← Back to projects
gerp is a C++ file search tool inspired by Unix grep. It recursively traverses a directory tree, indexes file contents, and supports both case-sensitive and case-insensitive search queries. The focus of this project is designing the indexing structure so queries stay fast even as the number of files grows.
The core idea is a nested hash-table index that separates case-normalized lookup from case-preserving results.
Indexing & Data Structures
- Efficient nested hash table index: insensitive → sensitive → occurrence count
- Directory traversal implemented with an n-ary tree representation of the filesystem
- Memory-conscious storage designed for large directory trees and datasets
Search Behavior
- Supports case-insensitive and case-sensitive queries
- Robust tokenization/string processing to handle formatting and special characters
- Optimized lookup path to minimize redundant comparisons during querying
Systems Engineering Takeaways
- File I/O patterns and recursive traversal at scale
- Hashing and custom data-structure design for fast query resolution
- Practical performance tradeoffs (index build time vs query speed)
Gerp: Recursive File Search Tool
A performance-oriented text indexing and search system in C++, implementing custom nested hash tables to process large directories efficiently.
C++Hash TableValgrindFile SearchData Structures
Key Highlights
- • Engineered a performance-oriented text indexing and search system in C++, implementing custom hash tables and data structures
- • Processed large text files with predictable latency, memory efficiency, and high throughput
- • Utilized Valgrind to verify memory leak-free execution and profile indexing performance
Project Details
gerp is a C++ file search tool inspired by Unix grep. It recursively traverses a directory tree, indexes file contents, and supports both case-sensitive and case-insensitive search queries. The focus of this project is designing the indexing structure so queries stay fast even as the number of files grows.
The core idea is a nested hash-table index that separates case-normalized lookup from case-preserving results.
Indexing & Data Structures
- Efficient nested hash table index: insensitive → sensitive → occurrence count
- Directory traversal implemented with an n-ary tree representation of the filesystem
- Memory-conscious storage designed for large directory trees and datasets
Search Behavior
- Supports case-insensitive and case-sensitive queries
- Robust tokenization/string processing to handle formatting and special characters
- Optimized lookup path to minimize redundant comparisons during querying
Systems Engineering Takeaways
- File I/O patterns and recursive traversal at scale
- Hashing and custom data-structure design for fast query resolution
- Practical performance tradeoffs (index build time vs query speed)