Feature: Add SARIF output support (#9078)

## Summary

Adds support for sarif v2.1.0 output to cli, usable via the
output-format paramter.

`ruff . --output-format=sarif` 

Includes a few changes I wasn't sure of, namely:
* Adds a few derives for Clone & Copy, which I think could be removed
with a little extra work as well.

## Test Plan

I built and ran this against several large open source projects and
verified that the output sarif was valid, using [Microsoft's SARIF
validator tool](https://sarifweb.azurewebsites.net/Validation)

I've also attached an output of the sarif generated by this version of
ruff on the main branch of django at commit: b287af5dc9

[django_main_b287af5dc9_sarif.json](https://github.com/astral-sh/ruff/files/13626222/django_main_b287af5dc9_sarif.json)

Note: this needs to be regenerated with the latest changes and
confirmed.


## Open Points
[ ] Convert to just using all Rules all the time
[ ] Fix the issue with getting the file URI when compiling for web
assembly

crates/ruff_cli/src/printer.rs

@@ -13,7 +13,7 @@ use ruff_linter::fs::relativize_path;
 use ruff_linter::logging::LogLevel;
 use ruff_linter::message::{
     AzureEmitter, Emitter, EmitterContext, GithubEmitter, GitlabEmitter, GroupedEmitter,
-    JsonEmitter, JsonLinesEmitter, JunitEmitter, PylintEmitter, TextEmitter,
+    JsonEmitter, JsonLinesEmitter, JunitEmitter, PylintEmitter, SarifEmitter, TextEmitter,
 };
 use ruff_linter::notify_user;
 use ruff_linter::registry::{AsRule, Rule};
@@ -291,6 +291,9 @@ impl Printer {
             SerializationFormat::Azure => {
                 AzureEmitter.emit(writer, &diagnostics.messages, &context)?;
             }
+            SerializationFormat::Sarif => {
+                SarifEmitter.emit(writer, &diagnostics.messages, &context)?;
+            }
         }
 
         writer.flush()?;

crates/ruff_linter/Cargo.toml

@@ -71,6 +71,7 @@ toml = { workspace = true }
 typed-arena = { version = "2.0.2" }
 unicode-width = { workspace = true }
 unicode_names2 = { workspace = true }
+url = { version = "2.2.2" }
 wsl = { version = "0.1.0" }
 
 [dev-dependencies]

crates/ruff_linter/src/message/mod.rs

@@ -17,6 +17,7 @@ use ruff_diagnostics::{Diagnostic, DiagnosticKind, Fix};
 use ruff_notebook::NotebookIndex;
 use ruff_source_file::{SourceFile, SourceLocation};
 use ruff_text_size::{Ranged, TextRange, TextSize};
+pub use sarif::SarifEmitter;
 pub use text::TextEmitter;
 
 mod azure;
@@ -28,6 +29,7 @@ mod json;
 mod json_lines;
 mod junit;
 mod pylint;
+mod sarif;
 mod text;
 
 #[derive(Debug, PartialEq, Eq)]

