Skip to content

Latest commit

 

History

History
273 lines (185 loc) · 7.13 KB

WP3.md

File metadata and controls

273 lines (185 loc) · 7.13 KB
Raw
layout title theme transition author period wp
slides_wp
WorkPackage Report
white
none
Luca De Feo
1st reporting period
3
<style>.reveal div { font-size: 36px } .reveal h3 { margin-top: 2ex; margin-bottom: 1ex; }</style>
## Component architecture

Components

  • User interfaces: Jupyter, SageMathCloud;
  • Systems: GAP, PARI/GP, SageMath, Singular;
  • Software libraries: LinBox, MPIR, ...;
  • Modules, packages, ...;
  • Databases: LMFDB, MathHub, ....

How do we want them? Flexible, Portable, Modular, Open.

--

Goals

  • Flexible. Enable components to be flexibly combined;
  • Portable. Enable components run smoothly across a wide range of environments (e.g., cloud, local, server, ...);
  • Modular. Develop, demonstrate, standardize APIs;
  • Open. Streamline user/developer workflows.

--

Examples

  • SageMath's elliptic curves module calls C functions in PARI/GP library.
  • Import data from LMFDB at the GAP command line.
  • Run commands on remote Singular server from a GAP client through MMT mediator.
  • Deploy SageMath virtual machines on Windows machines in Microsoft Azure cloud.
  • Deploy a computational mathematics software stack on a HPC cluster, controlled through Jupyter notebooks.

--

Inputs

  • WP4 → User interface components;
  • WP5 → HPC-enabled components;
  • WP6 → Interface languages, ontologies;
  • WP7 → User behaviour.

Outputs

  • Network protocol implementations → WP4, WP6;
  • Tools, infrastructure (for testing, bug reporting, packaging, distributing) → WP7;
  • Micromagnetic components → WP4.

Tasks

T3.1, Portability

  • (mostly) Windows;
  • Testing, continuous integration.

T3.2, Interfaces between systems

  • Application level interfaces;
  • Network protocols.

Delivered: D3.3, SCSCP support.

--

T3.3, Modularisation and packaging

  • Virtual machines, containers;
  • Packaging and distributing user code;
  • Packaging for popular repositories (Debian, Anaconda, ...).

Delivered: D3.1, Virtual images and containers.

T3.4, Simulagora integration

  • Commercial VRE, showcasing OpenDreamKit components.
  • Interest for the project: measure effort needed to produce a Debian-based VM.

Note​: Internal task, no public output.

--

T3.6, Document and modularise SageMathCloud's codebase

  • Collaboration on the open source software powering the commercial service SageMathCloud.
  • Share technologies, know-how.

Delivered: D3.2, Understand and docuement SageMathCloud backend

T3.7, Improving the development workflow in mathematical software

  • Experiments with improving user/developer experience in VREs;
  • Integration into SageMathCloud.

--

T3.8, Python interface for OOMMF micromagnetic simulation library

  • See H. Fanghor's talk.
  • Completed!

Note: no deliverables in this WP.

T3.5, Component architecture for High Performance Computing and Parallelism

  • Experiment: integrate outputs of WP5 into SageMath;

Note: Only starts in month 36.

Deliverables

{% comment %} The include below gets the finished deliverables of the work pakacge and put the list in "delivs" {% endcomment %}

{% include deliverables_slides.html %}

{{ delivs | size }} deliverables were submitted.

{% for d in delivs %}

  • {{ d }} {% endfor %}

--

{{ delivs[0] }}

Output: Virtual images and containers for most of the relevant components of OpenDreamKit.

Based on modern and popular technologies: Ansible, Docker, Vagrant, VirtualBox.

Partners involved: JacobsUni, UPSud, USouthampton, USTAN, UVersailles.

--

{{ delivs[1] }}

Output: Published documentation on the internals of SageMathCloud

  • Part of a planned interaction with SageMathCloud (SMC).
  • Development cycles in SMC are very short, architecture, technologies change quickly.
  • ODK benefits from keeping track of SMC internal details.
  • SMC benefits from publishing better documentation.

Partners involved: UPSud.

--

{{ delivs[2] }}

Output: Implementations of SCSCP in GAP, Python, MMT/MathHub

  • SCSCP = Symbolic Computation Software Composability Protocol.
  • Network protocol (XML-based) for RPC on semantically annotated mathematical data.
  • Developed in EU FP6 project 026133 SCIEnce, standardized by the OpenMath consortium.
  • Implementations already available in Java, C, C++. This deliverable: extend spectrum of available implementations.

Partners involved: CNRS, JacobsUni, USTAN, UVSQ.

--

GAP

Adds SCSCP support to: GAP, Singular

MathHub

--

Python

Adds SCSCP support to: Python, LMFDB, SageMath, PARI/GP (via D4.10)

Achievements

  • Two workshops on modularization and packaging, great occasion to interact with the community:

    • Sage Days 77, March 2016, Cernay (France);
    • Sage Days 85, March 2017, Cernay (France).

  • Interactions with other work packages: WP4, WP6.

  • OOMMF task (Python interface to OOMMF) completed. See H. Fanghor talk.

Upcoming work

Four deliverables due on month 24:

  • D3.4: Single user SageMathCloud.
  • D3.5: Integration between SageMathCloud and Trac (SageMath bug tracking system).
  • D3.6: Open package repository for SageMath.
  • D3.7: One-click install SageMath distribution for Windows.

--

D3.4: Single user SageMathCloud

Good news!

Already done by SMC developers, will re-allocate efforts.

D3.5: Integration between SageMathCloud and Trac

Ongoing discussions with SMC developers.

D3.6: Open package repository for SageMath

Current plan is to provide cookiecutter templates for hosting packages on PyPI (most popular Python package repository).

--

D3.7: One-click install SageMath distribution for Windows.

This deliverable is mostly done:

  • Erik Bray (UPsud) has already completed a one-click install distribution for Windows based on Docker (not original plan, but important milestone).
  • The latest version of SageMath compiles on Cygwin!

Longer term

  • Packaging: With help from the community, SageMath will be in the next Debian (Linux) release. More than 24 months ahead of schedule!

  • Packaging (bis): experimental Conda packages for SageMath and components are being made, with the help of the community. This is not a planned deliverable, but it would be a pity to miss this opportunity.

  • GAP/SageMath Interface: experimental work started in January 2016. First prototype complete.