TreesById missing from API response

[snacktavish]

I think I am forgetting something important, potentially related to Nexson versions.

I want to find out what trees are in any given study, without clicking on them in the curator.

TreesById seems like the place to look
e.g.

phylesystem-1/study/ot_03/ot_1003/ot_1003.json

Line 11700 in c33806c

"treesById": {

But if I get the study from the api

e.g.

curl https://api.opentreeoflife.org/v3/study/ot_1003 > ot_1003_api.json

But "TreesById" does not seem to be there anywhere...

grep treesById ot_1003_api.json

References to the tree I am looking for (Tr100304) seem to only be buried in the annotations. Pointers?

[jimallman]

Yes, looks like two radically different versions of Nexson. Both files show nexml version 0.0.4a, but each was converted with a different version of nexml2json (1.2.1 vs 0.0.0). I'm guessing that explains the radically different JSON (Nexson) structure.

It's easier to examine the JSON structure using the command-line jq:

% jq '.data.nexml."@nexml2json"' ot_1003-nexml2json-0.0.0.json | head -n6
"0.0.0"

You can also search for target text (like your grep example above) and return all paths where it's found:

jq 'paths(. | select(. == "0.0.0"))| join(".")' ot_1003-nexml2json-0.0.0.json
"data.nexml.@nexml2json"
"data.nexml.meta.10.agent.1.@version"

Using this, we can confirm the location of each tree in each format, by searching for its ID:

% jq 'paths(. | select(. == "Tr100304"))| join(".")' ot_1003-nexml2json-1.2.1.json
"nexml.treesById.Tb40460.^ot:treeElementOrder.0"

...and for the other nexson version:

% jq 'paths(. | select(. == "Tr100304"))| join(".")' ot_1003-nexml2json-0.0.0.json
"data.nexml.trees.tree.0.@id"
"data.nexml.meta.11.annotation.2.message.7.refersTo.@idref"
"data.nexml.meta.11.annotation.2.message.7.refersTo.@treeID"
"data.nexml.meta.11.annotation.2.message.8.refersTo.@idref"
"data.nexml.meta.11.annotation.2.message.8.refersTo.@treeID"

The first path returned includes the equivalent to the 'treesById' object at .data.nexml.trees.tree. (In this earlier Nexson version, it's a simple list of objects with the ID stored internally.)

% jq '.data.nexml.trees.tree' ot_1003-nexml2json-0.0.0.json
[
  {
    "@label": "ML phylogram",
    "@xsi:type": "nex:FloatTree",
    "node": [
      {
        "@root": true,
        "@id": "Tn14962993",
...

Hope this helps!

[mtholder]

I think you need to use the output_nexml2json parameter. The documentation is at https://github.com/OpenTreeOfLife/phylesystem-api/blob/master/docs/README.md#get-arguments
I'm not sure why we never decided to move those docs to the general API documentation page

[snacktavish]

Awesome!! Thanks to you both. I have used jq to make json pretty, but never to search it. Both of these pieces of info are super helpful. And I’ll at least add a link to that doc in the API docs.