TRAPI 1.5: new query property `bypass_cache`
colleenXu opened this issue · comments
TRAPI 1.5 introduces a new optional property in queries (same level as message/log_level/workflow/submitter), bypass_cache (lines 292-299, 351-358). It can be added to "sync" (/query) and async queries.
If this property is present and set to true
, then BTE shouldn't use cached data. Also, BTE should add bypass_cache: true
to its queries to TRAPI KPs, according to the implementation note in changelog
- Note: got clarification that this property can't/shouldn't be passed to NodeNorm (Translator Slack link)
Example TRAPI query with `bypass_cache: true`
{
"bypass_cache": true,
"message": {
"query_graph": {
"nodes": {
"SN": {
"categories":["biolink:ChemicalEntity"]
},
"ON": {
"ids":["MONDO:0001119"],
"categories":["biolink:Disease"]
}
},
"edges": {
"e01": {
"subject": "SN",
"object": "ON",
"predicates": ["biolink:treats"],
"knowledge_type": "inferred"
}
}
}
}
}
May be related to old feature, where caching
is a parameter in queries #224
Looks good.
Based on quick testing, this works as-intended: I ran queries 3 times. 1st to cache, 2nd to see if it uses the cache (it does), 3rd with bypass_cache to see if it correctly doesn't use the cache (yep). I tested a basic query (not-creative-mode) and a creative-mode query.
The code was deployed today to Prod as part of the Octopus release. I tested and it's live. I used the same method of 3 queries, seeing if I can toggle the value of bypass_cache to use the cache or not.