refactor: delegate memory models and table creation to memu-py by SparkZou · Pull Request #19 · NevaMind-AI/memU-server

SparkZou · 2026-02-09T04:45:10Z

Summary

Remove server-side memory model definitions and Alembic migrations in favor of
memu-py's built-in database management (ddl_mode: "create"). The server now
acts purely as a wrapper — it configures and initializes MemoryService from
memu-py, which handles all ORM models, table creation, and schema management
internally.

Changes

Added

config/settings.py — centralized Settings (pydantic-settings) declaring all
environment variables with typed defaults (DATABASE_*, EMBEDDING_*,
TEMPORAL_*, STORAGE_PATH, etc.)
config/memu.py — build_memu_config() / build_memu_llm_profiles() to
construct memu-py configuration from Settings
app/services/memu.py — create_memory_service() factory
tests/test_memu_config.py — tests for the new config builders

Modified

alembic/env.py — removed app.models imports; set target_metadata = None
app/main.py — added module docstring; extracted exception messages to
variables (EM101/EM102)
tests/test_env_validation.py — moved import os to top level; added
check=False to subprocess.run
tests/test_health.py — used HTTPStatus.OK instead of magic 200;
narrowed exception catch from Exception to (ImportError, RuntimeError)
pyproject.toml — changed memu-py to memu-py[postgres]; removed
sqlmodel, sqlalchemy, pgvector, ksuid, pendulum, openai (provided
transitively by memu-py); updated deptry ignore rules

Motivation

The previous implementation duplicated model definitions that memu-py already
provides, causing import errors (app.models.base did not exist) and
architectural misalignment. Per review feedback, the server should rely on
memu-py for all memory-related schema management.

- Add BaseModel with KSUID primary keys and automatic timestamps - Implement Memory model with pgvector support for embeddings - Add MemoryCreate, MemoryRead, MemoryUpdate schemas - Migrate from SQLAlchemy DeclarativeBase to SQLModel - Update database.py to use SQLModel - Add comprehensive unit tests for BaseModel and Memory - Add svix-ksuid dependency for KSUID generation All changes pass make check (ruff, mypy, deptry, pre-commit)

Copilot

Pull request overview

Implements new SQLModel-based data modeling foundations (a shared BaseModel with KSUID primary keys + timestamps) and introduces a Memory model with pgvector embeddings, along with corresponding create/read/update schemas and unit tests.

Changes:

Replace the previous SQLAlchemy DeclarativeBase approach with a SQLModel BaseModel (KSUID IDs + timestamps).
Add Memory table/model and MemoryCreate/MemoryRead/MemoryUpdate schemas with pgvector + JSONB fields.
Update database module exports and add unit tests; introduce svix-ksuid dependency.

Reviewed changes

Copilot reviewed 7 out of 8 changed files in this pull request and generated 5 comments.

Show a summary per file

File	Description
`app/models/base.py`	Introduces SQLModel-based `BaseModel` with KSUID primary key and timestamp fields.
`app/models/memory.py`	Adds `Memory` table + schemas, including pgvector embedding and JSONB links.
`app/models/__init__.py`	Re-exports models/schemas for simpler imports.
`app/database.py`	Switches exports toward SQLModel to align with the migration.
`tests/test_base_model.py`	Adds unit tests for KSUID IDs and timestamp defaults.
`tests/test_memory.py`	Adds unit tests for `Memory` model and schemas.
`pyproject.toml`	Adds `svix-ksuid` dependency.
`uv.lock`	Locks `svix-ksuid` and transitive dependencies.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Critical fixes addressing Copilot feedback: - Add Base alias in base.py for Alembic backward compatibility - Fix timezone-aware datetimes (use datetime.now(timezone.utc) instead of utcnow()) - Add onupdate lambda for automatic updated_at refresh - Fix links type annotation from bare list to list[Any] - Export both Base and SQLModel from database.py - Correct ksuid import (lowercase ksuid() function) Test fixes: - Update test files to use timezone.utc instead of UTC - Fix KSUID length assertion (40 chars, not 27) - Ensure all datetime comparisons are timezone-aware All tests now pass (20/20) and make check succeeds.

…existence Assert against SQLModel Field metadata (model_fields[].index) rather than hasattr() checks that would pass even if indexes were removed.

