Spaces:
Sleeping
Sleeping
project-samarth
Browse files- database/schema.py +70 -0
database/schema.py
ADDED
|
@@ -0,0 +1,70 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
from sqlalchemy import create_engine, Column, Integer, String, Text, DateTime, Float, JSON, Boolean
|
| 2 |
+
from sqlalchemy.ext.declarative import declarative_base
|
| 3 |
+
from sqlalchemy.orm import sessionmaker
|
| 4 |
+
from datetime import datetime
|
| 5 |
+
import os
|
| 6 |
+
|
| 7 |
+
Base = declarative_base()
|
| 8 |
+
|
| 9 |
+
class ChatSession(Base):
|
| 10 |
+
__tablename__ = 'chat_sessions'
|
| 11 |
+
|
| 12 |
+
id = Column(Integer, primary_key=True)
|
| 13 |
+
session_id = Column(String(100), unique=True, nullable=False, index=True)
|
| 14 |
+
created_at = Column(DateTime, default=datetime.utcnow)
|
| 15 |
+
last_active = Column(DateTime, default=datetime.utcnow, onupdate=datetime.utcnow)
|
| 16 |
+
|
| 17 |
+
class ChatMessage(Base):
|
| 18 |
+
__tablename__ = 'chat_messages'
|
| 19 |
+
|
| 20 |
+
id = Column(Integer, primary_key=True)
|
| 21 |
+
session_id = Column(String(100), nullable=False, index=True)
|
| 22 |
+
role = Column(String(20), nullable=False)
|
| 23 |
+
content = Column(Text, nullable=False)
|
| 24 |
+
sources = Column(JSON)
|
| 25 |
+
timestamp = Column(DateTime, default=datetime.utcnow)
|
| 26 |
+
|
| 27 |
+
class DatasetMetadata(Base):
|
| 28 |
+
__tablename__ = 'dataset_metadata'
|
| 29 |
+
|
| 30 |
+
id = Column(Integer, primary_key=True)
|
| 31 |
+
dataset_id = Column(String(100), unique=True, nullable=False, index=True)
|
| 32 |
+
name = Column(String(500))
|
| 33 |
+
category = Column(String(100))
|
| 34 |
+
description = Column(Text)
|
| 35 |
+
endpoint = Column(String(500))
|
| 36 |
+
fields = Column(JSON)
|
| 37 |
+
record_count = Column(Integer)
|
| 38 |
+
last_fetched = Column(DateTime)
|
| 39 |
+
is_active = Column(Boolean, default=True)
|
| 40 |
+
|
| 41 |
+
class EmbeddingMetadata(Base):
|
| 42 |
+
__tablename__ = 'embedding_metadata'
|
| 43 |
+
|
| 44 |
+
id = Column(Integer, primary_key=True)
|
| 45 |
+
dataset_id = Column(String(100), nullable=False, index=True)
|
| 46 |
+
chunk_id = Column(String(100), nullable=False, index=True)
|
| 47 |
+
content = Column(Text, nullable=False)
|
| 48 |
+
meta_info = Column(JSON)
|
| 49 |
+
embedding_model = Column(String(100))
|
| 50 |
+
created_at = Column(DateTime, default=datetime.utcnow)
|
| 51 |
+
|
| 52 |
+
def init_db():
|
| 53 |
+
database_url = os.getenv('DATABASE_URL')
|
| 54 |
+
if not database_url:
|
| 55 |
+
raise ValueError("DATABASE_URL environment variable not set")
|
| 56 |
+
|
| 57 |
+
engine = create_engine(database_url)
|
| 58 |
+
Base.metadata.create_all(engine)
|
| 59 |
+
|
| 60 |
+
Session = sessionmaker(bind=engine)
|
| 61 |
+
return Session()
|
| 62 |
+
|
| 63 |
+
def get_db_session():
|
| 64 |
+
database_url = os.getenv('DATABASE_URL')
|
| 65 |
+
if not database_url:
|
| 66 |
+
raise ValueError("DATABASE_URL environment variable not set")
|
| 67 |
+
|
| 68 |
+
engine = create_engine(database_url)
|
| 69 |
+
Session = sessionmaker(bind=engine)
|
| 70 |
+
return Session()
|