cledar-sdk 1.1.0

Cledar Python SDK

Cledar Python SDK

Project Description

Cledar Python SDK is a shared set of production‑ready services and utilities used across Cledar projects. It can be installed from PyPI (recommended), or consumed as a Git dependency or Git submodule.

Included modules:

• kafka_service: Kafka Producer/Consumer, helpers and DLQ handler
• storage_service: Object storage abstraction (S3/ABFS/local via fsspec)
• monitoring_service: FastAPI monitoring server with Prometheus metrics and healthchecks
• redis_service: Redis‑backed typed config store
• kserve_service: KServe helpers
• common_logging: Common logging utilities

---

Installation and Setup

1. From PyPI (recommended)

   Using pip:

   pip install cledar-sdk
   

   Using uv:

   uv add cledar-sdk
   

   Pin a specific version (example):

   pip install "cledar-sdk==1.0.1"
   

2. From Git (alternative)

   Using pip (SSH, specific tag):

   pip install "git+ssh://[e-mail]/Cledar/cledar-python-sdk.git@v1.0.1"
   

   Using uv (SSH, specific tag):

   uv add --git ssh://[e-mail]/Cledar/cledar-python-sdk.git@v1.0.1
   

   You can also point to a branch (e.g. main) instead of a tag.

3. As a Git submodule

   git submodule add [e-mail]:Cledar/cledar-python-sdk.git vendor/cledar-python-sdk
   git submodule update --init --recursive
   

   Optionally install it in editable mode from the submodule path:

   uv add -e ./vendor/cledar-python-sdk
   

4. Developing locally

   git clone [e-mail]/Cledar/cledar-python-sdk.git
   cd cledar-python-sdk
   uv sync
   

Python version required: 3.12.7

Testing

Unit tests are implemented using pytest and unittest.

1. Run tests:

   uv run pytest
   

2. Adding tests: Place tests under each module's tests directory (e.g. kafka_service/tests, storage_service/tests) or create files with the _test.py suffix.

---

Quick Start Examples

Kafka

Producer:

from kafka_service.clients.producer import KafkaProducer
from kafka_service.config.schemas import KafkaProducerConfig

cfg = KafkaProducerConfig(
    kafka_servers="localhost:9092",
    kafka_group_id="example",
    kafka_topic_prefix="dev",
    compression_type="snappy",
    kafka_partitioner="consistent_random",
)
producer = KafkaProducer(config=cfg)
producer.connect()
producer.send(topic="my-topic", value='{"id":"123","payload":"hello"}', key="123")

Consumer:

from kafka_service.clients.consumer import KafkaConsumer
from kafka_service.config.schemas import KafkaConsumerConfig

cfg = KafkaConsumerConfig(
    kafka_servers="localhost:9092",
    kafka_group_id="example",
    kafka_topic_prefix="dev",
    kafka_offset="earliest",
    kafka_auto_commit_interval_ms=5000,
)
consumer = KafkaConsumer(config=cfg)
consumer.connect()
consumer.subscribe(["my-topic"])
msg = consumer.consume_next()
if msg:
    consumer.commit(msg)

Object Storage (S3/ABFS/local)

from storage_service.object_storage import ObjectStorageService
from storage_service.models import ObjectStorageServiceConfig

cfg = ObjectStorageServiceConfig(
    s3_access_key="minioadmin",
    s3_secret_key="minioadmin",
    s3_endpoint_url="http://localhost:9000",
)
storage = ObjectStorageService(config=cfg)
storage.upload_file(
    file_path="README.md",
    destination_path="s3://bucket/path/README.md",
)

Monitoring Server

from monitoring_service.monitoring_server import MonitoringServer, MonitoringServerConfig

config = MonitoringServerConfig(
    readiness_checks={"s3": storage.is_alive},
    liveness_checks={"app": lambda: True},
)
server = MonitoringServer(host="0.0.0.0", port=8080, config=config)
server.start_monitoring_server()

Redis Config Store

from redis import Redis
from redis_service.redis_config_store import RedisConfigStore

redis = Redis(host="localhost", port=6379, db=0)
store = RedisConfigStore(redis=redis, prefix="example:")
# See redis_service/example.py for a full typed config provider example

Code Quality

• pydantic - settings management
• ruff, mypy - Linting, formatting, and static type checking
• pre-commit - Pre-commit file checks

Linting

If you want to run linting or type checker manually, you can use the following commands. Pre-commit will run these checks automatically before each commit.

uv run ruff format .
uv run ruff check .
uv run mypy .

Pre-commit setup

To get started follow these steps:

1. Install pre-commit by running the following command:

   pip install pre-commit
   

2. Once pre-commit is installed, set up the pre-commit hooks by running:

   pre-commit install
   

3. Pre-commit hooks will analyze only committed files. To analyze all files after installation run the following:

   pre-commit run --all-files
   

Automatic Fixing Before Commits:

pre-commit will run Ruff (format + lint) and mypy during the commit process:

git commit -m "Describe your changes"

To skip pre-commit hooks for a single commit, use the --no-verify flag:

```bash
git commit -m "Your commit message" --no-verify
```

---

Technologies and Libraries

Main Dependencies:

• python >= "3.12.7"
• pydantic-settings
• confluent-kafka
• fastapi
• prometheus-client
• uvicorn
• redis
• fsspec/s3fs/adlfs (S3/ABFS backends)
• boto3 and boto3-stubs

Developer Tools:

• uv - Dependency and environment management
• pydantic - settings management
• ruff - Linting and formatting
• mypy - Static type checker
• pytest, unittest - Unit tests
• pre-commit - Code quality hooks

---

Commit conventions

We use Conventional Commits for our commit messages. This helps us to create a better, more readable changelog.

Example of a commit message:

refactor(XXX-NNN): spaghetti code is now a carbonara