Spaces:
Sleeping
Sleeping
- .Dockerignore +6 -0
- Dockerfile +12 -0
- api/app.py +10 -0
- api/middleware/auth.py +0 -0
- api/routes/__init__.py +2 -0
- api/routes/clusterRoute.py +16 -0
- api/routes/embeddingRoute.py +20 -0
- requirements.txt +73 -0
.Dockerignore
ADDED
|
@@ -0,0 +1,6 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# Ignore Python cache files in all directories
|
| 2 |
+
**/__pycache__/
|
| 3 |
+
*.pyc
|
| 4 |
+
|
| 5 |
+
# Ignore virtual environment directory
|
| 6 |
+
venv/
|
Dockerfile
ADDED
|
@@ -0,0 +1,12 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
FROM python:3.9
|
| 2 |
+
|
| 3 |
+
RUN useradd -m -u 1000 user
|
| 4 |
+
USER user
|
| 5 |
+
ENV PATH="/home/user/.local/bin:$PATH"
|
| 6 |
+
|
| 7 |
+
|
| 8 |
+
WORKDIR /server-slave
|
| 9 |
+
COPY --chown=user . /server-slave
|
| 10 |
+
RUN pip install -r /server-slave/requirements.txt
|
| 11 |
+
EXPOSE 7000
|
| 12 |
+
CMD ["python", "api/app.py"]
|
api/app.py
ADDED
|
@@ -0,0 +1,10 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from flask import Flask, request, jsonify
|
| 2 |
+
|
| 3 |
+
from gevent.pywsgi import WSGIServer
|
| 4 |
+
app = Flask(__name__)
|
| 5 |
+
|
| 6 |
+
from routes import *
|
| 7 |
+
|
| 8 |
+
if __name__ == "__main__":
|
| 9 |
+
http_server = WSGIServer(('', 7000), app)
|
| 10 |
+
http_server.serve_forever()
|
api/middleware/auth.py
ADDED
|
File without changes
|
api/routes/__init__.py
ADDED
|
@@ -0,0 +1,2 @@
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from .clusterRoute import *
|
| 2 |
+
from .embeddingRoute import *
|
api/routes/clusterRoute.py
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from app import app
|
| 2 |
+
from flask import request, jsonify
|
| 3 |
+
@app.route('/api/cluster', methods=['POST'])
|
| 4 |
+
def cluster():
|
| 5 |
+
data = request.get_json()
|
| 6 |
+
embeddings = data['embeddings']
|
| 7 |
+
from sklearn.cluster import DBSCAN
|
| 8 |
+
import numpy as np
|
| 9 |
+
try:
|
| 10 |
+
dbscan = DBSCAN(eps=0.4, min_samples=2, metric='cosine', n_jobs=-1)
|
| 11 |
+
embeddings_array = np.array(embeddings)
|
| 12 |
+
labels = dbscan.fit_predict(embeddings_array)
|
| 13 |
+
labels = labels.tolist()
|
| 14 |
+
return jsonify({'labels': labels})
|
| 15 |
+
except Exception as e:
|
| 16 |
+
return jsonify({'error': str(e)})
|
api/routes/embeddingRoute.py
ADDED
|
@@ -0,0 +1,20 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from sentence_transformers import SentenceTransformer
|
| 2 |
+
from flask import Flask, request, jsonify
|
| 3 |
+
import logging
|
| 4 |
+
from app import app
|
| 5 |
+
embedding_model = SentenceTransformer("dangvantuan/vietnamese-document-embedding",
