From 952cd981b9b8370a4bfa231a7c5c528184e92fcf Mon Sep 17 00:00:00 2001
From: Maximilian Knespel <mxmlnkn@users.noreply.github.com>
Date: Tue, 15 Oct 2024 17:47:55 +0200
Subject: [PATCH] Speed up git backend (#1712)

---
 fsspec/implementations/git.py            | 66 ++++++++++--------------
 fsspec/implementations/tests/test_git.py | 33 ++++++++++++
 fsspec/spec.py                           |  2 +-
 3 files changed, 61 insertions(+), 40 deletions(-)

diff --git a/fsspec/implementations/git.py b/fsspec/implementations/git.py
index fde671b8c..7b9d3539a 100644
--- a/fsspec/implementations/git.py
+++ b/fsspec/implementations/git.py
@@ -55,6 +55,8 @@ def _path_to_object(self, path, ref):
         tree = comm.tree
         for part in parts:
             if part and isinstance(tree, pygit2.Tree):
+                if part not in tree:
+                    raise FileNotFoundError(path)
                 tree = tree[part]
         return tree
 
@@ -69,46 +71,32 @@ def _get_kwargs_from_urls(path):
             out["ref"], path = path.split("@", 1)
         return out
 
+    @staticmethod
+    def _object_to_info(obj, path=None):
+        # obj.name and obj.filemode are None for the root tree!
+        is_dir = isinstance(obj, pygit2.Tree)
+        return {
+            "type": "directory" if is_dir else "file",
+            "name": (
+                "/".join([path, obj.name or ""]).lstrip("/") if path else obj.name
+            ),
+            "hex": str(obj.id),
+            "mode": "100644" if obj.filemode is None else f"{obj.filemode:o}",
+            "size": 0 if is_dir else obj.size,
+        }
+
     def ls(self, path, detail=True, ref=None, **kwargs):
-        path = self._strip_protocol(path)
-        tree = self._path_to_object(path, ref)
-        if isinstance(tree, pygit2.Tree):
-            out = []
-            for obj in tree:
-                if isinstance(obj, pygit2.Tree):
-                    out.append(
-                        {
-                            "type": "directory",
-                            "name": "/".join([path, obj.name]).lstrip("/"),
-                            "hex": str(obj.id),
-                            "mode": f"{obj.filemode:o}",
-                            "size": 0,
-                        }
-                    )
-                else:
-                    out.append(
-                        {
-                            "type": "file",
-                            "name": "/".join([path, obj.name]).lstrip("/"),
-                            "hex": str(obj.id),
-                            "mode": f"{obj.filemode:o}",
-                            "size": obj.size,
-                        }
-                    )
-        else:
-            obj = tree
-            out = [
-                {
-                    "type": "file",
-                    "name": obj.name,
-                    "hex": str(obj.id),
-                    "mode": f"{obj.filemode:o}",
-                    "size": obj.size,
-                }
-            ]
-        if detail:
-            return out
-        return [o["name"] for o in out]
+        tree = self._path_to_object(self._strip_protocol(path), ref)
+        return [
+            GitFileSystem._object_to_info(obj, path)
+            if detail
+            else GitFileSystem._object_to_info(obj, path)["name"]
+            for obj in (tree if isinstance(tree, pygit2.Tree) else [tree])
+        ]
+
+    def info(self, path, ref=None, **kwargs):
+        tree = self._path_to_object(self._strip_protocol(path), ref)
+        return GitFileSystem._object_to_info(tree, path)
 
     def ukey(self, path, ref=None):
         return self.info(path, ref=ref)["hex"]
diff --git a/fsspec/implementations/tests/test_git.py b/fsspec/implementations/tests/test_git.py
index ffa7b47d9..2aeb544a1 100644
--- a/fsspec/implementations/tests/test_git.py
+++ b/fsspec/implementations/tests/test_git.py
@@ -61,6 +61,39 @@ def test_refs(repo):
         assert f.read() == b"data3"
 
 
+def _check_FileNotFoundError(f, *args, **kwargs):
+    with pytest.raises(FileNotFoundError):
+        f(*args, **kwargs)
+
+
+def test_file_existence_checks(repo):
+    d, sha = repo
+
+    fs, _ = fsspec.url_to_fs(f"git://{d}:abranch@")
+
+    assert fs.lexists("inner")
+    assert fs.exists("inner")
+    assert fs.isdir("inner")
+    assert fs.info("inner")
+    assert fs.ls("inner")
+
+    assert fs.lexists("inner/file1")
+    assert fs.exists("inner/file1")
+    assert fs.info("inner/file1")
+    assert fs.ls("inner/file1")
+
+    assert not fs.lexists("non-existing-file")
+    assert not fs.exists("non-existing-file")
+
+    assert not fs.isfile("non-existing-file")
+    assert not fs.isdir("non-existing-file")
+
+    _check_FileNotFoundError(fs.info, "non-existing-file")
+    _check_FileNotFoundError(fs.size, "non-existing-file")
+    _check_FileNotFoundError(fs.ls, "non-existing-file")
+    _check_FileNotFoundError(fs.open, "non-existing-file")
+
+
 def test_url(repo):
     d, sha = repo
     fs, _, paths = fsspec.core.get_fs_token_paths(f"git://file1::file://{d}")
diff --git a/fsspec/spec.py b/fsspec/spec.py
index e67d280c1..8284366ba 100644
--- a/fsspec/spec.py
+++ b/fsspec/spec.py
@@ -648,7 +648,7 @@ def info(self, path, **kwargs):
         Returns a single dictionary, with exactly the same information as ``ls``
         would with ``detail=True``.
 
-        The default implementation should calls ls and could be overridden by a
+        The default implementation calls ls and could be overridden by a
         shortcut. kwargs are passed on to ```ls()``.
 
         Some file systems might not be able to measure the file's size, in