The Day Four presentations of the 5th annual Python Web Conf are now up on LoudSwarm for re-watch.
KEYNOTE: Python and the Future of Fusion Energy
by DR. LISA REUSCH @ SHINE TECHNOLOGIES
Key points:
- Cool introduction to fusion science;
- Python is a standard tool for scientific computations in the fusion community;
- Data aggregation helps address the limitations of the harsh fusion environment;
- Lots of opportunities for experts in big data, ML and data driven processing technologies.
KEYNOTE: No Holds Barred Web Framework Battle
by DANIEL ROY GREENFELD @ KRAKEN TECHNOLOGIES
Key points:
- Pros and cons of the popular Python frameworks Django/FastAPI/Flask;
- It’s possible to build anything from small to big systems with any of them;
- Discussion on developer experience, speed metrics, databases for all three frameworks.
TUTORIAL: We've "REST"ed long enough -- let's have an API adventure!
by W. IAN DOUGLAS @ POSTMAN
Key points:
- REST is easy but limited; today's apps are event-driven and interacts with various servers;
- The leading options (GraphQL Subscriptions, websockets, and server-sent events) come with limitations;
- Enter gRPC: built with performance in mind + can be synchronous or asynchronous;
- When choosing which style of APIs to create, consider performance as well as your team’s capabilities.
Finding and fixing an N+1 problem in the Django docs
by ANTON PIRKER @ SENTRY
Key points:
- It’s easy to accidentally create an N+1 problem, especially when using an ORM;
- There are easy fixes in Django and SQLAlchemy;
- You don’t need to worry about an N+1 query that is not creating a problem;
- Django Debug Toolbar, nplusone, database-specific tools and monitoring tools like Sentry can be used to help find N+1 problems.
Proactive cluster autoscaling in Kubernetes
by DANIELE POLENCIC @ LEARNK8S
Key points:
- Elastic scaling nodes in Kubernetes can take several minutes on default settings;
- Scheduler plays Tetris perfecting the sizing of nodes and pods;
- For scaling, configure pod priorities to deploy and pre-warm a new node;
- Overprovision for occasional traffic spikes.
A fresh take on monorepos in Python
by DAVID VUJIC @ BIOLIT AB
Key points:
- A monorepo is a place you put all of your code for one or several apps;
- Polylith combines libraries and bricks (bases + components) to build projects;
- Poetry helps determine how everything is set up for each project, and when releases need to happen.,
Software Best Practices for Data Pipelines: Form Energy's Multi-Day Storage R&D Pipeline
by MARK FICKETT @ FORM ENERGY
Key points:
- Rapid feedback from pipelines is necessary for rapid innovation;
- Use structured events rather than logs;
- Example layering OpenTelemetry, HoneyComb, fluentbit and Grafana;
- Achieve near-real time pipelines with near-real time system feedback;
- Identify trends over arbitrary time-frames for telemetry data.
Optimizing Django deployments with a Continuous Observability Strategy
by CHAD CARLSON @ PLATFORM.SH
Key points:
- Fixing issues in production is a lot more expensive than in development;
- Blackfire fits in Platform.sh to form a Continuous Observability strategy that optimizes the monitoring, profiling and testing experience;
- Demo of Blackfire debugging a real performance issue on Django app.
Building & testing Cloud applications locally using LocalStack
by HARSH BARDHAN MISHRA @ LOCALSTACK
Key points:
- LocalStack is a core AWS cloud emulator that ships as a single docker container;
- LocalStack comes with a web UI with a simplified AWS console for 25+ services, and stack insights to analyze API usage;
- LocalStack can be used with tools like Terraform;
- LocalStack is highly configurable, with sensible defaults.
How to be More Influential in a World that’s Tired of Feeling Manipulated by Technology
by ERIN FUSARO @ MARCO POLO
Key points:
- People hate networking, largely because it is transactional;
- Typical social media has made us feel more isolated;
- In the brain, the feeling of exclusion is registered the same way as physical pain;
- To build meaningful connections, focus on quality over quantity;
- Build trust by showing vulnerability and being supportive.
Cloud Native Apache Pulsar Development 202 with Python
by TIMOTHY SPANN @ CLOUDERA
Key points:
- Apache Pulsar is a messaging and streaming platform with geo-replication, scaling, multi-tenancy, and queuing;
- Example setting up Python code to consume Pulsar service for turning live event data into live NLP results and sentiment analysis;
- Example for real-time NLP analytics to gain insight on how and when to use various streaming protocols.
