tokens or [])
             if isinstance(self._tokenizer, M2M100Tokenizer):
                 src_lang_token = self._tokenizer.lang_code_to_token.get(src_lang) if src_lang is not None else None
                 tgt_lang_token = self._tokenizer.lang_code_to_token.get(tgt_lang) if tgt_lang is not None else None

machine/translation/huggingface/hugging_face_nmt_model_trainer.py

@@ -274,9 +274,10 @@ def add_tokens(tokenizer: Any, missing_tokens: List[str]) -> Any:
             # For multilingual translation models like mBART-50 and M2M100 we need to force the target language token
             # as the first generated token. We ask the user to explicitly provide this as --forced_bos_token argument.
             forced_bos_token_id = tokenizer.convert_tokens_to_ids(self._tgt_lang)
-            model.config.forced_bos_token_id = forced_bos_token_id
             if model.generation_config is not None:
                 model.generation_config.forced_bos_token_id = forced_bos_token_id
+            else:
+                model.config.forced_bos_token_id = forced_bos_token_id
 
         prefix = ""
         if model.name_or_path.startswith("t5-") or model.name_or_path.startswith("google/mt5-"):