Normalize #75
Normalize #75
Conversation
| @@ -0,0 +1,170 @@ | |||
| { | |||
There was a problem hiding this comment.
Should this file be names something like 01_prenormalized_message.json instead of premerged?
| @@ -0,0 +1,245 @@ | |||
| { "workflow": null, | |||
There was a problem hiding this comment.
Same with the name of this file
operations/normalize_nodes.yml
Outdated
| description: This operation updates the identifiers on qgraph and kgraph nodes to their preferred identifiers, and adds equivalent identifiers in a property for knodes. When two kgraph nodes normalize to the same preferred identifier, the two knodes are merged. The new node contain the union of the properties of the two original nodes. All edges attached to either of the two original nodes are now subsequently attached to the new merged knode. Qnodes are not merged, so that the structure of the query can be preserved. The updates to kgraph node identifiers also necessitates the updating of result node bindings. | ||
| examples: | ||
| - input: normalize_nodes/messages/01_premerged_message.json | ||
| output: normalize_nodes/messages/02_postmerged_message.json |
There was a problem hiding this comment.
Should there be a hyphen here?
| @@ -0,0 +1,15 @@ | |||
| id: normalize_nodes | |||
There was a problem hiding this comment.
Let's make this a non-unique operation (so there is only one way to do the normalization)
|
Thanks @dkoslicki . I've updated the filenames / dash. I'm not sure how to designate unique/nonunique. I took a look at sort and some others that I thought should be nonunique and didn't see any kind of annotation there. Can you point me in the right direction? |
|
Here's an example of the |
|
@dkoslicki is right. This should be tagged as Here's an example: https://github.com/NCATSTranslator/OperationsAndWorkflows/blob/main/operations/bind.yml |
|
Aha, explains why I didn't see unique in the places I was looking. I went ahead and added it explicitly to this one. |
|
That shouldn't be a problem. However, before this is merged, the schema needs to be rebuilt first. This is done through |
|
Oh right. I've also updated the README with a reminder on the process. |
Define the normalize nodes operation