Files
compliance-scanner-agent/compliance-agent/src/pipeline/sbom/syft.rs
Sharang Parnerkar 3bb690e5bb
All checks were successful
CI / Format (push) Successful in 4s
CI / Clippy (push) Successful in 4m19s
CI / Security Audit (push) Successful in 1m44s
CI / Tests (push) Successful in 5m15s
CI / Detect Changes (push) Successful in 5s
CI / Deploy Agent (push) Successful in 2s
CI / Deploy Dashboard (push) Successful in 2s
CI / Deploy Docs (push) Has been skipped
CI / Deploy MCP (push) Successful in 2s
refactor: modularize codebase and add 404 unit tests (#13)
2026-03-13 08:03:45 +00:00

356 lines
10 KiB
Rust

use std::path::Path;
use compliance_core::models::SbomEntry;
use compliance_core::CoreError;
#[tracing::instrument(skip_all, fields(repo_id = %repo_id))]
pub(super) async fn run_syft(repo_path: &Path, repo_id: &str) -> Result<Vec<SbomEntry>, CoreError> {
let output = tokio::process::Command::new("syft")
.arg(repo_path)
.args(["-o", "cyclonedx-json"])
// Enable remote license lookups for all ecosystems
.env("SYFT_GOLANG_SEARCH_REMOTE_LICENSES", "true")
.env("SYFT_JAVASCRIPT_SEARCH_REMOTE_LICENSES", "true")
.env("SYFT_PYTHON_SEARCH_REMOTE_LICENSES", "true")
.env("SYFT_JAVA_USE_NETWORK", "true")
.output()
.await
.map_err(|e| CoreError::Scanner {
scanner: "syft".to_string(),
source: Box::new(e),
})?;
if !output.status.success() {
let stderr = String::from_utf8_lossy(&output.stderr);
return Err(CoreError::Scanner {
scanner: "syft".to_string(),
source: format!("syft exited with {}: {stderr}", output.status).into(),
});
}
let cdx: CycloneDxBom = serde_json::from_slice(&output.stdout)?;
let entries = cdx
.components
.unwrap_or_default()
.into_iter()
.map(|c| {
let package_manager = c
.purl
.as_deref()
.and_then(extract_ecosystem_from_purl)
.unwrap_or_else(|| "unknown".to_string());
let mut entry = SbomEntry::new(
repo_id.to_string(),
c.name,
c.version.unwrap_or_else(|| "unknown".to_string()),
package_manager,
);
entry.purl = c.purl;
entry.license = c.licenses.and_then(|ls| extract_license(&ls));
entry
})
.collect();
Ok(entries)
}
// CycloneDX JSON types
#[derive(serde::Deserialize)]
struct CycloneDxBom {
components: Option<Vec<CdxComponent>>,
}
#[derive(serde::Deserialize)]
struct CdxComponent {
name: String,
version: Option<String>,
#[serde(rename = "type")]
#[allow(dead_code)]
component_type: Option<String>,
purl: Option<String>,
licenses: Option<Vec<CdxLicenseWrapper>>,
}
#[derive(serde::Deserialize)]
struct CdxLicenseWrapper {
license: Option<CdxLicense>,
/// SPDX license expression (e.g. "MIT OR Apache-2.0")
expression: Option<String>,
}
#[derive(serde::Deserialize)]
struct CdxLicense {
id: Option<String>,
name: Option<String>,
}
/// Extract the best license string from CycloneDX license entries.
/// Handles three formats: expression ("MIT OR Apache-2.0"), license.id ("MIT"), license.name ("MIT License").
fn extract_license(entries: &[CdxLicenseWrapper]) -> Option<String> {
// First pass: look for SPDX expressions (most precise for dual-licensed packages)
for entry in entries {
if let Some(ref expr) = entry.expression {
if !expr.is_empty() {
return Some(expr.clone());
}
}
}
// Second pass: collect license.id or license.name from all entries
let parts: Vec<String> = entries
.iter()
.filter_map(|e| {
e.license.as_ref().and_then(|lic| {
lic.id
.clone()
.or_else(|| lic.name.clone())
.filter(|s| !s.is_empty())
})
})
.collect();
if parts.is_empty() {
return None;
}
Some(parts.join(" OR "))
}
/// Extract the ecosystem/package-manager from a PURL string.
