Add 10 Greptile-inspired review intelligence modules with 246 tests

Implement graph-based symbol indexing, NL code summaries, function-level
diff chunking, multi-pass hotspot review, convention learning from
feedback, PR comment history analysis, git churn risk scoring,
composable pipeline architecture, offline/self-hosted model support,
and eval benchmark suites. Integrates all modules into the review
pipeline via enhanced_review.rs.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: haasonsaas

src/core/code_summary.rs

@@ -1,21 +1,35 @@
 pub mod changelog;
+pub mod code_summary;
 pub mod comment;
 pub mod commit_prompt;
+pub mod composable_pipeline;
 pub mod context;
+pub mod convention_learner;
 pub mod diff_parser;
+pub mod enhanced_review;
+pub mod eval_benchmarks;
+pub mod function_chunker;
 pub mod git;
+pub mod git_history;
 pub mod interactive;
+pub mod multi_pass;
+pub mod offline;
+pub mod pr_history;
 pub mod pr_summary;
 pub mod prompt;
 pub mod rules;
 pub mod smart_review_prompt;
+pub mod symbol_graph;
 pub mod symbol_index;
 
 pub use changelog::ChangelogGenerator;
 pub use comment::{Comment, CommentSynthesizer};
 pub use commit_prompt::CommitPromptBuilder;
 pub use context::{ContextFetcher, ContextType, LLMContextChunk};
 pub use diff_parser::{DiffParser, UnifiedDiff};
+pub use enhanced_review::{
+    build_enhanced_context, apply_enhanced_filters, generate_enhanced_guidance,
+};
 pub use git::GitIntegration;
 pub use pr_summary::{PRSummaryGenerator, SummaryOptions};
 pub use prompt::PromptBuilder;