Add safe redundant hierarchical relations auto-removal

Implement distinction between safe (auto-fixable) and unsafe (manual review)
redundant hierarchical relations in the lint command.

Safe redundant relations are those with a single unambiguous chain
(A→B→C with redundant direct A→C). These can be safely auto-removed
when using `reqvire lint --fix`.

Unsafe redundant relations occur when multiple paths converge to the
same ancestor (A→B→C and A→D→C with redundant direct A→C). These
require manual review as both paths may have semantic meaning.

Implementation:
- Add SafeRedundantHierarchicalRelations variant to AutoFixableIssue
- Implement path counting logic to distinguish single-chain from multi-path
- Split detect_maybe_redundant_hierarchical_relations into safe/unsafe cases
- Update apply_fixes to handle safe hierarchical relation removal
- Add sorting for deterministic output in lint reports
- Update test suite with comprehensive test cases

Testing:
- Extended test-lint-command with single-chain and multi-path examples
- Verified diff-based test validation for file modifications
- All tests pass successfully

Traceability:
- Satisfies: specifications/SystemRequirements/Requirements.md#safe-redundant-hierarchical-relations-auto-removal
- Verified by: specifications/Verifications/Misc.md#lint-command-verification

Author: ilijaljubicic

cli/src/cli.rs

@@ -209,6 +209,18 @@ pub enum Commands {
         json: bool,
     },
 
+    /// Generate model structure diagram with optional filtering
+    #[clap(override_help = "Generate model structure diagram with optional filtering\n\nMODEL OPTIONS:\n      --root-id <ID>              Filter model from specific root element (forward relations only)\n      --json                      Output results in JSON format")]
+    Model {
+        /// Filter model from specific root element using forward-only relation traversal
+        #[clap(long, value_name = "ID", help_heading = "MODEL OPTIONS")]
+        root_id: Option<String>,
+
+        /// Output results in JSON format
+        #[clap(long, help_heading = "MODEL OPTIONS")]
+        json: bool,
+    },
+
     /// Analyze model quality and detect issues in requirements relations
     #[clap(override_help = "Analyze model quality and detect issues in requirements relations\n\nLINT OPTIONS:\n      --fixable                   Show only auto-fixable issues\n      --auditable                 Show only issues requiring manual review\n      --fix                       Apply automatic fixes for auto-fixable issues\n      --json                      Output results in JSON format")]
     Lint {
@@ -431,6 +443,7 @@ fn wants_json(args: &Args) -> bool {
         Some(Commands::Matrix { json, .. }) => *json,
         Some(Commands::Traces { json, .. }) => *json,
         Some(Commands::Coverage { json }) => *json,
+        Some(Commands::Model { json, .. }) => *json,
         Some(Commands::Lint { json, .. }) => *json,
         _ => false,
     }
@@ -655,6 +668,18 @@ pub fn handle_command(
             coverage_report.print(json);
             return Ok(0);
         },
+        Some(Commands::Model { root_id, json }) => {
+            // Generate model diagram with optional filtering
+            let forward_only = root_id.is_some(); // Use forward-only when filtering by root element
+            let output = diagrams::generate_model_report(
+                &model_manager.graph_registry,
+                root_id.as_deref(),
+                forward_only,
+                json
+            )?;
+            println!("{}", output);
+            return Ok(0);
+        },
         Some(Commands::Lint { fixable, auditable, fix, json }) => {
             // Run lint analysis
             let lint_report = lint::analyze_model(&model_manager.graph_registry);

core/src/error.rs

@@ -106,4 +106,7 @@ pub enum ReqvireError {
     #[error("Validation failed with {} errors", .0.len())]
     ValidationError(Vec<ReqvireError>),
 
+    #[error("Serialization error: {0}")]
+    SerializationError(String),
+
 }