Obtaining Existing Reference Nodes #390
Replies: 5 comments 3 replies
-
|
Hi @c-heat16, there are a few questions here, let me see if I can unpack it and address them. If I miss anything please feel free to let me know.
import jsonschema
import requests
def get_db_schema():
json_schema = requests.get(
url="https://lb-stage.mycriptapp.org/api/v1/schema/"
).json()["data"]
return json_schema
def validate_giant_json():
db_schema = get_db_schema()
# this field needs to be overwritten with whatever is needed
node_type: str = "citation"
http_method: str = "Post"
db_schema["$ref"] = f"#/$defs/{node_type.title()}{http_method.title()}"
node = {
"node": [
"Citation"
],
"uid": "_:33d273ac-f63f-4d07-a77d-8ba17e2e3c01",
"uuid": "33d273ac-f63f-4d07-a77d-8ba17e2e3c01",
"type": "extracted_by_human",
"reference": {
"uid": "_:bd4da2ca-f9d5-493c-9f29-7e194ce88698",
"uuid": "uuid()",
"model_version": "1.0.0",
"updated_at": "now()",
"created_at": "now()",
"public": True
},
"model_version": "1.0.0",
"updated_at": "now()",
"created_at": "now()"
}
jsonschema.validate(instance=node, schema=db_schema)
if __name__ == "__main__":
validate_giant_json()I got Failed validating 'anyOf' in schema['properties']['reference']:
{'anyOf': [{'$ref': 'EdgeUUID'},
{'$ref': 'ReferencePost'},
{'$ref': 'Edge'}]}
On instance['reference']:
{'created_at': 'now()',
'model_version': '1.0.0',
'public': True,
'uid': '_:bd4da2ca-f9d5-493c-9f29-7e194ce88698',
'updated_at': 'now()',
'uuid': 'uuid()'}I think the issue is that the reference node is coming out incorrectly in JSON format. The reference node is supposed to look like the JSON given in the documentation. How are you creating the reference node? Is it being created incorrectly after being made in Python or are you getting the reference node from the API and it is being deserialized into a node incorrectly? |
Beta Was this translation helpful? Give feedback.
-
|
Hi @nh916 Not being able to fetch by name makes sense as references do not have a 'name'. Well, I don't need to grab a reference by name, but AFAIK, the only way to obtain the UID of an existing reference is search through the ever-growing list of references on CRIPT - is there an easier way? We have 35 different references associated with just this one project, and additional in future projects, so it would be great if there is a way to programmatically fetch existing references without having to manually comb through the existing references in CRIPT. Thanks for looking in to validating each JSON. Reference nodes are created as shown below. I haven't been able to retrieve references with the API, so I presume it is some issue with reference creation, but I couldn't identify any issues. Where In looking at the documentation, the only potential issue I see is that we don't provide a ISSN. The documentation says an ISSN is conditionally required, but doesn't describe the conditions. It also has a default value of "", so that requirement should be satisfied regardless? Ie it isn't |
Beta Was this translation helpful? Give feedback.
-
|
I think the error may arise from 2+ experiments in the same project using the same reference node. The script below works as intended when only expt1 has the reference and fails if the reference is added to expt2. Is something checking for duplicate references/citations? Seems unnecessary for citations? I also noticed we can't have duplicate process names, which may also be unnecessary? All of our processes are pyrolysis, and naming them Error: |
Beta Was this translation helpful? Give feedback.
-
|
@c-heat16 yeah that makes sense. I reported that issue to the API and asked if we can have an endpoint to grab reference nodes by DOI. |
Beta Was this translation helpful? Give feedback.
-
|
I think that would be great! Thanks! |
Beta Was this translation helpful? Give feedback.
-
|
@c-heat16, regarding the ISSN conditionality, I believe it might depend on the reference type. However, with the recent updates to the vocabulary, I'm not seeing that specification clearly mentioned. I also checked the DB schema and it appears to treat ISSN more as an optional attribute rather than conditional. I've noted this and will work on converting all |
Beta Was this translation helpful? Give feedback.
-
|
@c-heat16, I've looked into the issue you've described where the API returns an error when two experiments use the same reference node. If only one experiment uses the reference, there's no error. I suspect this might be an API-related issue. I've reported this to the API team and am awaiting their feedback. If they confirm there's no error on their end, I'll further investigate by examining the JSON requests and responses through tools like postman to get a deeper understanding and pinpoint the root cause. Regarding using same process names in different experiments, this is a known issue from the API that if the user inputs the same process name into 2 different experiments the API gives an error, I have already reported this in the past and am waiting for it to be fixed. However, CRIPT does want scientists to use different descriptive names for their nodes that helps readers easily understand each step, so for that reason CRIPT tries to stop users from entering the same name for the same node within another node, for example CRIPT does not want users to input the same process name for many processes within an experiment. I hope that makes sense, let me know if you have any questions. |
Beta Was this translation helpful? Give feedback.
-
|
Thank you for providing you script in each issue, that makes it very transparent and easier to track down issues |
Beta Was this translation helpful? Give feedback.
-
Is there a correct way to fetch existing reference nodes? I've looked around but don't see any examples, and trying to fetch a reference as if it were a material is unsuccessful. For example, there is a reference I previously created with the title 'Mechanistic Modeling of Polymer Pyrolysis: Polypropylene'
Trying to retrieve the existing reference as described below, either by name or DOI, yields errors.
Complete CLI output:
We are constructing a repository of previously published results, so having references to the original source is imperative. Creating new reference nodes with the same attributes as existing reference nodes yields a very long, nondescript error message (below). If no references are added to the project being saved, the error goes away.
Beta Was this translation helpful? Give feedback.
All reactions