BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//pretalx//pretalx.com//euroscipy-2024//speaker//QXKHMS
BEGIN:VTIMEZONE
TZID:Europe/Berlin
BEGIN:DAYLIGHT
DTSTART:20230829T000000
TZNAME:CEST
TZOFFSETFROM:+0200
TZOFFSETTO:+0200
END:DAYLIGHT
BEGIN:STANDARD
DTSTART:20231029T030000
RDATE:20241027T030000
TZNAME:CET
TZOFFSETFROM:+0200
TZOFFSETTO:+0100
END:STANDARD
BEGIN:DAYLIGHT
DTSTART:20240331T030000
RDATE:20250330T030000
TZNAME:CEST
TZOFFSETFROM:+0100
TZOFFSETTO:+0200
END:DAYLIGHT
END:VTIMEZONE
BEGIN:VEVENT
SUMMARY:Understanding NetworkX's API Dispatching with a parallel backend -
  Erik Welch\, Aditi Juneja
DTSTART;TZID=Europe/Berlin:20240828T110500
DTEND;TZID=Europe/Berlin:20240828T113500
DTSTAMP:20260907T151923Z
UID:pretalx-euroscipy-2024-QLVBYY@pretalx.com
DESCRIPTION:Hi! Have you ever wished your pure Python libraries were faste
 r? Or wanted to fundamentally improve a Python library by rewriting everyt
 hing in a faster language like C or Rust? Well\, wish no more... NetworkX'
 s backend dispatching mechanism redirects your plain old NetworkX function
  calls to a FASTER implementation present in a separate backend package by
  leveraging the Python's [`entry_point`](https://packaging.python.org/en/l
 atest/specifications/entry-points) specification!\n\nNetworkX is a popular
 \, pure Python library used for graph(aka network) analysis. But when the 
 graph size increases (like a network of everyone in the world)\, then Netw
 orkX algorithms could take days to solve a simple graph analysis problem. 
 So\, to address these performance issues this backend dispatching mechanis
 m was recently developed. In this talk\, we will unveil this dispatching m
 echanism and its implementation details\, and how we can use it just by sp
 ecifying a `backend` kwarg like this:\n\n    >>> nx.betweenness_centrality
 (G\, backend=“parallel”)\n\nor by passing the backend graph object(typ
 e-based dispatching):\n\n    >>> H = nxp.ParallelGraph(G)\n    >>> nx.betw
 eenness_centrality(H)\n\nWe'll also go over the limitations of this dispat
 ch mechanism. Then we’ll use the example of nx-parallel as a guide to bu
 ilding our own custom NetworkX backend. And then\, using NetworkX's existi
 ng test suite\, we'll test this backend that we build. Ending with a quick
  dive into the details of the nx-parallel backend.
LOCATION:Room 6
URL:https://pretalx.com/euroscipy-2024/talk/QLVBYY/
END:VEVENT
BEGIN:VEVENT
SUMMARY:Dispatching\, Backend Selection\, and Compatibility APIs - Guillau
 me Lemaitre\, Joris Van den Bossche\, Tim Head\, Erik Welch\, Marco Gorell
 i\, Sebastian Berg\, Aditi Juneja\, Stéfan van der Walt
DTSTART;TZID=Europe/Berlin:20240829T132000
DTEND;TZID=Europe/Berlin:20240829T150000
DTSTAMP:20260907T151923Z
UID:pretalx-euroscipy-2024-8MXPRW@pretalx.com
DESCRIPTION:Scientific python libraries struggle with the existence of sev
 eral array and dataframe providers.  Many important libraries currently ma
 inly support NumPy arrays or pandas dataframes.\nHowever\, as library auth
 ors we wish to allow users to smoothly use other array provides and simpli
 fy for example the use of GPUs without the need for explicit use of cuda e
 nabled libraries.\n\nThis session will be split into three related discuss
 ions around efforts to tackle this situation:\n* Dispatching and backend s
 election discussion\n* Array API adoption progress and discussion\n* Dataf
 rame compatibility layer discussion
LOCATION:Room 5
URL:https://pretalx.com/euroscipy-2024/talk/8MXPRW/
END:VEVENT
END:VCALENDAR