compliance-scanner-agent/compliance-core/src/models/graph.rs

use chrono::{DateTime, Utc};
use serde::{Deserialize, Serialize};

/// Type of code node in the knowledge graph
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq, Hash)]
#[serde(rename_all = "snake_case")]
pub enum CodeNodeKind {
    Function,
    Method,
    Class,
    Struct,
    Enum,
    Interface,
    Trait,
    Module,
    File,
}

impl std::fmt::Display for CodeNodeKind {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            Self::Function => write!(f, "function"),
            Self::Method => write!(f, "method"),
            Self::Class => write!(f, "class"),
            Self::Struct => write!(f, "struct"),
            Self::Enum => write!(f, "enum"),
            Self::Interface => write!(f, "interface"),
            Self::Trait => write!(f, "trait"),
            Self::Module => write!(f, "module"),
            Self::File => write!(f, "file"),
        }
    }
}

/// A node in the code knowledge graph
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct CodeNode {
    #[serde(rename = "_id", skip_serializing_if = "Option::is_none")]
    pub id: Option<bson::oid::ObjectId>,
    pub repo_id: String,
    pub graph_build_id: String,
    /// Unique identifier within the graph (e.g., "src/main.rs::main")
    pub qualified_name: String,
    pub name: String,
    pub kind: CodeNodeKind,
    pub file_path: String,
    pub start_line: u32,
    pub end_line: u32,
    /// Language of the source file
    pub language: String,
    /// Community ID from Louvain clustering
    pub community_id: Option<u32>,
    /// Whether this is a public entry point (main, exported fn, HTTP handler, etc.)
    pub is_entry_point: bool,
    /// Internal petgraph node index for fast lookups
    #[serde(skip_serializing_if = "Option::is_none")]
    pub graph_index: Option<u32>,
}

/// Type of relationship between code nodes
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq, Hash)]
#[serde(rename_all = "snake_case")]
pub enum CodeEdgeKind {
    Calls,
    Imports,
    Inherits,
    Implements,
    Contains,
    /// A type reference (e.g., function parameter type, return type)
    TypeRef,
}

impl std::fmt::Display for CodeEdgeKind {
    fn fmt(&self, f: &mut std::fmt::Formatter<'_>) -> std::fmt::Result {
        match self {
            Self::Calls => write!(f, "calls"),
            Self::Imports => write!(f, "imports"),
            Self::Inherits => write!(f, "inherits"),
            Self::Implements => write!(f, "implements"),
            Self::Contains => write!(f, "contains"),
            Self::TypeRef => write!(f, "type_ref"),
        }
    }
}

/// An edge in the code knowledge graph
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct CodeEdge {
    #[serde(rename = "_id", skip_serializing_if = "Option::is_none")]
    pub id: Option<bson::oid::ObjectId>,
    pub repo_id: String,
    pub graph_build_id: String,
    /// Qualified name of source node
    pub source: String,
    /// Qualified name of target node
    pub target: String,
    pub kind: CodeEdgeKind,
    /// File where this relationship was found
    pub file_path: String,
    pub line_number: Option<u32>,
}

/// Status of a graph build operation
#[derive(Debug, Clone, Serialize, Deserialize, PartialEq, Eq)]
#[serde(rename_all = "snake_case")]
pub enum GraphBuildStatus {
    Running,
    Completed,
    Failed,
}

/// Tracks a graph build operation for a repo/commit
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct GraphBuildRun {
    #[serde(rename = "_id", skip_serializing_if = "Option::is_none")]
    pub id: Option<bson::oid::ObjectId>,
    pub repo_id: String,
    pub commit_sha: Option<String>,
    pub status: GraphBuildStatus,
    pub node_count: u32,
    pub edge_count: u32,
    pub community_count: u32,
    pub languages_parsed: Vec<String>,
    pub error_message: Option<String>,
    #[serde(with = "super::serde_helpers::bson_datetime")]
    pub started_at: DateTime<Utc>,
    #[serde(default, with = "super::serde_helpers::opt_bson_datetime")]
    pub completed_at: Option<DateTime<Utc>>,
}

impl GraphBuildRun {
    pub fn new(repo_id: String) -> Self {
        Self {
            id: None,
            repo_id,
            commit_sha: None,
            status: GraphBuildStatus::Running,
            node_count: 0,
            edge_count: 0,
            community_count: 0,
            languages_parsed: Vec::new(),
            error_message: None,
            started_at: Utc::now(),
            completed_at: None,
        }
    }
}

/// Impact analysis result for a finding
#[derive(Debug, Clone, Serialize, Deserialize)]
pub struct ImpactAnalysis {
    #[serde(rename = "_id", skip_serializing_if = "Option::is_none")]
    pub id: Option<bson::oid::ObjectId>,
    pub repo_id: String,
    pub finding_id: String,
    pub graph_build_id: String,
    /// Number of nodes reachable from the finding location
    pub blast_radius: u32,
    /// Entry points affected by this finding (via reverse call chain)
    pub affected_entry_points: Vec<String>,
    /// Call chains from entry points to the finding location
    pub call_chains: Vec<Vec<String>>,
    /// Community IDs affected
    pub affected_communities: Vec<u32>,
    /// Direct callers of the affected function
    pub direct_callers: Vec<String>,
    /// Direct callees of the affected function
    pub direct_callees: Vec<String>,
    #[serde(with = "super::serde_helpers::bson_datetime")]
    pub created_at: DateTime<Utc>,
}

impl ImpactAnalysis {
    pub fn new(repo_id: String, finding_id: String, graph_build_id: String) -> Self {
        Self {
            id: None,
            repo_id,
            finding_id,
            graph_build_id,
            blast_radius: 0,
            affected_entry_points: Vec::new(),
            call_chains: Vec::new(),
            affected_communities: Vec::new(),
            direct_callers: Vec::new(),
            direct_callees: Vec::new(),
            created_at: Utc::now(),
        }
    }
}