|
| 6 |
+
trust_remote_code=True,
|
| 7 |
+
device='cpu')
|
| 8 |
+
|
| 9 |
+
@app.route('/api/embedding', methods=['POST'])
|
| 10 |
+
def embedding():
|
| 11 |
+
data = request.get_json()
|
| 12 |
+
logging.info(data)
|
| 13 |
+
try:
|
| 14 |
+
content = data['text']
|
| 15 |
+
embedding = embedding_model.encode(content)
|
| 16 |
+
embedding = embedding.tolist()
|
| 17 |
+
return jsonify({'embedding': embedding})
|
| 18 |
+
except Exception as e:
|
| 19 |
+
return jsonify({'error': str(e)})
|
| 20 |
+
|
requirements.txt
ADDED
|
@@ -0,0 +1,73 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
accelerate==1.2.1
|
| 2 |
+
aiohappyeyeballs==2.4.4
|
| 3 |
+
aiohttp==3.11.11
|
| 4 |
+
aiosignal==1.3.2
|
| 5 |
+
async-timeout==5.0.1
|
| 6 |
+
attrs==24.3.0
|
| 7 |
+
blinker==1.9.0
|
| 8 |
+
certifi==2024.12.14
|
| 9 |
+
cffi==1.17.1
|
| 10 |
+
charset-normalizer==3.4.1
|
| 11 |
+
click==8.1.8
|
| 12 |
+
colorama==0.4.6
|
| 13 |
+
coloredlogs==15.0.1
|
| 14 |
+
datasets==3.2.0
|
| 15 |
+
dill==0.3.8
|
| 16 |
+
evaluate==0.4.3
|
| 17 |
+
filelock==3.16.1
|
| 18 |
+
Flask==3.1.0
|
| 19 |
+
flatbuffers==24.12.23
|
| 20 |
+
frozenlist==1.5.0
|
| 21 |
+
fsspec==2024.9.0
|
| 22 |
+
gevent==24.11.1
|
| 23 |
+
greenlet==3.1.1
|
| 24 |
+
huggingface-hub==0.27.1
|
| 25 |
+
humanfriendly==10.0
|
| 26 |
+
idna==3.10
|
| 27 |
+
importlib_metadata==8.5.0
|
| 28 |
+
itsdangerous==2.2.0
|
| 29 |
+
Jinja2==3.1.5
|
| 30 |
+
joblib==1.4.2
|
| 31 |
+
MarkupSafe==3.0.2
|
| 32 |
+
mpmath==1.3.0
|
| 33 |
+
multidict==6.1.0
|
| 34 |
+
multiprocess==0.70.16
|
| 35 |
+
networkx==3.2.1
|
| 36 |
+
numpy==2.0.2
|
| 37 |
+
onnx==1.17.0
|
| 38 |
+
onnxruntime==1.19.2
|
| 39 |
+
optimum==1.23.3
|
| 40 |
+
packaging==24.2
|
| 41 |
+
pandas==2.2.3
|
| 42 |
+
pillow==11.1.0
|
| 43 |
+
propcache==0.2.1
|
| 44 |
+
protobuf==5.29.3
|
| 45 |
+
psutil==6.1.1
|
| 46 |
+
pyarrow==18.1.0
|
| 47 |
+
pycparser==2.22
|
| 48 |
+
pyreadline3==3.5.4
|
| 49 |
+
python-dateutil==2.9.0.post0
|
| 50 |
+
pytz==2024.2
|
| 51 |
+
PyYAML==6.0.2
|
| 52 |
+
regex==2024.11.6
|
| 53 |
+
requests==2.32.3
|
| 54 |
+
safetensors==0.5.2
|
| 55 |
+
scikit-learn==1.6.1
|
| 56 |
+
scipy==1.13.1
|
| 57 |
+
sentence-transformers==3.3.1
|
| 58 |
+
six==1.17.0
|
| 59 |
+
sympy==1.13.1
|
| 60 |
+
threadpoolctl==3.5.0
|
| 61 |
+
tokenizers==0.20.3
|
| 62 |
+
torch==2.5.1
|
| 63 |
+
tqdm==4.67.1
|
| 64 |
+
transformers==4.46.3
|
| 65 |
+
typing_extensions==4.12.2
|
| 66 |
+
tzdata==2024.2
|
| 67 |
+
urllib3==2.3.0
|
| 68 |
+
Werkzeug==3.1.3
|
| 69 |
+
xxhash==3.5.0
|
| 70 |
+
yarl==1.18.3
|
| 71 |
+
zipp==3.21.0
|
| 72 |
+
zope.event==5.0
|
| 73 |
+
zope.interface==7.2
|