GeezerVison

matrix:

os: ["ubuntu", "windows", "macos"]

python: ["3.7", "3.8", "3.9", "3.10", "3.11", "3.12"] # https://devguide.python.org/versions/#versions

exclude:

- os: "macos"

python: "3.7" # Prevent "The version '3.7' with architecture 'arm64' was not found for macOS 14.4.1"

runs-on: ${{ matrix.os }}-latest

steps:

**/__pycache__

**/__pycache__

env/

ObjectBox Python ChangeLog

4.0.0 (2024-05-16)

* ObjectBox now supports vector search ("vector database") to enable efficient similarity searches.

This is particularly useful for AI/ML/RAG applications, e.g. image, audio, or text similarity.

Other use cases include sematic search or recommendation engines.

See https://docs.objectbox.io/ann-vector-search for details.

* Queries: support for Property-based conditions and logic combinations

* Convenient "Store" API deprecates ObjectBox and Builder API

* New examples added, illustrating an VectorSearch and AI/RAG application

* Dependency flatbuffers: Updated to 24.3.50

* Adjusting the version number to match the core version (4.0); we will be aligning on major versions from now on.

Older Versions

Please check https://github.com/objectbox/objectbox-python/releases for details.

rm -rf *.egg-info

publish: ## Publish the package built by `make build`

set -e ; \

${PYTHON} -m twine upload --verbose dist/objectbox*.whl

set -e

@echo "****************************************************************"

@echo ">>> Please enter the API token when asked for a password. <<<"

@echo ">>> The API token starts with the prefix 'pypi-'. <<<"

@echo ">>> See https://pypi.org/help/#apitoken for details. <<<"

@echo "****************************************************************"

${PYTHON} -m twine upload -u "__token__" --verbose dist/objectbox*.whl

ObjectBox Python API

[ObjectBox](https://objectbox.io) is a superfast database for objects, now also available for Python (3.4+) with a simple CRUD API.

ObjectBox Python

[ObjectBox](https://objectbox.io) Python is a lightweight yet powerful on-device database & vector database.

Store Python objects and vectors directly with an easy-to-use CRUD API while enjoying exceptional speed and efficiency.

And because it's an embedded database, there's no setup required.

- [Getting Started](#getting-started)

- [Model IDs and UIDs](#model-ids-and-uids)

- [model.py](#modelpy)

- [Using ObjectBox](#using-objectbox)

- [Some features](#some-features)

- [Coming in the future](#coming-in-the-future)

- [Help wanted](#help-wanted)

- [Feedback](#feedback)

- [License](#license)

Its advanced vector search empowers AI for a variety of applications, including RAG AI, generative AI,

and similarity searches.

pip install --upgrade objectbox

To start using ObjectBox as a storage for your data, you need to define your model first.

The model consists of Python classes annotated with `@Entity` decorator.

_Table of Contents_

Each Entity has to have an ID (unique among entities).

Properties need an ID as well (unique inside one Entity).

Both Entities and Properties must also have an UID, which is a globally unique identifier.

For other ObjectBox supported languages, the binding takes care of assigning these IDs/UIDs but this feature is not yet implemented for Python.

To learn more, see [ObjectBox Java documentation](https://docs.objectbox.io/advanced/meta-model-ids-and-uids)

- [Feature Highlights](#feature-highlights)

- [Code Example (CRUD - Create, Read, Update, Delete)](#code-example-crud---create-read-update-delete)

- [Getting Started](#getting-started)

- [Alpha Notes](#alpha-notes)

- [Help wanted](#help-wanted)

- [Feedback](#feedback)

- [License](#license)

from objectbox.model import *

@Entity(id=1, uid=1)

class Person:

id = Id(id=1, uid=1001)

name = Property(str, id=2, uid=1002)

is_enabled = Property(bool, id=3, uid=1003)

# int can be stored with 64 (default), 32, 16 or 8 bit precision.

int64 = Property(int, id=4, uid=1004)

int32 = Property(int, type=PropertyType.int, id=5, uid=1005)

int16 = Property(int, type=PropertyType.short, id=6, uid=1006)

int8 = Property(int, type=PropertyType.byte, id=7, uid=1007)

# float can be stored with 64 or 32 (default) bit precision.

float64 = Property(float, id=8, uid=1008)

float32 = Property(float, type=PropertyType.float, id=9, uid=1009)

byte_array = Property(bytes, id=10, uid=1010)

# Regular properties are not stored.

transient = ""

Feature Highlights

🏁 **On-device vector database** - for AI apps that work any place.\

🏁 **High performance** - superfast response rates enabling real-time applications.\

🪂 **ACID compliant** - Atomic, Consistent, Isolated, Durable.\

🌱 **Scalable** - grows with your app, handling millions of objects with ease.\

💚 **Sustainable** - frugal on CPU, Memory and battery / power use, reducing CO2 emissions.\

💐 **[Queries](https://docs.objectbox.io/queries)** - filter data as needed, even across relations.\

💻 **Multiplatform** - Get native speed on your favorite platforms.\

* Linux x86-64 (64-bit)

* Linux ARMv6hf (e.g. Raspberry PI Zero)

* Linux ARMv7hf (e.g. Raspberry PI 3)

* Linux ARMv8 (e.g. Raspberry PI 4, 5, etc.)

