Website/src/teaching/cse-350/2024sp/assignments/p1.md

P1 - Binary Search (On Disk)

Deadline: Sunday, Feb 11; 11:59 PM

Accept Assignment: https://classroom.github.com/a/hsxNiYoj

Submit Assignment: https://autolab.cse.buffalo.edu/courses/cse410-s24/assessments/P1-Binary

In this assignment, we will implement binary search using O(1) memory using file handles.

This assignment is intended to:

• Familiarize you with Rust and Cargo
• Familiarize you with Rust's File API, including Seek
• Familiarize you with working with binary data encodings
• Familiarize you with implementing bounded-memory algorithms.

You should expect to spend approximately 10-15 hours on this assignment. Plan accordingly.

---

To complete this assignment, you should:

1. Accept this assignment through GitHub Classroom.
2. Modify the file src/data_file.rs, implementing the functions labeled todo!().
3. Commit your changes and push them to Github.
4. Go to Autolab, select your repository, acknowledge the course AI Policy, and click Submit.

You may repeat steps 2-4 as many times as desired

---

Overview

In this assignment, you will be provided with a data file consisting of an arbitrary number of serialized Record objects, each consisting of a key and a value. Each record will have a unique key, and records will be stored in ascending sorted order of their key.

Your data_file::DataFile implementation should be able to:

• Open the file
• Retrieve the nth record from the file
• Perform an O(1)-memory binary search over the file to find a specific key

---

Documentation

You may find the following documentation useful:

• The Rust Book
• std::fs::File
• std::fs::Metadata

---

The following utility methods are provided for your convenience:

buffer_to_record(buffer)

Given a buffer, exactly the size of one record, this function will transmute it into a Record object.

---

Objectives

In this assignment, you will implement three functions:

DataFile::open(path)

This method should instantiate a DataFile object using the file at the provided path. Note the four fields of a DataFile:

• file: A File reference storing an open, read-only filehandle.
• number_of_records: The number of records in the file.
• min_key: The least key of any record in the file (the key of the first record)
• max_key: The greatest key of any record in the file (the key of the last record)

You should derive the number_of_records, min_key, and max_key attributes directly from the file. The length of the file (in bytes) is given as part of the file's Metadata.

Complexity:

• Runtime: O(1)
• Memory: O(1)
• IO: O(1)

data_file.get(idx)

This method should return the idxth record stored in the file. If idx is out of bounds, you should panic.

Note the buffer_to_record helper function.

Note also the bound on memory.

Complexity:

• Runtime: O(1)
• Memory: O(1)
• IO: O(1)

data_file.find(key)

If a record with key key is present in the file, this method should return it. If a record is not present, this function should return:

• The successor of key (the record with the next highest key) if one exists
• None if key has no successor

You may assume that the records in the file are stored in sorted order.

Note the bound on memory.

Complexity:

• Runtime: O(log_2(N))
• Memory: O(1)
• IO: O(log_2(N))