feat: Implement efficient node lookup table using node indices (#463)

Major Changes:
- Add generic Lookup<T> structure for efficient O(1) module-level data access
- Combine separate scope and order lookups into unified QuerySchedule structure
- Add query_schedule field to Interpreter for dedicated user query scheduling
- Refactor loop hoising to separate module
- Use efficient lookup for loop vars
- Also added more tests for loops

Key Concept:
- Ensure module context and indexing stay synchronized during function calls

Testing:
- All scheduler and interpreter tests passing

Signed-off-by: Anand Krishnamoorthi <anakrish@microsoft.com>

src/tests/scheduler/analyzer/mod.rs

@@ -52,11 +52,27 @@ fn analyze_file(regos: &[String], expected_scopes: &[Scope]) -> Result<()> {
 
     let analyzer = Analyzer::new();
     let schedule = analyzer.analyze(&modules)?;
-    let mut scopes: Vec<(Ref<Query>, &crate::scheduler::Scope)> = schedule
-        .scopes
-        .iter()
-        .map(|(r, s)| (r.clone(), s))
-        .collect();
+
+    // Collect all queries from modules to create the mapping
+    let mut all_queries = Vec::new();
+    for (module_idx, module) in modules.iter().enumerate() {
+        for rule in &module.policy {
+            if let Rule::Spec { bodies, .. } = rule.as_ref() {
+                for body in bodies {
+                    all_queries.push((module_idx as u32, body.query.qidx, body.query.clone()));
+                }
+            }
+        }
+    }
+
+    let mut scopes = Vec::new();
+    for (module_idx, qidx, query) in all_queries.iter() {
+        // Find the corresponding query schedule
+        if let Some(query_schedule) = schedule.queries.get(*module_idx, *qidx) {
+            scopes.push((query.clone(), &query_schedule.scope));
+        }
+    }
+
     scopes.sort_by(|a, b| a.0.span.line.cmp(&b.0.span.line));
     for (idx, (_, scope)) in scopes.iter().enumerate() {
         if idx > expected_scopes.len() {