The examples in scripts/examples/ as well as the tutorials in doc/tutorial/ are quite outdated by now. Most of them focus on showcasing minor syntactical aspects of DaphneDSL and DaphneLib, but I'd say they are not really interesting from a user's point of view and don't reflect what can meanwhile be achieved using DAPHNE.
It would be great to show some actually meaningful integrated data analysis pipelines, which
Read some real-world data set from a file (or receive it from a Python library).
Execute some meaningful ML algorithm, query processing, and/or simulation on the data (ideally using reusable high-level primitives from scripts/algorithms/).
Output something meaningful (print an insight, write some data to a file, or transfer data back to Python).
These examples can still focus on certain aspects (like using hardware accelerators, Python data exchange, distributed computing, etc.).
Ideally there should be corresponding test cases that make sure that DAPHNE can successfully compile and run the examples at any point in time.
pdamme
commented
3 months ago
The examples in
scripts/examples/as well as the tutorials indoc/tutorial/are quite outdated by now. Most of them focus on showcasing minor syntactical aspects of DaphneDSL and DaphneLib, but I'd say they are not really interesting from a user's point of view and don't reflect what can meanwhile be achieved using DAPHNE.It would be great to show some actually meaningful integrated data analysis pipelines, which
scripts/algorithms/).These examples can still focus on certain aspects (like using hardware accelerators, Python data exchange, distributed computing, etc.).
Ideally there should be corresponding test cases that make sure that DAPHNE can successfully compile and run the examples at any point in time.