Merge pull request #433 from sillsdev/upgrade_eflomal

Upgrade eflomal

Author: mshannon-sil

Dockerfile

@@ -40,6 +40,7 @@ RUN apt-get install --no-install-recommends -y \
     git \
     python$PYTHON_VERSION \
     python3-pip \
+    python3-dev \
     wget \
     build-essential \
     gdb \
@@ -59,12 +60,8 @@ RUN ln -sfn /usr/bin/python${PYTHON_VERSION} /usr/bin/python3  & \
 COPY --from=builder /src/requirements.txt .
 RUN pip install --no-cache-dir -r requirements.txt && rm requirements.txt
 
-# Install eflomal
-RUN git clone https://github.com/robertostling/eflomal.git
-RUN make -C eflomal/src
-RUN make -C eflomal/src install
-RUN rm -rf eflomal
-ENV EFLOMAL_PATH=/usr/local/bin
+# Set eflomal path
+ENV EFLOMAL_PATH=/usr/local/lib/python3.8/dist-packages/eflomal/bin
 
 # Install fast_align
 RUN apt-get update && \

poetry.lock

@@ -81,7 +81,7 @@ sacremoses = "^0.0.53"
 evaluate = "^0.3.0"
 python-docx = "^0.8.11"
 iso639-lang = "^2.1.0"
-eflomal = { version = "^0.1", optional = true }
+eflomal = { version = "^2.0.0", optional = true }
 accelerate = "^0.23.0"
 transformers = "^4.36.2"
 optimum = "^1.16.0"

pyproject.toml

@@ -8,7 +8,6 @@
 from ..common.environment import SIL_NLP_ENV
 from ..common.flatcat_stemmer import FlatCatStemmer
 from ..common.null_stemmer import NullStemmer
-from ..common.packages_utils import is_eflomal_available
 from ..common.snowball_stemmer import SnowballStemmer
 from ..common.stemmer import Stemmer
 from ..common.utils import merge_dict
@@ -24,10 +23,7 @@
     Ibm4DotnetMachineAligner,
     ParatextDotnetMachineAligner,
 )
-
-if is_eflomal_available():
-    from .eflomal import EflomalAligner
-
+from .eflomal import EflomalAligner
 from .fast_align import FastAlign
 from .giza_aligner import HmmGizaAligner, Ibm1GizaAligner, Ibm2GizaAligner, Ibm3GizaAligner, Ibm4GizaAligner
 from .machine_aligner import (
@@ -67,11 +63,9 @@
     "clear3_fa": (ClearAligner, "Clear-3-FA"),
     "clear3_hmm": (ClearAligner, "Clear-3-HMM"),
     "clab_fast_align": (FastAlign, "clab-FastAlign"),
+    "eflomal": (EflomalAligner, "Eflomal"),
 }
 
-if is_eflomal_available():
-    ALIGNERS["eflomal"] = (EflomalAligner, "Eflomal")
-
 STEMMERS: Dict[str, Type[Stemmer]] = {
     "snowball": SnowballStemmer,
     "wordnet": WordNetStemmer,

silnlp/alignment/config.py

@@ -6,20 +6,26 @@
 from tempfile import TemporaryDirectory
 from typing import IO, Iterable
 
-from eflomal import read_text, write_text
 from machine.corpora import AlignedWordPair
 from machine.translation import SymmetrizationHeuristic, WordAlignmentMatrix
 
 from ..common.corpus import load_corpus
+from ..common.packages_utils import is_eflomal_available
 from .aligner import Aligner
 from .lexicon import Lexicon
 from .tools import execute_atools, execute_eflomal, is_atools_available
 
+if is_eflomal_available():
+    from eflomal import read_text, write_text
+
 LOGGER = logging.getLogger(__name__)
 
 
 class EflomalAligner(Aligner):
     def __init__(self, model_dir: Path) -> None:
+        if not is_eflomal_available():
+            raise RuntimeError("eflomal is not installed.")
+
         super().__init__("eflomal", model_dir)
 
     def train(self, src_file_path: Path, trg_file_path: Path) -> None: