Add --list-dates option

CHANGELOG.md

@@ -2,6 +2,9 @@ In Development
 --------------
 - Add `--compress-filter-msgs` option
 - Support all documented S3 Inventory fields in inventory lists
+- Add `--list-dates` option
+- The `<outdir>` command-line argument is now optional and defaults to the
+  current directory
 
 v0.1.0-alpha.2 (2025-01-06)
 ---------------------------

README.md

@@ -67,10 +67,11 @@ for further options.
 Usage
 =====
 
-    s3invsync [<options>] <inventory-base> <outdir>
+    s3invsync [<options>] <inventory-base> [<outdir>]
 
 `s3invsync` downloads the contents of an S3 bucket, including old versions of
-objects, to the directory `<outdir>` using S3 Inventory files located at
+objects, to the directory `<outdir>` (defaulting to the current working
+directory if not specified) using S3 Inventory files located at
 `<inventory-base>`.
 
 `<inventory-base>` must be of the form `s3://{bucket}/{prefix}/`, where
@@ -112,6 +113,9 @@ Options
 - `-I <INT>`, `--inventory-jobs <INT>` — Specify the maximum number of inventory
   list files to download & process at once [default: 20]
 
+- `--list-dates` — List available inventory manifest dates instead of
+  backing anything up
+
 - `-l <level>`, `--log-level <level>` — Set the log level to the given value.
   Possible values are  "`ERROR`", "`WARN`", "`INFO`", "`DEBUG`", and "`TRACE`"
   (all case-insensitive).  [default value: `DEBUG`]

src/main.rs

@@ -11,6 +11,7 @@ use crate::syncer::Syncer;
 use crate::timestamps::DateMaybeHM;
 use anyhow::Context;
 use clap::Parser;
+use futures_util::TryStreamExt;
 use std::io::{stderr, IsTerminal};
 use std::num::NonZeroUsize;
 use std::path::PathBuf;
@@ -43,6 +44,10 @@ struct Arguments {
     #[arg(short = 'I', long, default_value = "20")]
     inventory_jobs: NonZeroUsize,
 
+    /// List available inventory manifest dates instead of backing anything up
+    #[arg(long)]
+    list_dates: bool,
+
     /// Set logging level
     #[arg(
         short,
@@ -75,8 +80,9 @@ struct Arguments {
     /// a key for a manifest file).
     inventory_base: S3Location,
 
-    /// Directory in which to download the S3 objects
-    outdir: PathBuf,
+    /// Directory in which to download the S3 objects.  Defaults to the current
+    /// working directory.
+    outdir: Option<PathBuf>,
 }
 
 // See
@@ -111,20 +117,31 @@ async fn run(args: Arguments) -> anyhow::Result<()> {
     let region = get_bucket_region(args.inventory_base.bucket()).await?;
     tracing::info!(%bucket, %region, "Found S3 bucket region");
     let client = S3Client::new(region, args.inventory_base, args.trace_progress).await?;
-    tracing::info!("Fetching manifest ...");
-    let (manifest, manifest_date) = client.get_manifest_for_date(args.date).await?;
-    let syncer = Syncer::new(
-        client,
-        args.outdir,
-        manifest_date,
-        start_time,
-        args.inventory_jobs,
-        args.object_jobs,
-        args.path_filter,
-        args.compress_filter_msgs,
-    );
-    tracing::info!("Starting backup ...");
-    syncer.run(manifest).await?;
-    tracing::info!("Backup complete");
+    if args.list_dates {
+        let mut stream = client.list_all_manifest_timestamps();
+        while let Some(date) = stream.try_next().await? {
+            println!("{date}");
+        }
+    } else {
+        let outdir = match args.outdir {
+            Some(p) => p,
+            None => std::env::current_dir().context("failed to determine current directory")?,
+        };
+        tracing::info!("Fetching manifest ...");
+        let (manifest, manifest_date) = client.get_manifest_for_date(args.date).await?;
+        let syncer = Syncer::new(
+            client,
+            outdir,
+            manifest_date,
+            start_time,
+            args.inventory_jobs,
+            args.object_jobs,
+            args.path_filter,
+            args.compress_filter_msgs,
+        );
+        tracing::info!("Starting backup ...");
+        syncer.run(manifest).await?;
+        tracing::info!("Backup complete");
+    }
     Ok(())
 }

src/s3/mod.rs

@@ -119,6 +119,11 @@ impl S3Client {
         Ok((manifest, ts))
     }
 
+    /// Returns a stream yielding all available inventory manifest timestamps
+    pub(crate) fn list_all_manifest_timestamps(&self) -> ListManifestDates {
+        ListManifestDates::new(self, &self.inventory_base)
+    }
+
     /// Return the full timestamp for the latest manifest, either (if `day` is
     /// `None`) out of all manifests or else the latest on the given date.
     #[tracing::instrument(skip_all, fields(day = day.map(|d| d.to_string())))]

src/s3/streams.rs

@@ -19,7 +19,7 @@ type InnerListError = SdkError<ListObjectsV2Error, HttpResponse>;
 /// parses their names as [`DateHM`] values, yielding the successful parses.
 #[derive(Debug)]
 #[must_use = "streams do nothing unless polled"]
-pub(super) struct ListManifestDates {
+pub(crate) struct ListManifestDates {
     url: S3Location,
     inner: Option<PaginationStream<Result<ListObjectsV2Output, InnerListError>>>,
     results: VecDeque<DateHM>,