* MacOS x86-64 and arm64 (Intel 64-bit and Apple Silicon)

* Windows x86-64 (64-bit)

To actually use the database, you launch (or "build") it with the model you've just defined.

Afterwards, you can reuse the instance (`ob` in the example below) and use it to access "Entity Boxes" which hold your objects.

import objectbox

model = objectbox.Model()

model.entity(Person, last_property_id=objectbox.model.IdUid(10, 1010))

model.last_entity_id = objectbox.model.IdUid(1, 1)

ob = objectbox.Builder().model(model).directory("db").build()

store = objectbox.Store(model=model)

box = objectbox.Box(ob, Person)

box = store.box(Person)

person = Person()

person.name = "Joe Green"

box.remove(person) # Delete

Additionally, see the [TaskList example app](https://github.com/objectbox/objectbox-python/tree/main/example). After checking out this repository to run the example:

Getting started

Latest version: 4.0.0a0 (2024-05-15)

To install or update the latest version of ObjectBox, run this:

pip install --upgrade --pre objectbox # "--pre" because you want to get the 4.0.0 alpha version

Now you are ready to use ObjectBox in your Python project.

Head over to the **[ObjectBox documentation](https://docs.objectbox.io)**

and learn how to setup your first entity classes.

Do you prefer to dive right into working examples?

We have you covered in the [example](example/) folder.

It comes with a task list application and a vector search example using cities.

Additionally, for AI enthusiasts, we provide an "ollama" example,

which integrates a local LLM (via [ollama](https://ollama.com))

with ObjectBox to manage and search embeddings effectively.

For more information and code examples, see the tests folder. The docs for other languages may also help you understand the basics.

* ObjectBox Java/Dart/Flutter - https://docs.objectbox.io

* ObjectBox Go - https://golang.objectbox.io

* ObjectBox Swift - https://swift.objectbox.io

Some features

* Automatic transactions (ACID compliant)

* Bulk operations

* Vector types, e.g. for AI vector embeddings

* Platforms supported with native speed:

* Linux x86-64 (64-bit)

* Linux ARMv6hf (e.g. Raspberry PI Zero)

* Linux ARMv7hf (e.g. Raspberry PI 3; available only on request)

* Linux ARMv8 (e.g. Raspberry PI 4)

* MacOS x86-64 and arm64 (Intel 64-bit and Apple Silicon)

* Windows x86-64 (64-bit)

Coming in the future

The goodness you know from the other ObjectBox language-bindings, e.g.,

* model management (no need to manually set id/uid)

* automatic model migration (no schema upgrade scripts etc.)

* powerful queries

* relations (to-one, to-many)

* asynchronous operations

* secondary indexes

Alpha Notes

While ObjectBox Python is powered by a rock stable core written in C/C++, we label our Python binding still "alpha."

We do this to manage expectations as some quality of life improvements are yet to come to our Python binding.

This is mostly about "model management," which still requires you to do some manual coding setup, e.g. for model IDs.

The final release will take care of this for you automatically.

Help wanted

ObjectBox for Python is still in an early stage with limited feature set (compared to our other supported languages).

To bring all these features to Python, we're asking the community to help out. PRs are more than welcome!

ObjectBox for Python is open to contributions.

The ObjectBox team will try its best to guide you and answer questions.

See [CONTRIBUTING.md](https://github.com/objectbox/objectbox-python/blob/main/CONTRIBUTING.md) to get started.

Feedback

Also, please let us know your feedback by opening an issue: for example, if you experience errors or if you have ideas

for how to improve the API. Thanks!

We are looking for your feedback!

Please let us know what you think about ObjectBox for Python and how we can improve it.

License

import objectbox

import time

from tests.model import TestEntity

from tests.common import remove_test_dir, load_empty_test_objectbox

from tests.common import remove_test_dir, load_empty_test_default_store

class ObjectBoxPerf:

def __init__(self):

self.ob = load_empty_test_objectbox()

self.box = objectbox.Box(self.ob, TestEntity)

self.store = load_empty_test_default_store()

self.box = store.box(TestEntity)

def remove_all(self):

self.box.remove_all()

version = "v0.21.0" # see objectbox/c.py required_version

version = "v4.0.0" # see objectbox/c.py required_version

variant = 'objectbox' # or 'objectbox-sync'

base_url = "https://github.com/objectbox/objectbox-c/releases/download/"

# Download the file from `url`, save it in a temporary directory and get the path to it (e.g. '/tmp/tmpb48zma')

source_url = url_for(rel_path);

print(f"URL {source_url}")

tmp_file, headers = urllib.request.urlretrieve(source_url)

# extract the file

The following examples are available from this repository.

This is our classic Tasks application using a CLI.

This example application starts with a pre-defined set of capital cities and their geo coordinates.

It allows to search for nearest neighbors of a city (`city_neighbors`) or by coordinates (`neighbors`) as well as adding more locations (`add`).