Apply #803

.github/workflows/build.yml

@@ -47,7 +47,7 @@ jobs:
       - name: Build
         run: |
           pip install -U wheel
-          pip install .[all,dev] fasttext==0.9.2
+          pip install .[all,dev] fasttext==0.9.2 pillow==10.4.0
           python -c "import nltk; nltk.download(['punkt', 'punkt_tab', 'averaged_perceptron_tagger_eng'])"
           python --version
           make data coverage

setup.py

@@ -130,7 +130,7 @@ extras["all"] = (
 
 setup(
     name="txtai",
-    version="7.5.0",
+    version="7.5.1",
     author="NeuML",
     description="All-in-one open-source embeddings database for semantic search, LLM orchestration and language model workflows",
     long_description=DESCRIPTION,

src/python/txtai/pipeline/text/translation.py

@@ -2,6 +2,8 @@
 Translation module
 """
 
+import os
+
 # Conditional import
 try:
     import fasttext
@@ -10,7 +12,7 @@ try:
 except ImportError:
     FASTTEXT = False
 
-from huggingface_hub import cached_download
+from huggingface_hub import hf_hub_download
 from huggingface_hub.hf_api import HfApi
 from transformers import AutoModelForSeq2SeqLM, AutoTokenizer
 
@@ -24,7 +26,7 @@ class Translation(HFModel):
     """
 
     # Default language detection model
-    DEFAULT_LANG_DETECT = "https://dl.fbaipublicfiles.com/fasttext/supervised-models/lid.176.ftz"
+    DEFAULT_LANG_DETECT = "julien-c/fasttext-language-id/lid.176.ftz"
 
     def __init__(self, path=None, quantize=False, gpu=True, batch=64, langdetect=None, findmodels=True):
         """
@@ -155,7 +157,7 @@ class Translation(HFModel):
             path = self.langdetect if self.langdetect else Translation.DEFAULT_LANG_DETECT
 
             # Load language detection model
-            path = cached_download(path, legacy_cache_layout=True)
+            path = path if os.path.exists(path) else self.download(path)
             self.detector = fasttext.load_model(path)
 
         # Transform texts to format expected by language detection model
@@ -163,6 +165,26 @@ class Translation(HFModel):
 
         return [x[0].split("__")[-1] for x in self.detector.predict(texts)[0]]
 
+    def download(self, path):
+        """
+        Downloads path from the Hugging Face Hub.
+
+        Args:
+            path: full model path
+
+        Returns:
+            local cached model path
+        """
+
+        # Split into parts
+        parts = path.split("/")
+
+        # Calculate repo id split
+        repo = 2 if len(parts) > 2 else 1
+
+        # Download and cache file
+        return hf_hub_download(repo_id="/".join(parts[:repo]), filename="/".join(parts[repo:]))
+
     def translate(self, texts, source, target, showmodels=False):
         """
         Translates text from source to target language.

src/python/txtai/version.py

@@ -3,4 +3,4 @@ Version strings
 """
 
 # Current version tag
-__version__ = "7.5.0"
+__version__ = "7.5.1"