Where communities thrive


  • Join over 1.5M+ people
  • Join over 100K+ communities
  • Free without limits
  • Create your own community
People
Repo info
Activity
  • 13:19
    eliagbayani synchronize #158
  • 13:09
    eliagbayani synchronize #158
  • 09:15
    eliagbayani synchronize #158
  • Jan 26 15:10
    eliagbayani synchronize #158
  • Jan 25 03:05
    eliagbayani synchronize #158
  • Jan 25 03:03
    eliagbayani synchronize #158
  • Jan 25 03:01
    eliagbayani synchronize #158
  • Jan 25 02:53
    eliagbayani synchronize #158
  • Jan 25 02:48
    eliagbayani synchronize #158
  • Jan 25 01:11
    eliagbayani synchronize #158
  • Jan 25 00:33
    eliagbayani synchronize #158
  • Jan 24 16:36
    eliagbayani synchronize #158
  • Jan 24 16:29
    eliagbayani synchronize #158
  • Jan 24 16:22
    eliagbayani synchronize #158
  • Jan 24 16:09
    eliagbayani synchronize #158
  • Jan 24 15:50
    eliagbayani synchronize #158
  • Jan 24 15:44
    eliagbayani synchronize #158
  • Jan 24 15:35
    eliagbayani synchronize #158
  • Jan 24 15:34
    eliagbayani synchronize #158
  • Jan 24 15:24
    eliagbayani synchronize #158
Jeremy Rice
@JRice
Command ... just ... worked. SO I can only speculate at why it failed when it did. :\ Prrrrrobably a network drive error.
Those should be quite rare (and if we had a real devops position, might be solvable), so I assume re-trying will fix.
...Unlocked the resource. You can try again.
("resume" should work.)
Jen Hammock
@jhammock
Thanks, it’s running now
Jeremy Rice
@JRice
One of the perils of having a child, particularly when one has a weak immune system: I am sick. Again. :| I really need to take the day today. I will check in as often as I can, though. Sorry! :S
Jen Hammock
@jhammock
roger that, nothing urgent here
@mvitale GBIF is published in beta and your GBIF search link works beautifully. One issue remains: it does show for >60 results, when we decided it should be hidden
Jorrit Poelen
@jhpoelen
@hyanwong crazy that Mick Elliot is affiliated with my alma mater, https://rug.nl . Great place to write fast newick parsers ; ) Did you try contacting him?
Jeremy Rice
@JRice
Sorry, I'm just feeling worse today; another day out. :(
Yan Wong
@hyanwong
@jhpoelen I haven't yet. I could do and copy you in?
(apologies for the off-topic conversation!)
Jeremy Rice
@JRice
(Aaaaaand, sorry, I have to take another day off today.) :(
Michael Vitale
@mvitale
@jhammock I see what the problem is with the GBIF page limit. I’m realizing it would be a lot easier to make the limit 50, since that’s all we fetch at a given time for a page of results. Would that be ok?
Jen Hammock
@jhammock
Yes, thanks
Jorrit Poelen
@jhpoelen
@hyanwong I hope you'll bring your favorite newick parser back to the surface! No need to copy me.
Jeremy Rice
@JRice
Apologies, looks like I'm making this a full week off. :| If there are high-priority tasks that I'm missing, though, please do let me know.
Jen Hammock
@jhammock
No, nothing that can't wait, thanks for asking. Keep drinking fluids!
Jorrit Poelen
@jhpoelen
@jhammock hey y'all traitbankers . . . do you have any favored format for documenting taxon-level traits (e.g., body mass, habitat, length) these days? Do you have examples of your favorite trait dataset format?
and @JRice hope you feel better soon . . .
Jen Hammock
@jhammock
we like anything structured, with metadata attached at the records level. URIs are a bonus, but pretty rare. Why?
Jorrit Poelen
@jhpoelen
@jhammock Just looking for existing examples I can point to.
would it look something like: | taxon name | trait name | trait value | trait unit | citation | ?
e.g., | Homo sapiens | habitat | kitchen table | n/a | Jorrit Poelen pers. comm. 2020
Jen Hammock
@jhammock
Sure, that would work, but… infinite diversity in infinite combinations...
Jorrit Poelen
@jhpoelen
ok, how about | taxon name | taxon name id | trait name | trait id | trait value | trait value id | trait unit | trait unit id | citation ?
| Homo sapiens | NCBI:9606 | habitat | ENVO:XXXX | kitchen table | ENVO:1234 | null | null | Jorrit Poelen pers. comm. 2020 ?
Jen Hammock
@jhammock
that too. But what I’m trying to say is that not all datasets will fit in that format. Mostly (for us) because of metadata, including provenance. It’s not always the same shape
Jorrit Poelen
@jhpoelen
do you have a favorite format?
Jen Hammock
@jhammock
(for which the generic solution is Add Columns As Needed)
nope :)
Jorrit Poelen
@jhpoelen
how about the format that would be least effort to integrate into traitbank ?
Jen Hammock
@jhammock
@eagbayani may have opinions. He does most of the heavy lifting, apart from terms assignment. In our reality, the easiest way for a dataset to become to hard to deal with is if it contains many different terms that require mapping.
Jorrit Poelen
@jhpoelen
Ok makes sense. Last question . . . ; ) Do you have any opinion on binary trait matrices? e.g., columns - name | habitat_kitchen_table | habitat_office | habitat_coffee_shop with rows like | Homo sapiens | 1 | 0 | 0 |
Jen Hammock
@jhammock
I’m not parsing your example, sorry. From what I’ve heard of binary matrices, they work great for certain types of data, but those types are a narrow slice of what we handle
Jorrit Poelen
@jhpoelen
| name | habitat_kitchen_table | habitat_office | habitat_coffee_shop | Homo sapiens | 1 | 0 | 0 |
(oops)
Jen Hammock
@jhammock
Oh, of course- yes, that’s what I think of for “binary matrix"
Jorrit Poelen
@jhpoelen
name habitat_kitchen_table habitat_office habitat_coffeshop habitat_couch
Homo sapiens 1 0 0 0
Felis catus 0 0 0 1
Jen Hammock
@jhammock
:+1:
Jorrit Poelen
@jhpoelen
(it must be Friday afternoon ; ) )
Jen Hammock
@jhammock
naturally
Jorrit Poelen
@jhpoelen
for what kind of data do they work well ?
(in your opinion)
Jen Hammock
@jhammock
if you picture it as a predicate:value pair instead, wherever there aren’t too many possible values per predicate
Jorrit Poelen
@jhpoelen
I don't follow.
Jen Hammock
@jhammock
If there were 20 different possible habitats, listing each and turning it on or off if you only want one seems like a lot of trouble
Or, say, 1500 possible values
Jorrit Poelen
@jhpoelen
Ok, thanks for clarifying.
Jeremy Rice
@JRice
I'm still in recovery, but plan to work today!