Native C++ Path

This project includes an optional native toolchain for high-performance sections.

Why this exists

Python remains the orchestration layer. C++ is used for hotspots where throughput matters (large text scans, metrics, or heavy parsing loops).

Current native tool

• chapter_metrics (C++11+)
• Computes UTF-8 text metrics quickly:
  • bytes
  • codepoints
  • lines
  • non-empty lines
  • dialogue lines
  • dialogue density
• story_feature_metrics (C++11+)
• Computes chapter feature primitives used by story_features.v1:
  • source length (UTF-8 codepoints)
  • sentence count
  • token count
  • average sentence length
  • dialogue line ratio

Build with CMake

cmake -S . -B build/cpp
cmake --build build/cpp --config Release

Native quality checks

clang-format --dry-run --Werror cpp/*.cpp
cppcheck --enable=warning,style,performance,portability --error-exitcode=2 cpp

CMake also auto-enables clang-tidy and cppcheck if those tools are installed.

Run demo

ctest --test-dir build/cpp -C Release -R chapter_metrics_demo --output-on-failure
ctest --test-dir build/cpp -C Release -R story_feature_metrics_demo --output-on-failure

Example usage

chapter_metrics --input chapter.txt
story_feature_metrics --input chapter.txt

Or pipe from stdin:

cat chapter.txt | chapter_metrics
cat chapter.txt | story_feature_metrics

Prerequisites

• CMake 3.20+
• A C++11+ compiler (MSVC Build Tools, clang++, or g++)

Make targets

• make cpp-configure
• make cpp-build
• make cpp-test
• make cpp-demo
• make cpp-format
• make cpp-format-check
• make cpp-cppcheck

What should move to C++ next

Highest-priority candidates based on current pipeline hotspots:

1. Segment/chapter token and sentence scanning loops in feature extraction.
2. Narrative segmentation pass over large event lists.
3. Timeline merge/sort for very large extracted event graphs.

Lower-priority for now:

1. API orchestration and persistence adapters (I/O-bound, not CPU-bound).
2. Dashboard read-model shaping (mostly serialization/mapping work).