Seeders

Seeders populate the database with initial or test data. One per feature, executed in dependency order.

Table of contents

• What they do
• Creating a seeder
• Real example: profile seeder
• How seed() works
• Running seeders
• Execution order
• Convention
• See also

---

What they do

When you need users, sample data, or configuration records in the database, you write a seeder. The CLI runs all seeders in the right order so dependencies are satisfied (auth before profile, notes before notes_tags).

---

Creating a seeder

# splent_feature_auth/seeders.py
from splent_framework.seeders.BaseSeeder import BaseSeeder
from splent_io.splent_feature_auth.models import User

class AuthSeeder(BaseSeeder):
    def run(self):
        users = [
            User(email="user1@example.com", password="1234"),
            User(email="user2@example.com", password="1234"),
        ]
        self.seed(users)

Implement run() with your seeding logic. Call self.seed(data) to bulk insert.

---

Real example: profile seeder

Profile depends on auth — users must exist first. The seeder can safely assume users are already seeded because execution order follows UVL constraints:

# splent_feature_profile/seeders.py
from splent_framework.seeders.BaseSeeder import BaseSeeder
from splent_io.splent_feature_auth.models import User
from splent_io.splent_feature_profile.models import UserProfile

class ProfileSeeder(BaseSeeder):
    def run(self):
        users = User.query.all()
        names = [("John", "Doe"), ("Jane", "Doe")]

        profiles = []
        for user, (name, surname) in zip(users, names):
            if not user.profile:
                profiles.append(
                    UserProfile(user_id=user.id, name=name, surname=surname)
                )

        if profiles:
            self.seed(profiles)

No need to check if auth ran first — UVL guarantees it.

---

How seed() works

1. Validates all objects are the same model type
2. Bulk inserts via db.session.add_all()
3. Commits the transaction
4. Returns the list with database-assigned IDs

On IntegrityError (duplicate key, constraint violation), rolls back and raises SeederError.

---

Running seeders

splent db:seed                          # all seeders in UVL order
splent db:seed splent_feature_auth      # only auth
splent db:seed --reset -y               # wipe DB first, then seed all

---

Execution order

db:seed discovers all seeders.py files across declared features and runs them in UVL topological order — the same order features load at runtime. If profile requires auth in the UVL model, AuthSeeder always runs before ProfileSeeder.

No manual priority attribute. The UVL constraints handle it.

---

Convention

• One seeders.py per feature
• One seeder class per file
• db:seed discovers it by finding the first class that extends BaseSeeder
• Name it whatever you want (AuthSeeder, NotesSeed, etc.)

---

See also

• db:seed — the CLI command that runs seeders
• Repositories — the data layer seeders write to