/// e.g. "pkg:npm/lodash@4.17.21" -> "npm", "pkg:cargo/serde@1.0" -> "cargo"
fn extract_ecosystem_from_purl(purl: &str) -> Option<String> {
let rest = purl.strip_prefix("pkg:")?;
let ecosystem = rest.split('/').next()?;
if ecosystem.is_empty() {
return None;
}
// Normalise common PURL types to user-friendly names
let normalised = match ecosystem {
"golang" => "go",
"pypi" => "pip",
_ => ecosystem,
};
Some(normalised.to_string())
}
#[cfg(test)]
mod tests {
use super::*;
// --- extract_ecosystem_from_purl tests ---
#[test]
fn purl_npm() {
assert_eq!(
extract_ecosystem_from_purl("pkg:npm/lodash@4.17.21"),
Some("npm".to_string())
);
}
#[test]
fn purl_cargo() {
assert_eq!(
extract_ecosystem_from_purl("pkg:cargo/serde@1.0.197"),
Some("cargo".to_string())
);
}
#[test]
fn purl_golang_normalised() {
assert_eq!(
extract_ecosystem_from_purl("pkg:golang/github.com/gin-gonic/gin@1.9.1"),
Some("go".to_string())
);
}
#[test]
fn purl_pypi_normalised() {
assert_eq!(
extract_ecosystem_from_purl("pkg:pypi/requests@2.31.0"),
Some("pip".to_string())
);
}
#[test]
fn purl_maven() {
assert_eq!(
extract_ecosystem_from_purl("pkg:maven/org.apache.commons/commons-lang3@3.14.0"),
Some("maven".to_string())
);
}
#[test]
fn purl_missing_prefix() {
assert_eq!(extract_ecosystem_from_purl("npm/lodash@4.17.21"), None);
}
#[test]
fn purl_empty_ecosystem() {
assert_eq!(extract_ecosystem_from_purl("pkg:/lodash@4.17.21"), None);
}
#[test]
fn purl_empty_string() {
assert_eq!(extract_ecosystem_from_purl(""), None);
}
#[test]
fn purl_just_prefix() {
assert_eq!(extract_ecosystem_from_purl("pkg:"), None);
}
// --- extract_license tests ---
#[test]
fn license_from_expression() {
let entries = vec![CdxLicenseWrapper {
license: None,
expression: Some("MIT OR Apache-2.0".to_string()),
}];
assert_eq!(
extract_license(&entries),
Some("MIT OR Apache-2.0".to_string())
);
}
#[test]
fn license_from_id() {
let entries = vec![CdxLicenseWrapper {
license: Some(CdxLicense {
id: Some("MIT".to_string()),
name: None,
}),
expression: None,
}];
assert_eq!(extract_license(&entries), Some("MIT".to_string()));
}
#[test]
fn license_from_name_fallback() {
let entries = vec![CdxLicenseWrapper {
license: Some(CdxLicense {
id: None,
name: Some("MIT License".to_string()),
}),
expression: None,
}];
assert_eq!(extract_license(&entries), Some("MIT License".to_string()));
}
#[test]
fn license_expression_preferred_over_id() {
let entries = vec![
CdxLicenseWrapper {
license: Some(CdxLicense {
id: Some("MIT".to_string()),
name: None,
}),
expression: None,
},
CdxLicenseWrapper {
license: None,
expression: Some("MIT AND Apache-2.0".to_string()),
},
];
// Expression should be preferred (first pass finds it)
assert_eq!(
extract_license(&entries),
Some("MIT AND Apache-2.0".to_string())
);
}
#[test]
fn license_multiple_ids_joined() {
let entries = vec![
CdxLicenseWrapper {
license: Some(CdxLicense {
id: Some("MIT".to_string()),
name: None,
}),
expression: None,
},
CdxLicenseWrapper {
license: Some(CdxLicense {
id: Some("Apache-2.0".to_string()),
name: None,
}),
expression: None,
},
];
assert_eq!(
extract_license(&entries),
Some("MIT OR Apache-2.0".to_string())
);
}
#[test]
fn license_empty_entries() {
let entries: Vec<CdxLicenseWrapper> = vec![];
assert_eq!(extract_license(&entries), None);
}
#[test]
fn license_all_empty_strings() {
let entries = vec![CdxLicenseWrapper {
license: Some(CdxLicense {
id: Some(String::new()),
name: Some(String::new()),
}),
expression: Some(String::new()),
}];
assert_eq!(extract_license(&entries), None);
}
#[test]
fn license_none_fields() {
let entries = vec![CdxLicenseWrapper {
license: None,
expression: None,
}];
assert_eq!(extract_license(&entries), None);
}
// --- CycloneDX deserialization tests ---
#[test]
fn deserialize_cyclonedx_bom() {
let json = r#"{
"components": [
{
"name": "serde",
"version": "1.0.197",
"type": "library",
"purl": "pkg:cargo/serde@1.0.197",
"licenses": [
{"expression": "MIT OR Apache-2.0"}
]
}
]
}"#;
let bom: CycloneDxBom = serde_json::from_str(json).unwrap();
let components = bom.components.unwrap();
assert_eq!(components.len(), 1);
assert_eq!(components[0].name, "serde");
assert_eq!(components[0].version, Some("1.0.197".to_string()));
assert_eq!(
components[0].purl,
Some("pkg:cargo/serde@1.0.197".to_string())
);
}
#[test]
fn deserialize_cyclonedx_no_components() {
let json = r#"{}"#;
let bom: CycloneDxBom = serde_json::from_str(json).unwrap();
assert!(bom.components.is_none());
}
#[test]
fn deserialize_cyclonedx_minimal_component() {
let json = r#"{"components": [{"name": "foo"}]}"#;
let bom: CycloneDxBom = serde_json::from_str(json).unwrap();
let c = &bom.components.unwrap()[0];
assert_eq!(c.name, "foo");
assert!(c.version.is_none());
assert!(c.purl.is_none());
assert!(c.licenses.is_none());
}
}