docs: Add documentation

Signed-off-by: Dheshan Mohandass <dheshan@mohandass.com>

Author: dheshanm

src/bin/diskusage/main.rs

@@ -19,6 +19,18 @@ struct Arguments {
  debug: bool,
 }
 
+/// Ensure that the user exists in the database.
+/// If the user does not exist, insert the user into the database.
+/// Users are stored in a cache to prevent querying the database for the same user multiple times.
+///
+/// Arguments
+/// * `owner` - The user id to check.
+/// * `pool` - The database connection pool.
+/// * `cache` - The cache to store user ids.
+///
+/// Returns
+/// * Ok(()) if the user exists or is inserted successfully.
+/// * Err(sqlx::Error) if an error occurs while inserting the user.
 async fn ensure_user_exists(
  owner: Option<i32>,
  pool: std::sync::Arc<sqlx::Pool<sqlx::Postgres>>,
@@ -47,6 +59,17 @@ async fn ensure_user_exists(
  Ok(())
 }
 
+/// Process a directory entry.
+/// Insert the directory into the database. If the directory already exists, update the directory.
+///
+/// Arguments
+/// * `entry` - The directory entry to process.
+/// * `pool` - The database connection pool.
+/// * `handle` - The tokio runtime handle.
+/// * `cache` - The cache to store user ids.
+///
+/// Returns
+/// * None
 fn process_directory(
  entry: walkdir::DirEntry,
  pool: std::sync::Arc<sqlx::Pool<sqlx::Postgres>>,
@@ -74,6 +97,19 @@ fn process_directory(
  });
 }
 
+/// Process a file entry.
+/// Insert the file into the database. If the file already exists, update the file.
+/// If the file is not inserted, retry the insert. Assume the file is not inserted due to
+/// its parent directory not being inserted.
+///
+/// Arguments
+/// * `entry` - The file entry to process.
+/// * `pool` - The database connection pool.
+/// * `handle` - The tokio runtime handle.
+/// * `cache` - The cache to store user ids.
+///
+/// Returns
+/// * None
 fn process_file(
  entry: walkdir::DirEntry,
  pool: std::sync::Arc<sqlx::Pool<sqlx::Postgres>>,

src/bin/estimate/main.rs

@@ -11,8 +11,10 @@ struct Arguments {
  /// The path to the directory to estimate.
  #[clap(short, long)]
  path: String,
+ /// The number of large files to display.
  #[clap(short, long, default_value = "5")]
  large_files_count: usize,
+ /// The offset to start displaying large files.
  #[clap(short, long, default_value = "0")]
  offset: usize,
 }
@@ -82,7 +84,14 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
  let size_gb: BigDecimal = size_mb.clone() / 1024;
  let size_tb: BigDecimal = size_gb.clone() / 1024;
 
- log::info!("Estimated size: {:.2} TB = {:.2} GB = {:.2} MB = {:.2} KB = {:.2} bytes", size_tb, size_gb, size_mb, size_kb, total_size);
+ log::info!(
+ "Estimated size: {:.2} TB = {:.2} GB = {:.2} MB = {:.2} KB = {:.2} bytes",
+ size_tb,
+ size_gb,
+ size_mb,
+ size_kb,
+ total_size
+ );
 
  let largest_files_query = format!(
  r#"
@@ -131,7 +140,8 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
  for i in 0..df_row_count {
  let mut row: Vec<String> = vec![];
  for series in df.get_columns() {
- let value: Result<polars::prelude::AnyValue, polars::prelude::PolarsError> = series.get(i);
+ let value: Result<polars::prelude::AnyValue, polars::prelude::PolarsError> =
+ series.get(i);
  let value_str = match value {
  Ok(value) => value.to_string(),
  Err(_) => "".to_string(),
@@ -143,7 +153,8 @@ async fn main() -> Result<(), Box<dyn std::error::Error>> {
 
  println!("Largest files in directory:");
  println!("{table}")
- }).await?;
+ })
+ .await?;
 
  Ok(())
 }