Copilot

Pull request overview

Copilot reviewed 7 out of 8 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

- Set nullable=False on content Column to match required str type - Import app.models in alembic/env.py so metadata is populated - Fix bare list to list[Any] in MemoryUpdate schema

Copilot

Pull request overview

Copilot reviewed 8 out of 9 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

- Remove redundant index=True on primary key (PG indexes PKs by default) - Add sa_type=DateTime(timezone=True) to happened_at field - Update test to reflect primary key no longer has explicit index flag

Copilot

Pull request overview

Copilot reviewed 8 out of 9 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 2 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 9 out of 10 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

… intent

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Use Python logging framework instead of traceback.print_exc() so error output respects log levels and integrates with log aggregation tools in production.

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated 7 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

…ents - Add detailed NOTE comment explaining target_metadata = None in alembic/env.py and how to re-enable autogenerate for server tables. - Assert full OPENAI_API_KEY error message in test_env_validation.py. - Narrow test_health.py exception catch to (RuntimeError, ImportError). - Update psycopg deptry comment to reflect actual usage. - Remove stale pydantic_core from DEP003 ignore list.

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated 3 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

- Settings docstring now shows full resolution order: init kwargs > environment variable > .env file > default. - Add 'config' to hatch wheel packages so the built wheel includes the config package at runtime.

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated 1 comment.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated 2 comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

Replaced hand-rolled DATABASE_HOST/USER/PASSWORD/NAME fallback with Settings().DATABASE_URL which reads the same POSTGRES_* variables as the application and already normalises the DSN prefix.

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated no new comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

…parser interpolation error

Copilot

Pull request overview

Copilot reviewed 16 out of 18 changed files in this pull request and generated no new comments.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

ankaisen

@SparkZou Thank you! LGTM 👍

Copilot AI review requested due to automatic review settings February 9, 2026 04:45

Copilot started reviewing on behalf of SparkZou February 9, 2026 04:45 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread app/database.py Outdated

Comment thread app/models/memory.py Outdated

Comment thread tests/test_memory.py Outdated

Comment thread app/models/base.py Outdated

Comment thread app/models/base.py Outdated

SparkZou added 2 commits February 9, 2026 16:04

test(memory): verify actual index configuration instead of attribute …

8db1fd6

…existence Assert against SQLModel Field metadata (model_fields[].index) rather than hasattr() checks that would pass even if indexes were removed.

Copilot AI review requested due to automatic review settings February 9, 2026 07:18

Copilot started reviewing on behalf of SparkZou February 9, 2026 07:19 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread app/models/memory.py Outdated

Comment thread app/models/base.py Outdated

Comment thread pyproject.toml

fix(models): address remaining Copilot review feedback

a9e721c

- Set nullable=False on content Column to match required str type - Import app.models in alembic/env.py so metadata is populated - Fix bare list to list[Any] in MemoryUpdate schema

SparkZou requested a review from Copilot February 9, 2026 08:04

Copilot started reviewing on behalf of SparkZou February 9, 2026 08:05 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread app/models/base.py Outdated

Comment thread app/models/memory.py Outdated

Comment thread alembic/env.py Outdated

refactor(models): remove redundant index and add timezone to happened_at

fdfd3da

- Remove redundant index=True on primary key (PG indexes PKs by default) - Add sa_type=DateTime(timezone=True) to happened_at field - Update test to reflect primary key no longer has explicit index flag

SparkZou requested a review from Copilot February 9, 2026 08:20

Copilot started reviewing on behalf of SparkZou February 9, 2026 08:20 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread tests/test_base_model.py Outdated

chore(alembic): add migration to create memories table

61c8ac2

SparkZou requested a review from Copilot February 9, 2026 09:08

Copilot started reviewing on behalf of SparkZou February 9, 2026 09:09 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread alembic/versions/52875d5f3609_create_memories_table.py Outdated

Comment thread pyproject.toml Outdated

fix(alembic): ensure pgvector extension before creating memories table

2403a7f

SparkZou requested a review from Copilot February 9, 2026 09:22

Copilot started reviewing on behalf of SparkZou February 9, 2026 09:23 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread tests/test_memory.py Outdated

fix(tests): correct docstring to match list[Any] type for links field

6d39b94

SparkZou requested a review from Copilot February 9, 2026 09:34

Copilot started reviewing on behalf of SparkZou February 9, 2026 09:34 View session

Copilot AI reviewed Feb 9, 2026

View reviewed changes

Comment thread alembic/versions/52875d5f3609_create_memories_table.py Outdated

docs(alembic): document pgvector extension prerequisite and downgrade…

3834ab1

… intent

Copilot started reviewing on behalf of SparkZou February 16, 2026 00:35 View session