crates/ruff_linter/src/message/sarif.rs

@@ -0,0 +1,212 @@
+use std::io::Write;
+
+use anyhow::Result;
+use serde::{Serialize, Serializer};
+use serde_json::json;
+
+use ruff_source_file::OneIndexed;
+
+use crate::codes::Rule;
+use crate::fs::normalize_path;
+use crate::message::{Emitter, EmitterContext, Message};
+use crate::registry::{AsRule, Linter, RuleNamespace};
+use crate::VERSION;
+
+use strum::IntoEnumIterator;
+
+pub struct SarifEmitter;
+
+impl Emitter for SarifEmitter {
+    fn emit(
+        &mut self,
+        writer: &mut dyn Write,
+        messages: &[Message],
+        _context: &EmitterContext,
+    ) -> Result<()> {
+        let results = messages
+            .iter()
+            .map(SarifResult::from_message)
+            .collect::<Result<Vec<_>>>()?;
+
+        let output = json!({
+            "$schema": "https://json.schemastore.org/sarif-2.1.0.json",
+            "version": "2.1.0",
+            "runs": [{
+                "tool": {
+                    "driver": {
+                        "name": "ruff",
+                        "informationUri": "https://github.com/astral-sh/ruff",
+                        "rules": Rule::iter().map(SarifRule::from).collect::<Vec<_>>(),
+                        "version": VERSION.to_string(),
+                    }
+                },
+                "results": results,
+            }],
+        });
+        serde_json::to_writer_pretty(writer, &output)?;
+        Ok(())
+    }
+}
+
+#[derive(Debug, Clone)]
+struct SarifRule<'a> {
+    name: &'a str,
+    code: String,
+    linter: &'a str,
+    summary: &'a str,
+    explanation: Option<&'a str>,
+    url: Option<String>,
+}
+
+impl From<Rule> for SarifRule<'_> {
+    fn from(rule: Rule) -> Self {
+        let code = rule.noqa_code().to_string();
+        let (linter, _) = Linter::parse_code(&code).unwrap();
+        Self {
+            name: rule.into(),
+            code,
+            linter: linter.name(),
+            summary: rule.message_formats()[0],
+            explanation: rule.explanation(),
+            url: rule.url(),
+        }
+    }
+}
+
+impl Serialize for SarifRule<'_> {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        json!({
+            "id": self.code,
+            "shortDescription": {
+                "text": self.summary,
+            },
+            "fullDescription": {
+                "text": self.explanation,
+            },
+            "help": {
+                "text": self.summary,
+            },
+            "helpUri": self.url,
+            "properties": {
+                "id": self.code,
+                "kind": self.linter,
+                "name": self.name,
+                "problem.severity": "error".to_string(),
+            },
+        })
+        .serialize(serializer)
+    }
+}
+
+#[derive(Debug)]
+struct SarifResult {
+    rule: Rule,
+    level: String,
+    message: String,
+    uri: String,
+    start_line: OneIndexed,
+    start_column: OneIndexed,
+    end_line: OneIndexed,
+    end_column: OneIndexed,
+}
+
+impl SarifResult {
+    #[cfg(not(target_arch = "wasm32"))]
+    fn from_message(message: &Message) -> Result<Self> {
+        let start_location = message.compute_start_location();
+        let end_location = message.compute_end_location();
+        let path = normalize_path(message.filename());
+        Ok(Self {
+            rule: message.kind.rule(),
+            level: "error".to_string(),
+            message: message.kind.name.clone(),
+            uri: url::Url::from_file_path(&path)
+                .map_err(|()| anyhow::anyhow!("Failed to convert path to URL: {}", path.display()))?
+                .to_string(),
+            start_line: start_location.row,
+            start_column: start_location.column,
+            end_line: end_location.row,
+            end_column: end_location.column,
+        })
+    }
+
+    #[cfg(target_arch = "wasm32")]
+    #[allow(clippy::unnecessary_wraps)]
+    fn from_message(message: &Message) -> Result<Self> {
+        let start_location = message.compute_start_location();
+        let end_location = message.compute_end_location();
+        let path = normalize_path(message.filename());
+        Ok(Self {
+            rule: message.kind.rule(),
+            level: "error".to_string(),
+            message: message.kind.name.clone(),
+            uri: path.display().to_string(),
+            start_line: start_location.row,
+            start_column: start_location.column,
+            end_line: end_location.row,
+            end_column: end_location.column,
+        })
+    }
+}
+
+impl Serialize for SarifResult {
+    fn serialize<S>(&self, serializer: S) -> Result<S::Ok, S::Error>
+    where
+        S: Serializer,
+    {
+        json!({
+            "level": self.level,
+            "message": {
+                "text": self.message,
+            },
+            "locations": [{
+                "physicalLocation": {
+                    "artifactLocation": {
+                        "uri": self.uri,
+                    },
+                    "region": {
+                        "startLine": self.start_line,
+                        "startColumn": self.start_column,
+                        "endLine": self.end_line,
+                        "endColumn": self.end_column,
+                    }
+                }
+            }],
+            "ruleId": self.rule.noqa_code().to_string(),
+        })
+        .serialize(serializer)
+    }
+}
+
+#[cfg(test)]
+mod tests {
+
+    use crate::message::tests::{capture_emitter_output, create_messages};
+    use crate::message::SarifEmitter;
+
+    fn get_output() -> String {
+        let mut emitter = SarifEmitter {};
+        capture_emitter_output(&mut emitter, &create_messages())
+    }
+
+    #[test]
+    fn valid_json() {
+        let content = get_output();
+        serde_json::from_str::<serde_json::Value>(&content).unwrap();
+    }
+
+    #[test]
+    fn test_results() {
+        let content = get_output();
+        let sarif = serde_json::from_str::<serde_json::Value>(content.as_str()).unwrap();
+        let rules = sarif["runs"][0]["tool"]["driver"]["rules"]
+            .as_array()
+            .unwrap();
+        let results = sarif["runs"][0]["results"].as_array().unwrap();
+        assert_eq!(results.len(), 3);
+        assert!(rules.len() > 3);
+    }
+}

crates/ruff_linter/src/settings/types.rs

@@ -423,6 +423,7 @@ pub enum SerializationFormat {
     Gitlab,
     Pylint,
     Azure,
+    Sarif,
 }
 
 impl Default for SerializationFormat {

docs/configuration.md

@@ -481,7 +481,7 @@ Options:
       --ignore-noqa
           Ignore any `# noqa` comments
       --output-format <OUTPUT_FORMAT>
-          Output serialization format for violations [env: RUFF_OUTPUT_FORMAT=] [possible values: text, json, json-lines, junit, grouped, github, gitlab, pylint, azure]
+          Output serialization format for violations [env: RUFF_OUTPUT_FORMAT=] [possible values: text, json, json-lines, junit, grouped, github, gitlab, pylint, azure, sarif]
   -o, --output-file <OUTPUT_FILE>
           Specify file to write the linter output to (default: stdout)
       --target-version <TARGET_VERSION>