fetch: add --num option to fetch the last N commits

dvc/commands/data_sync.py

@@ -93,6 +93,7 @@ def run(self):
                 all_branches=self.args.all_branches,
                 all_tags=self.args.all_tags,
                 all_commits=self.args.all_commits,
+                num=self.args.num,
                 with_deps=self.args.with_deps,
                 recursive=self.args.recursive,
                 run_cache=self.args.run_cache,
@@ -135,7 +136,7 @@ def shared_parent_parser():
     return parent_parser
 
 
-def add_parser(subparsers, _parent_parser):
+def add_parser(subparsers, _parent_parser):  # noqa: PLR0915
     from dvc.commands.status import CmdDataStatus
 
     # Pull
@@ -309,6 +310,18 @@ def add_parser(subparsers, _parent_parser):
         default=False,
         help="Fetch cache for all commits.",
     )
+    fetch_parser.add_argument(
+        "--num",
+        type=int,
+        default=1,
+        dest="num",
+        metavar="<number>",
+        help=(
+            "Fetch cache for the last `num` commits. Defaults to the current "
+            "branch when used on its own, or applies to each branch with "
+            "`--all-branches`."
+        ),
+    )
     fetch_parser.add_argument(
         "-d",
         "--with-deps",

dvc/repo/fetch.py

@@ -1,6 +1,6 @@
 from typing import TYPE_CHECKING
 
-from dvc.exceptions import DownloadError
+from dvc.exceptions import DownloadError, InvalidArgumentError
 from dvc.log import logger
 from dvc.stage.cache import RunCacheNotSupported
 from dvc.ui import ui
@@ -35,6 +35,7 @@ def _collect_indexes(  # noqa: PLR0913
     recursive=False,
     all_commits=False,
     revs=None,
+    num=1,
     workspace=True,
     max_size=None,
     types=None,
@@ -66,6 +67,7 @@ def outs_filter(out: "Output") -> bool:
         all_branches=all_branches,
         all_tags=all_tags,
         all_commits=all_commits,
+        num=num,
         workspace=workspace,
     ):
         try:
@@ -110,6 +112,7 @@ def fetch(  # noqa: PLR0913
     all_commits=False,
     run_cache=False,
     revs=None,
+    num=1,
     workspace=True,
     max_size=None,
     types=None,
@@ -136,6 +139,13 @@ def fetch(  # noqa: PLR0913
     if isinstance(targets, str):
         targets = [targets]
 
+    if num < 1:
+        raise InvalidArgumentError("`--num` must be a positive integer.")
+
+    if num > 1 and not (revs or all_branches or all_tags or all_commits):
+        # `--num` on its own fetches the last `num` commits of the current branch.
+        revs = ["HEAD"]
+
     failed_count = 0
     transferred_count = 0
 
@@ -157,6 +167,7 @@ def fetch(  # noqa: PLR0913
         recursive=recursive,
         all_commits=all_commits,
         revs=revs,
+        num=num,
         workspace=workspace,
         max_size=max_size,
         types=types,

tests/func/test_data_cloud.py

@@ -537,6 +537,36 @@ def test_push_pull_all(tmp_dir, scm, dvc, local_remote, key, expected):
     }
 
 
+def test_fetch_num(tmp_dir, scm, dvc, local_remote):
+    # Three commits, each tracking a distinct version (hash) of the same file.
+    tmp_dir.dvc_gen("foo", "first", commit="first")
+    tmp_dir.dvc_gen("foo", "second", commit="second")
+    tmp_dir.dvc_gen("foo", "third", commit="third")
+
+    assert dvc.push(all_commits=True) == 3
+
+    # `--num` on its own fetches the last `num` commits of the current branch:
+    # HEAD ("third") and HEAD~1 ("second"), but not HEAD~2 ("first").
+    clean(["foo"], dvc)
+    assert dvc.fetch(num=2) == 2
+
+    # default (`num=1`) only fetches the workspace revision ("third").
+    clean(["foo"], dvc)
+    assert dvc.fetch() == 1
+
+    # `--num` larger than history is capped at the available commits.
+    clean(["foo"], dvc)
+    assert dvc.fetch(num=10) == 3
+
+
+def test_fetch_num_invalid(tmp_dir, dvc, local_remote):
+    from dvc.exceptions import InvalidArgumentError
+
+    tmp_dir.dvc_gen({"foo": "foo"})
+    with pytest.raises(InvalidArgumentError):
+        dvc.fetch(num=0)
+
+
 def test_push_pull_fetch_pipeline_stages(tmp_dir, dvc, run_copy, local_remote):
     tmp_dir.dvc_gen("foo", "foo")
     run_copy("foo", "bar", name="copy-foo-bar")

tests/unit/command/test_data_sync.py

@@ -15,6 +15,8 @@ def test_fetch(mocker, dvc):
             "--all-branches",
             "--all-tags",
             "--all-commits",
+            "--num",
+            "2",
             "--with-deps",
             "--recursive",
             "--run-cache",
@@ -40,6 +42,7 @@ def test_fetch(mocker, dvc):
         all_branches=True,
         all_tags=True,
         all_commits=True,
+        num=2,
         with_deps=True,
         recursive=True,
         run_cache=True,