MINUTES, 2009.01.13

Present: Chris, Istvan, Jenny, Marek, Namshin


1.    0.8 status

Close to at least tagging an alpha. Some most recent changes (i.e.
issue 34) haven't had corresponding documentation updated yet. Request
from Chris regarding practical testing of his recent Blast-related
changes.

Last major issue before alpha: short-term changes to Pygr.Data. Jenny
expressed concern about excessive length of names after change - not a
problem if namespaces are just a convention, not forced. Regarding
timing, Jenny and Istvan suggest considering postponing this until 0.9
(long-term changes are scheduled for 1.0), Chris however would like to
at least introduce new interfaces in 0.8 so that they do not change any
more. Note that all changes will be backward-compatible.


2.    Testing etc.

General overview:
 - megatests - running, set-up documented, MS generalising scripts;
 - small tests - running but bench not documented; MS will ask Titus to
document this;
 - building, compatibility - Titus's student was supposed to work on
that, MS will ask T. for a progress report.

Istvan asked about megatest requirements. Namshin: memory < 2 GB, disc
space c.a. 200 GB, running time: short version about 1 hour on leelab2,
long originally estimated at about 6 h but the number is uncertain
(Namshin will test this).

Namshin: Required resource reorganisation on UCLA megatest bench! MS
will talk to Chris, summarise resources, get things ready.

Chris requested detailed timing information to be added to megatests so
that performance changes can be tracked. Namshin will implement them.


2a.    Test framework rewrite by Istvan

Almost everything in testing framework transferred to new structure. To
do: blast code, megatests; Istvan needs a few more days to finish this.
After that, changes can be reviewed and possibly incorporated into
master. Note from Chris: we should have Titus around for that.

Istvan: In addition to addressing Windows incompatibilities the rewrite
makes it easier for developers to add their own test (framework more
simple), also high modularity (tests can now be run one by one, i.e.
directly from the editor).

Chris: what tests can be skipped and not treated as errors? Istvan: at
present only SQL tests are skippable (this includes some non-Pygr
errors such as permission issues).


2b.    Pygr message clean-up

Proposal from Istvan: replace all prints with calls to the logger
module! See pygrlogger.py (sp?) in his repository on github for usage
example. Istvan will open an issue on the tracker regarding this as
well as work on migrating most of the current code. Note: once the
module has been imported and formatters for different log levels have
been defined, actual message-related changes can be made gradually.

Request: all devs please get in habit of using logger instead of print!

Question of verbosity: presently long messages by default which can be
suppressed. Istvan: can cause problems because in some cases it's
all-or-nothing. Suggestion: use short error messages with links to long
explanations. Will add an issue to tracker.

Jenny will propagate those changes to PyEnsembl.


2c.    Misc. testing

Issue 33: can be marked closed but needs testing. Chris has written an
order-by test for GraphView but not for this. Jenny will follow up.


3.    Code Review

First review will happen on SeqDB. Need Titus for that! Then subdivide
modules between people. MS: how hard will it be to test all modules,
esp. in case of devs not that familiar with them? Results from Titus
may or may not shed light on that.

Will likely tag alpha before the review, final version will wait for
finish.

Schedule needed!
Volunteers requested for modules they would like to review.
MS: follow up with Titus.


4.    Documentation

Chris will upgrade most out-of-code documentation. Issue: things like
Pygr Cookbook need to be checked AND moved to new location - most
likely committed to the Git repository as doctests, then generate HTML
from those doctests during testing runs.

Chris: how many Cookbook recipes are too large to become doctests?
Namshin thinks only about 5 are large.

Question of code examples posted on the Wiki. These should probably end
up as doctests too, with Wiki devoted to general issues and non-code
notes.

Lack of direct Git support in Google Code a hindrance, hopefully this
will be addressed by Google relatively soon.

Documentation/testing task force: Istvan, MS, Namshin, Titus


5.    Getting people involved

 - use tracker a lot, including posting progress reports on issues
 - general call to pygr-dev: if you see an issue you feel like being
able to fix, go ahead!
 - pull in more students!
 - after 0.8 release: Advertise Pygr to sites with Python recipes,
git.or.cz list of projects, freshmeat, ... Long-term: write some
articles and paper(s) about this. Concentrate on others can't do.

Istvan: data from new high-throughput sequencers presents an
opportunity for Pygr, we should think about detailed work-flow
documentation aimed at such audience.


6.    Next meeting

Tuesdays 17:00 Pacific fine for everyone. Tentative date of next
meeting: 2009.01.27.


-- 
MS

--~--~---------~--~----~------------~-------~--~----~
You received this message because you are subscribed to the Google Groups 
"pygr-dev" group.
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/pygr-dev?hl=en
-~----------~----~----~----~------~----~------~--~---

Reply via email to