From bd3faf81339a800f878db931873419cdfa29e117 Mon Sep 17 00:00:00 2001
From: Fam Zheng <fam@euphon.net>
Date: Tue, 4 Nov 2025 23:06:33 +0000
Subject: [PATCH] add hf models in docker

---
 Dockerfile                  |  2 ++
 api/products/rag_service.py | 17 +++++++++++++++--
 requirements.txt            |  1 +
 scripts/download_models.py  | 24 ++++++++++++++++++++++++
 4 files changed, 42 insertions(+), 2 deletions(-)
 create mode 100755 scripts/download_models.py

diff --git a/Dockerfile b/Dockerfile
index 593944c..976ead4 100644
--- a/Dockerfile
+++ b/Dockerfile
@@ -18,6 +18,8 @@ ADD requirements.txt requirements.txt
 RUN curl -LsSf https://astral.sh/uv/install.sh | sh && \
     rm -f /usr/lib/python*/EXTERNALLY-MANAGED && \
     /root/.local/bin/uv pip install --system --no-cache -r requirements.txt
+ADD scripts/download_models.py /tmp/download_models.py
+RUN python3 /tmp/download_models.py && rm /tmp/download_models.py
 ADD api /emblem/api
 ADD web /emblem/web
 RUN cd /emblem/api && ./manage.py collectstatic --noinput
diff --git a/api/products/rag_service.py b/api/products/rag_service.py
index a59534e..cb6199c 100644
--- a/api/products/rag_service.py
+++ b/api/products/rag_service.py
@@ -5,7 +5,7 @@ RAG Service Module for Themblem
 """
 
 from langchain_text_splitters import RecursiveCharacterTextSplitter
-from langchain_community.embeddings import HuggingFaceEmbeddings
+from langchain_huggingface import HuggingFaceEmbeddings
 from langchain_community.vectorstores import FAISS
 from langchain_core.documents import Document
 from .models import Article
@@ -20,8 +20,21 @@ class CachedLangChainRAG:
     """基于缓存的实时LangChain RAG服务"""
     
     def __init__(self):
+        # Use pre-downloaded model from container (relative to this file)
+        import os
+        from pathlib import Path
+        
+        # Model path relative to this file: api/products/rag_service.py -> ../../models/
+        model_path = Path(__file__).parent.parent.parent / "models" / "text2vec-base-chinese"
+        
+        # Fall back to downloading if local model doesn't exist (for development)
+        if not model_path.exists():
+            model_path = "shibing624/text2vec-base-chinese"
+        else:
+            model_path = str(model_path)
+        
         self.embeddings = HuggingFaceEmbeddings(
-            model_name="shibing624/text2vec-base-chinese",
+            model_name=model_path,
             model_kwargs={'device': 'cpu'},
             encode_kwargs={'normalize_embeddings': True}
         )
diff --git a/requirements.txt b/requirements.txt
index e1dec28..4cdb36b 100644
--- a/requirements.txt
+++ b/requirements.txt
@@ -14,6 +14,7 @@ langchain>=0.1.0
 langchain-community>=0.1.0
 langchain-core>=0.1.0
 langchain-text-splitters>=0.1.0
+langchain-huggingface>=0.1.0
 sentence-transformers>=2.2.0
 faiss-cpu>=1.7.4
 beautifulsoup4>=4.12.0
diff --git a/scripts/download_models.py b/scripts/download_models.py
new file mode 100755
index 0000000..c7663f6
--- /dev/null
+++ b/scripts/download_models.py
@@ -0,0 +1,24 @@
+#!/usr/bin/env python3
+"""
+Download HuggingFace models for offline use
+"""
+from sentence_transformers import SentenceTransformer
+from pathlib import Path
+
+# Create models directory relative to this script: scripts/ -> ../models/
+script_dir = Path(__file__).parent
+models_dir = script_dir.parent / "models"
+models_dir.mkdir(parents=True, exist_ok=True)
+
+# Download the model
+model_name = "shibing624/text2vec-base-chinese"
+model_path = models_dir / "text2vec-base-chinese"
+
+print(f"Downloading model: {model_name}")
+print(f"Saving to: {model_path}")
+
+model = SentenceTransformer(model_name)
+model.save(str(model_path))
+
+print(f"Model successfully downloaded to {model_path}")
+