Feature request: --outputsarif support

[tusharsadhwani]

Is your feature request related to a problem? Please describe.

SARIF is a well supported output format for linters, there exist integrations (eg. to view issues in a web browser), and I think it is really easy to convert the current json output into SARIF:

Current output

{
    "version": "1.1.337",
    "time": "1701179145050",
    "generalDiagnostics": [
        {
            "file": "/tmp/asd.py",
            "severity": "error",
            "message": "Expression of type \"Literal['42']\" cannot be assigned to return type \"int\"\n  \"Literal['42']\" is incompatible with \"int\"",
            "range": {
                "start": {
                    "line": 1,
                    "character": 11
                },
                "end": {
                    "line": 1,
                    "character": 15
                }
            },
            "rule": "reportGeneralTypeIssues"
        }
    ],
    "summary": {
        "filesAnalyzed": 1,
        "errorCount": 1,
        "warningCount": 0,
        "informationCount": 0,
        "timeInSec": 0.224
    }
}

Equivalent SARIF

{
  "version": "2.1.0",
  "$schema": "https://schemastore.azurewebsites.net/schemas/json/sarif-2.1.0.json",
  "runs": [
    {
      "tool": {
        "driver": {
          "name": "pyright",
          "version": "1.1.337"
        }
      },
      "results": [
        {
          "ruleId": "reportGeneralTypeIssues",
          "level": "error",
          "message": {
            "text": "Expression of type \"Literal['42']\" cannot be assigned to return type \"int\"\n  \"Literal['42']\" is incompatible with \"int\""
          },
          "locations": [
            {
              "physicalLocation": {
                "artifactLocation": {
                  "uri": "/tmp/asd.py"
                },
                "region": {
                  "startLine": 1,
                  "startColumn": 11,
                  "endLine": 1,
                  "endColumn": 15
                }
              }
            }
          ]
        }
      ]
    }
  ]
}

Describe the solution you’d like

An --outputsarif option, or adding a new --output enum flag with text/json/sarif options, whichever is preferrable.

[erictraut]

Pyright already outputs a JSON output option. From this, you can transform to any output format you'd prefer with a short script. We have no plans to support alternative formats at this time.

[VIKTORVAV99]

Is this still the case?

Having SARIF outputs would allow it to plug into the GitHub code scanning tool and allow people to run CI pipelines that only block new type errors (which would allow legacy codebases to enable pyright and gradually fix issues).

Seems like something that should be built into the tool itself, especially since it seems like the current output has all the needed values so no change in how this is collected would be needed?

I would even argue that this could and should be the standard output format in v2.

[dgutson]

Pyright already outputs a JSON output option. From this, you can transform to any output format you'd prefer with a short script. We have no plans to support alternative formats at this time.

yet another "short script" to convert from XXX to a well stablished standard with plenty of tools to support it. This "short script" will require maintenance, running it, downloading, will present yet another attack surface, will require dependencies to be regularly updated...
Not using stablished standards incur in isolation and artificial bridges to the rest of the industry.