Spaces:
Sleeping
Sleeping
Upload 5 files
Browse files- .gitattributes +1 -0
- build_vector_store.py +47 -0
- chroma.sqlite3 +3 -0
- index_metadata.pickle +3 -0
- 高等数学上册.md +0 -0
- 高等数学下册.md +0 -0
.gitattributes
CHANGED
@@ -33,3 +33,4 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
36 |
+
chroma.sqlite3 filter=lfs diff=lfs merge=lfs -text
|
build_vector_store.py
ADDED
@@ -0,0 +1,47 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import os
|
2 |
+
from langchain.text_splitter import RecursiveCharacterTextSplitter
|
3 |
+
from langchain_community.embeddings import HuggingFaceEmbeddings
|
4 |
+
from langchain_community.vectorstores import Chroma
|
5 |
+
|
6 |
+
# 1. 设置路径
|
7 |
+
BASE_DIR = os.path.dirname(os.path.abspath(__file__)) # 当前脚本所在路径
|
8 |
+
PERSIST_DIR = os.path.abspath(os.path.join(BASE_DIR, "../vector_store")) # 向量库存储路径
|
9 |
+
SOURCE_DIR = BASE_DIR # 你的 .md 文件就在当前 vector_build/ 目录下
|
10 |
+
|
11 |
+
# 2. 加载 Embedding 模型
|
12 |
+
embed_model = HuggingFaceEmbeddings(
|
13 |
+
model_name="sentence-transformers/paraphrase-multilingual-mpnet-base-v2"
|
14 |
+
)
|
15 |
+
|
16 |
+
# 3. 加载 Markdown 文件 & 切分为小段
|
17 |
+
text_splitter = RecursiveCharacterTextSplitter(
|
18 |
+
chunk_size=500, chunk_overlap=50
|
19 |
+
)
|
20 |
+
|
21 |
+
docs = []
|
22 |
+
for fname in os.listdir(SOURCE_DIR):
|
23 |
+
if fname.endswith(".md"):
|
24 |
+
with open(os.path.join(SOURCE_DIR, fname), "r", encoding="utf-8") as f:
|
25 |
+
raw_text = f.read()
|
26 |
+
chunks = text_splitter.split_text(raw_text)
|
27 |
+
for chunk in chunks:
|
28 |
+
docs.append({
|
29 |
+
"text": chunk,
|
30 |
+
"source": fname
|
31 |
+
})
|
32 |
+
|
33 |
+
print(f"🐣 共切分出 {len(docs)} 个文本块,准备向量化...")
|
34 |
+
|
35 |
+
# 4. 创建 Chroma 向量库
|
36 |
+
texts = [d["text"] for d in docs]
|
37 |
+
metas = [{"source": d["source"]} for d in docs]
|
38 |
+
|
39 |
+
vectordb = Chroma.from_texts(
|
40 |
+
texts=texts,
|
41 |
+
embedding=embed_model,
|
42 |
+
metadatas=metas,
|
43 |
+
persist_directory=PERSIST_DIR
|
44 |
+
)
|
45 |
+
vectordb.persist()
|
46 |
+
|
47 |
+
print(f"🎉 向量库生成完毕,已保存在:{PERSIST_DIR}")
|
chroma.sqlite3
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:7f7a7a249bf3b1b4e7dd730cde0985cdf0220c849e5deeca31d4df7c912720f2
|
3 |
+
size 22417408
|
index_metadata.pickle
ADDED
@@ -0,0 +1,3 @@
|
|
|
|
|
|
|
|
|
1 |
+
version https://git-lfs.github.com/spec/v1
|
2 |
+
oid sha256:5d2130d448d160bb76ed65984a84859d3d6b645d4a3bf09971b32e2eb4defe63
|
3 |
+
size 213668
|
高等数学上册.md
ADDED
The diff for this file is too large to render.
See raw diff
|
|
高等数学下册.md
ADDED
The diff for this file is too large to render.
See raw diff
|
|