runner.aws_batch: Download .snakemake/metadata/ too

Snakemake stores state information per input/output here and uses it to
determine if it needs to re-run rules or not.  It seems akin to the file
mtimes which we already take care to preserve on upload/download.
Additionally, the metadata recorded is used in Snakemake's report
generation and is generally useful for looking at workflow statistics.

Continue to not download all of .snakemake/ en masse because it can
potentially contain files that interfere with local usage and/or are
large and unnecessary.

Resolves: <#373>
Related-to: <nextstrain/docker-base#220>

CHANGES.md

@@ -13,6 +13,17 @@ development source code and as such may not be routinely kept up to date.
 
 # __NEXT__
 
+## Improvements
+
+* Snakemake's per-input/output file metadata (stored in `.snakemake/metadata/`)
+  is now downloaded from AWS Batch builds by default.  Like file modification
+  times (mtimes), which are already preserved from the remote build, this
+  additional metadata is used by Snakemake to track when inputs have changed
+  and when it should regenerate outputs.  The metadata is also used in
+  [Snakemake report generation](https://snakemake.readthedocs.io/en/v8.14.0/snakefiles/reporting.html#rendering-reports)
+  and can be useful for gathering ad-hoc workflow statistics.
+  ([#374](https://github.com/nextstrain/cli/pull/374))
+
 
 # 8.4.0 (29 May 2024)
 

nextstrain/cli/runner/aws_batch/s3.py

@@ -119,8 +119,12 @@ def download_workdir(remote_workdir: S3Object, workdir: Path, patterns: List[str
     ])
 
     included = path_matcher([
-        # But we do want the Snakemake logs to come over.
+        # But we do want the Snakemake logs to come over…
         ".snakemake/log/",
+
+        # …and the input/output metadata Snakemake tracks (akin to mtimes,
+        # which we also preserve).
+        ".snakemake/metadata/",
     ])
 
     if patterns: