- Improvements to enable deployment in the cloud, mainly changing file access to Minio
- Improvements on RAG logging, and some debugging in that area
This commit is contained in:
86
common/utils/minio_utils.py
Normal file
86
common/utils/minio_utils.py
Normal file
@@ -0,0 +1,86 @@
|
||||
from minio import Minio
|
||||
from minio.error import S3Error
|
||||
from flask import Flask
|
||||
import io
|
||||
from werkzeug.datastructures import FileStorage
|
||||
|
||||
class MinioClient:
|
||||
def __init__(self):
|
||||
self.client = None
|
||||
|
||||
def init_app(self, app: Flask):
|
||||
self.client = Minio(
|
||||
app.config['MINIO_ENDPOINT'],
|
||||
access_key=app.config['MINIO_ACCESS_KEY'],
|
||||
secret_key=app.config['MINIO_SECRET_KEY'],
|
||||
secure=app.config.get('MINIO_USE_HTTPS', False)
|
||||
)
|
||||
app.logger.info(f"MinIO client initialized with endpoint: {app.config['MINIO_ENDPOINT']}")
|
||||
|
||||
def generate_bucket_name(self, tenant_id):
|
||||
return f"tenant-{tenant_id}-bucket"
|
||||
|
||||
def create_tenant_bucket(self, tenant_id):
|
||||
bucket_name = self.generate_bucket_name(tenant_id)
|
||||
try:
|
||||
if not self.client.bucket_exists(bucket_name):
|
||||
self.client.make_bucket(bucket_name)
|
||||
return bucket_name
|
||||
return bucket_name
|
||||
except S3Error as err:
|
||||
raise Exception(f"Error occurred while creating bucket: {err}")
|
||||
|
||||
def generate_object_name(self, document_id, language, version_id, filename):
|
||||
return f"{document_id}/{language}/{version_id}/{filename}"
|
||||
|
||||
def upload_document_file(self, tenant_id, document_id, language, version_id, filename, file_data):
|
||||
bucket_name = self.generate_bucket_name(tenant_id)
|
||||
object_name = self.generate_object_name(document_id, language, version_id, filename)
|
||||
|
||||
try:
|
||||
if isinstance(file_data, FileStorage):
|
||||
file_data = file_data.read()
|
||||
elif isinstance(file_data, io.BytesIO):
|
||||
file_data = file_data.getvalue()
|
||||
elif isinstance(file_data, str):
|
||||
file_data = file_data.encode('utf-8')
|
||||
elif not isinstance(file_data, bytes):
|
||||
raise TypeError('Unsupported file type. Expected FileStorage, BytesIO, str, or bytes.')
|
||||
|
||||
self.client.put_object(
|
||||
bucket_name, object_name, io.BytesIO(file_data), len(file_data)
|
||||
)
|
||||
return True
|
||||
except S3Error as err:
|
||||
raise Exception(f"Error occurred while uploading file: {err}")
|
||||
|
||||
def download_document_file(self, tenant_id, document_id, language, version_id, filename):
|
||||
bucket_name = self.generate_bucket_name(tenant_id)
|
||||
object_name = self.generate_object_name(document_id, language, version_id, filename)
|
||||
try:
|
||||
response = self.client.get_object(bucket_name, object_name)
|
||||
return response.read()
|
||||
except S3Error as err:
|
||||
raise Exception(f"Error occurred while downloading file: {err}")
|
||||
|
||||
def list_document_files(self, tenant_id, document_id, language=None, version_id=None):
|
||||
bucket_name = self.generate_bucket_name(tenant_id)
|
||||
prefix = f"{document_id}/"
|
||||
if language:
|
||||
prefix += f"{language}/"
|
||||
if version_id:
|
||||
prefix += f"{version_id}/"
|
||||
try:
|
||||
objects = self.client.list_objects(bucket_name, prefix=prefix, recursive=True)
|
||||
return [obj.object_name for obj in objects]
|
||||
except S3Error as err:
|
||||
raise Exception(f"Error occurred while listing files: {err}")
|
||||
|
||||
def delete_document_file(self, tenant_id, document_id, language, version_id, filename):
|
||||
bucket_name = self.generate_bucket_name(tenant_id)
|
||||
object_name = self.generate_object_name(document_id, language, version_id, filename)
|
||||
try:
|
||||
self.client.remove_object(bucket_name, object_name)
|
||||
return True
|
||||
except S3Error as err:
|
||||
raise Exception(f"Error occurred while deleting file: {err}")
|
||||
@@ -141,7 +141,7 @@ def select_model_variables(tenant):
|
||||
default_headers=portkey_headers)
|
||||
tool_calling_supported = False
|
||||
match llm_model:
|
||||
case 'gpt-4-turbo' | 'gpt-4o':
|
||||
case 'gpt-4-turbo' | 'gpt-4o' | 'gpt-4o-mini':
|
||||
tool_calling_supported = True
|
||||
case _:
|
||||
raise Exception(f'Error setting model variables for tenant {tenant.id} '
|
||||
|
||||
Reference in New Issue
Block a user