luigi
direktiv
luigi | direktiv | |
---|---|---|
14 | 13 | |
17,327 | 464 | |
0.5% | 1.5% | |
6.3 | 10.0 | |
9 days ago | 3 days ago | |
Python | TypeScript | |
Apache License 2.0 | Apache License 2.0 |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
luigi
-
Ask HN: What is the correct way to deal with pipelines?
I agree there are many options in this space. Two others to consider:
- https://airflow.apache.org/
- https://github.com/spotify/luigi
There are also many Kubernetes based options out there. For the specific use case you specified, you might even consider a plain old Makefile and incrond if you expect these all to run on a single host and be triggered by a new file showing up in a directory…
-
In the context of Python what is a Bob Job?
Maybe if your use case is “smallish” and doesn’t require the whole studio suite you could check out apscheduler for doing python “tasks” on a schedule and luigi to build pipelines.
-
Lessons Learned from Running Apache Airflow at Scale
What are you trying to do? Distributed scheduler with a single instance? No database? Are you sure you don't just mean "a scheduler" ala Luigi? https://github.com/spotify/luigi
-
Apache Airflow. How to make the complex workflow as an easy job
It's good to know what Airflow is not the only one on the market. There are Dagster and Spotify Luigi and others. But they have different pros and cons, be sure that you did a good investigation on the market to choose the best suitable tool for your tasks.
-
DevOps Fundamentals for Deep Learning Engineers
MLOps is a HUGE area to explore, and not surprisingly, there are many startups showing up in this space. If you want to get it on the latest trends, then I would look at workflow orchestration frameworks such as Metaflow (started off at Netflix, is now spinning off into its own enterprise business, https://metaflow.org/), Kubeflow (used at Google, https://www.kubeflow.org/), Airflow (used at Airbnb, https://airflow.apache.org/), and Luigi (used at Spotify, https://github.com/spotify/luigi). Then you have the model serving itself, so there is Seldon (https://www.seldon.io/), Torchserve (https://pytorch.org/serve/), and TensorFlow Serving (https://www.tensorflow.org/tfx/guide/serving). You also have the actual export and transfer of DL models, and ONNX is the most popular here (https://onnx.ai/). Spark (https://spark.apache.org/) still holds up nicely after all these years, especially if you are doing batch predictions on massive amount of data. There is also the GitFlow way of doing things and Data Version Control (DVC, https://dvc.org/) is taken a pole position there.
-
Data pipelines with Luigi
At Wonderflow we're doing a lot of ML / NLP using Python and recently we are enjoying writing data pipelines using Spotify's Luigi.
- Noobie who is trying to use K8s needs confirmation to know if this is the way or he is overestimating Kubernetes.
-
Open Source ETL Project For Startups
💡【About Luigi】 https://github.com/spotify/luigi Luigi was built at Spotify since 2012, it's open source and mainly used for getting data insights by showing recommendations, toplists, A/B test analysis, external reports, internal dashboards, etc.
- Resources/tutorials to help me learn about ETL?
-
Using Terraform to make my many side-projects 'pick up and play'
So to sum that up, I went from having nothing for my side-project set up in AWS to having a Kubernetes cluster with the basic metrics and dashboard, a proper IAM-linked ServiceAccount support for a smooth IAM experience in K8s, and Luigi deployed so that I could then run a Luigi workflow using an ad-hoc run of a CronJob. That's quite remarkable to me. All that took hours to figure out and define when I first did it, over six months ago.
direktiv
-
Preloading Ollama Models
That went okay, but there is still the startup problem - it took ages to run the lifecycle hook, plus it won't function on Kubernetes nodes with no internet access. At Direktiv were are using Knative a lot as well which does not support lifecycle events. So, my plan was to create a container using the Ollama image as base with the model pre-downloaded.
-
Knative Serverless in 2024
When deciding which option to choose, consider your specific environment, requirements, and preferences. At Direktiv, we typically opt for Contour due to its simplicity. However, your choice may vary depending on your use case and infrastructure setup.
-
Lessons Learned from Running Apache Airflow at Scale
So being completely transparent, we're the creators of Direktiv (https://github.com/direktiv/direktiv). We're genuinely curious to have users who have previously used Airflow and other DAGs (mentioned in here is Argo workflows) try Direktiv and give us more feedback.
- direktiv runs containers as part of workflows from any compliant container registry, passing JSON structured data between workflow states.
- Encrypting server-side emails using serverless workflows using Direktiv
- Encrypting server-side emails using serverless workflows
-
Step Functions Wait Loop w/ Timeout Feature
If you want a portable step functions take a look at https://github.com/vorteil/direktiv developers are very helpful and responsive
-
Direktiv: Docker development environment, VSCode plugin & Infrastructure-as-a-Chatbot
Another update to our Direktiv event-driven serverless workflow engine - but this one focused on development. Release v0.3.1 included some bug fixes, improved stability and security enhancements, but more notably:
-
Update to our serverless workflow engine Direktiv
We've previously posted on our serverless workflow / automation engine called Direktiv and wanted to share a couple of updates:
What are some alternatives?
Airflow - Apache Airflow - A platform to programmatically author, schedule, and monitor workflows
direktiv-apps - Direktiv Application Containers
Kedro - Kedro is a toolbox for production-ready data science. It uses software engineering best practices to help you create data engineering and data science pipelines that are reproducible, maintainable, and modular.
states-language-cadence - States Language on Cadence
Apache Spark - Apache Spark - A unified analytics engine for large-scale data processing
argo - Workflow Engine for Kubernetes
mrjob - Run MapReduce jobs on Hadoop or Amazon Web Services
windmill - Open-source developer platform to turn scripts into workflows and UIs. Fastest workflow engine (5x vs Airflow). Open-source alternative to Airplane and Retool.
Dask - Parallel computing with task scheduling
toil - A scalable, efficient, cross-platform (Linux/macOS) and easy-to-use workflow engine in pure Python.
Pinball
stepwise - Clojure AWS Step Functions library