clap - Command Line Argument Parser
To make a user friendly CLI, even in simple situations, the popular library clap is the best tool for the job. I use it to generate the --help section to the cli, get the provided path to the transaction file, and let a user know if it's missing.
polars - DataFrame Library based on Apache Arrow
Having a good dataframe makes all the difference for data driven applications. Polars is to a large extend immutable, so evey change to the dataframe happens by generating a new one and releasing the old - this will in many cases allow for parallelization. It's also columnar because it's build on Apache Arror, and that allows for great performance - dispite demanding an entire different mindset when using it instead of row-based solutions. The list of accounts in this library is a dataframe, where the transactions are not sice they are streamed from the csv to avoid keeping them all in memory at a time.
serde - serialization/deserialization
The best libarry out there for serialization, deserialization. I use it to read the transacrions, and through Polars, I use it to output the accounts dataframe as csv formatted data to the stdout bufwriter as well.
csv - read/write csv files
It used to accounts data in csv-format in output.rs and to read accounts data in csv-format for testing each type of transacrion in load_accounts in the accounts.rs file.
Each type of transaction has it's own file and function, which will take a reference to the existing list of accounts and disputes, and a execute the transaction. A new list of accounts is being returned. I chose this structure because it's easy to test each type of transaction, and because it avoids having too much responsibility in a single file.
Testing is done by executing a set of transactions, and compaing the account output to the expected output. For this a set of fixtures are use, which can be found in the ./fixtures directory.