Stream: IG creation
Topic: FHIR Validator in VSCode
Eirik Myklebost (Oct 10 2021 at 15:34):
This might be useful for those of you using VSCode during IG creation: You can create a custom-task in tasks.json that runs the fhir-validator on your resources and reports the output as problems with highlighting in vscode using a custom problemMatcher:
vscode-validator-example.png
Note: This problemMatcher only works when validating at least 2 resources, because the console-output when validating a single resource differs from multi-resource validation.
tasks.json:
{
"version": "2.0.0",
"tasks": [
{
"label": "Run FHIR Validator",
"type": "shell",
"command": "java",
"args": [
"-jar", "validator_cli.jar",
"${workspaceFolder}/test-resources",
"-version", "4.0",
],
"presentation": {
"clear": true
},
"problemMatcher": [
{
"source": "FHIR Validator",
"fileLocation": "absolute",
"pattern": [
{
"regexp": "^-- (.+) -{5,}$",
"file": 1,
},
{
"regexp": "^.*$",
},
{
"regexp": "^ ((?:Info|Warning|Error)).* \\(line ?(\\d+), col ?(\\d+)\\) : (.+)$",
"severity": 1,
"line": 2,
//"column": 3,
"message": 4,
"loop": true
}
]
}
]
}
]
}
Grahame Grieve (Oct 11 2021 at 01:08):
This problemMatcher only works when validating at least 2 resources, because the console-output when validating a single resource differs from multi-resource validation
umm. really? I would've thought it would be the other way around?
Eirik Myklebost (Oct 11 2021 at 07:59):
Grahame Grieve said:
This problemMatcher only works when validating at least 2 resources, because the console-output when validating a single resource differs from multi-resource validation
umm. really? I would've thought it would be the other way around?
Output from single resource validation:
Validating
Validate fsh-generated\resources\Questionnaire-NAV-06-03-04.json ..Detect format for fsh-generated\resources\Questionnaire-NAV-06-03-04.json
00:01.0003
Done. Times: Loading: 00:12.0465, validation: 00:01.0004. Memory = 398Mb
Success: 0 errors, 5 warnings, 2 notes
Information @ Questionnaire.item[0].answerValueSet (line 21, col68) : Canonical URL 'http://fhir.nav.no/ValueSet/icd10-vs-nr2' does not resolve
Output from multi resource validation:
Validating
Validate fsh-generated\resources\Patient.json ..Detect format for fsh-generated\resources\Patient.json
00:00.0083
Validate fsh-generated\resources\Questionnaire-NAV-06-03-04.json ..Detect format for fsh-generated\resources\Questionnaire-NAV-06-03-04.json
00:00.0836
Done. Times: Loading: 00:10.0786, validation: 00:00.0923 (#2). Memory = 353Mb
-- fsh-generated\resources\Patient.json --------------------------------------------
Success: 0 errors, 0 warnings, 1 notes
Information @ ?? : All OK
------------------------------------------------------------------------------------
-- fsh-generated\resources\Questionnaire-NAV-06-03-04.json ---------------------------------------------------------------
Success: 0 errors, 5 warnings, 2 notes
Information @ Questionnaire.item[0].answerValueSet (line 21, col68) : Canonical URL 'http://fhir.nav.no/ValueSet/icd10-vs-nr2' does not resolve
The problemMatcher uses regex and looks for -- {filepath} -----, which is excluded from the single-resource validation output.
Eirik Myklebost (Oct 12 2021 at 15:34):
@Grahame Grieve , is it possible to change the log format, or add support for different styles? Using an already established format will make it easier to integrate the validator in IDEs and tools, e.g. as a linter. ESlint's compact or stylish format are some alternatives.
Grahame Grieve (Oct 12 2021 at 18:28):
it's possible, sure. I'd accept a PR or you can create a task. I'
Grahame Grieve (Oct 12 2021 at 18:28):
I'm not going to change the default behaviour but we'd sure support an -output-style paramter
Eirik Myklebost (Oct 12 2021 at 21:46):
Alright, will look into it. :+1:
Grahame Grieve (Oct 21 2021 at 01:56):
ok so I added a parameter -output-style and a possible value eslint-compact. And it's now much easier to add any kind of other output style because I refactored the code to make it easy
Grahame Grieve (Oct 21 2021 at 01:56):
@Eirik Myklebost should be in today's release
Last updated: Apr 12 2022 at 19:14 UTC
