d30b22b50c
- FastAPI + Python 3.12 backend - Meta WhatsApp Business API client (official) - OpenAI GPT-4o with function calling - RAG vector store with pgvector - ERPNext Frappe REST client - Celery + Redis async task queue - PostgreSQL with migrations (Alembic) - Docker Compose full stack - Enterprise logging, metrics, health checks
101 lines
4.2 KiB
Python
101 lines
4.2 KiB
Python
"""Initial migration: conversations, messages, knowledge_chunks.
|
|
|
|
Revision ID: 001
|
|
Revises:
|
|
Create Date: 2026-04-28 00:00:00.000000
|
|
|
|
"""
|
|
from typing import Sequence, Union
|
|
|
|
from alembic import op
|
|
import sqlalchemy as sa
|
|
from sqlalchemy.dialects import postgresql
|
|
|
|
# revision identifiers, used by Alembic.
|
|
revision: str = "001"
|
|
down_revision: Union[str, None] = None
|
|
branch_labels: Union[str, Sequence[str], None] = None
|
|
depends_on: Union[str, Sequence[str], None] = None
|
|
|
|
|
|
def upgrade() -> None:
|
|
# Enable pgvector extension
|
|
op.execute("CREATE EXTENSION IF NOT EXISTS vector")
|
|
|
|
# Conversations table
|
|
op.create_table(
|
|
"conversations",
|
|
sa.Column("id", sa.String(36), primary_key=True),
|
|
sa.Column("phone_number", sa.String(20), nullable=False, index=True),
|
|
sa.Column("patient_id", sa.String(100), nullable=True, index=True),
|
|
sa.Column("patient_name", sa.String(255), nullable=True),
|
|
sa.Column(
|
|
"status",
|
|
sa.Enum("active", "paused", "resolved", "escalated", "appointment_confirmed", name="conversationstatus"),
|
|
nullable=False,
|
|
server_default="active",
|
|
),
|
|
sa.Column("context", postgresql.JSONB(astext_type=sa.Text()), server_default="{}"),
|
|
sa.Column("last_message_at", sa.DateTime(timezone=True), nullable=True),
|
|
sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now()),
|
|
sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now(), onupdate=sa.func.now()),
|
|
)
|
|
|
|
# Messages table
|
|
op.create_table(
|
|
"messages",
|
|
sa.Column("id", sa.String(36), primary_key=True),
|
|
sa.Column("conversation_id", sa.String(36), nullable=False, index=True),
|
|
sa.Column("direction", sa.String(10), nullable=False),
|
|
sa.Column("role", sa.String(20), nullable=False),
|
|
sa.Column("message_type", sa.String(50), server_default="text"),
|
|
sa.Column("content", sa.Text(), nullable=False),
|
|
sa.Column("whatsapp_message_id", sa.String(100), nullable=True),
|
|
sa.Column("tool_calls", postgresql.JSONB(astext_type=sa.Text()), nullable=True),
|
|
sa.Column("tool_results", postgresql.JSONB(astext_type=sa.Text()), nullable=True),
|
|
sa.Column("tokens_used", sa.Integer(), server_default="0"),
|
|
sa.Column("metadata", postgresql.JSONB(astext_type=sa.Text()), server_default="{}"),
|
|
sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now()),
|
|
)
|
|
|
|
# Knowledge chunks table (for RAG)
|
|
op.create_table(
|
|
"knowledge_chunks",
|
|
sa.Column("id", sa.String(36), primary_key=True, server_default=sa.text("gen_random_uuid()::text")),
|
|
sa.Column("content", sa.Text(), nullable=False),
|
|
sa.Column("metadata", postgresql.JSONB(astext_type=sa.Text()), server_default="{}"),
|
|
sa.Column("category", sa.String(50), server_default="general"),
|
|
sa.Column("source", sa.String(255), server_default=""),
|
|
sa.Column("embedding", sa.String(), nullable=True), # Stored as string; pgvector uses special type
|
|
sa.Column("created_at", sa.DateTime(timezone=True), server_default=sa.func.now()),
|
|
sa.Column("updated_at", sa.DateTime(timezone=True), server_default=sa.func.now()),
|
|
)
|
|
|
|
# Create pgvector column properly using raw SQL
|
|
op.execute("""
|
|
ALTER TABLE knowledge_chunks
|
|
ALTER COLUMN embedding TYPE vector(1536)
|
|
USING embedding::vector(1536)
|
|
""")
|
|
|
|
# Indexes
|
|
op.create_index("idx_knowledge_category", "knowledge_chunks", ["category"])
|
|
op.create_index("idx_knowledge_source", "knowledge_chunks", ["source"])
|
|
op.execute("""
|
|
CREATE INDEX idx_knowledge_embedding
|
|
ON knowledge_chunks
|
|
USING ivfflat (embedding vector_cosine_ops)
|
|
WITH (lists = 100)
|
|
""")
|
|
|
|
|
|
def downgrade() -> None:
|
|
op.drop_index("idx_knowledge_embedding", table_name="knowledge_chunks")
|
|
op.drop_index("idx_knowledge_source", table_name="knowledge_chunks")
|
|
op.drop_index("idx_knowledge_category", table_name="knowledge_chunks")
|
|
op.drop_table("knowledge_chunks")
|
|
op.drop_table("messages")
|
|
op.drop_table("conversations")
|
|
op.execute("DROP TYPE IF EXISTS conversationstatus")
|
|
op.execute("DROP EXTENSION IF EXISTS vector")
|