Fix memory allocation issue for large files by capping StaticSparseDAG capacity (#126)

* Initial plan

* Fix memory allocation issue for large files by capping StaticSparseDAG capacity

Co-authored-by: messense <1556054+messense@users.noreply.github.com>

* Remove new tests and restore sparse_dag module privacy

Co-authored-by: messense <1556054+messense@users.noreply.github.com>

* Update static sparse dag max capacity to ~32MB

---------

Co-authored-by: copilot-swe-agent[bot] <198982749+Copilot@users.noreply.github.com>
Co-authored-by: messense <1556054+messense@users.noreply.github.com>
Co-authored-by: messense <messense@icloud.com>

Author: Copilot

src/sparse_dag.rs

@@ -33,8 +33,14 @@ impl Iterator for EdgeIter<'_> {
 
 impl StaticSparseDAG {
     pub(crate) fn with_size_hint(hint: usize) -> Self {
+        // Cap the allocation to prevent memory issues with very large inputs
+        // The theoretical maximum should be much smaller than hint * 5 for most practical cases
+        const MAX_CAPACITY: usize = 4_000_000; // 1M elements = ~32MB on 64-bit systems
+        const MULTIPLIER: usize = 5;
+
+        let capacity = std::cmp::min(hint * MULTIPLIER, MAX_CAPACITY);
         StaticSparseDAG {
-            array: Vec::with_capacity(hint * 5),
+            array: Vec::with_capacity(capacity),
             start_pos: HashMap::default(),
             size_hint_for_iterator: 0,
             curr_insertion_len: 0,