Add support for rm_file in GitHubFileSystem implementation

fsspec/implementations/github.py

@@ -265,3 +265,63 @@ def _open(
             cache_options=cache_options,
             **kwargs,
         )
+
+    def rm(self, path, **kwargs):
+        self.rm_file(path, **kwargs)
+
+    def rm_file(self, path, branch=None, message=None, **kwargs):
+        """
+        Remove a file from a specified branch using a given commit message.
+
+        Parameters
+        ----------
+        path : str
+            The file's location relative to the repository root.
+        branch : str, optional
+            The branch containing the file. Defaults to the repository's default branch if not provided.
+        message : str, optional
+            The commit message for the deletion.
+        """
+
+        if not self.username:
+            raise ValueError("Authentication required")
+
+        path = self._strip_protocol(path)
+
+        # Attempt to get SHA from cache or Github API
+        sha = self._get_sha_from_cache(path)
+        if not sha:
+            url = self.content_url.format(
+                org=self.org, repo=self.repo, path=path.lstrip("/"), sha=self.root
+            )
+            r = requests.get(url, timeout=self.timeout, **self.kw)
+            if r.status_code == 404:
+                raise FileNotFoundError(path)
+            r.raise_for_status()
+            sha = r.json()["sha"]
+
+        # Delete the file
+        delete_url = self.content_url.format(
+            org=self.org, repo=self.repo, path=path, sha=self.root
+        )
+        data = {
+            "message": message or f"Delete {path}",
+            "sha": sha,
+            **({"branch": branch} if branch else {}),
+        }
+        r = requests.delete(delete_url, json=data, timeout=self.timeout, **self.kw)
+        r.raise_for_status()
+
+        self.invalidate_cache(path)
+
+    def _get_sha_from_cache(self, path):
+        sha = None
+        for entries in self.dircache.values():
+            for entry in entries:
+                entry_path = entry.get("name")
+                if entry_path and entry_path == path and "sha" in entry:
+                    sha = entry["sha"]
+                    break
+            if sha:
+                break
+        return sha

fsspec/implementations/tests/test_github.py

@@ -1,4 +1,5 @@
 import fsspec
+import pytest
 
 
 def test_github_open_small_file():
@@ -46,3 +47,17 @@ def test_github_ls():
     expected = {"brain_networks.csv", "mpg.csv", "penguins.csv", "README.md", "raw"}
     # check if the result is a subset of the expected files
     assert expected.issubset(ls_result)
+
+
+def test_github_rm():
+    # trying to remove a file without passing authentication should raise ValueError
+    fs = fsspec.filesystem("github", org="mwaskom", repo="seaborn-data")
+    with pytest.raises(ValueError):
+        fs.rm("mpg.csv")
+
+    # trying to remove a file which doesn't exist should raise FineNotFoundError
+    fs = fsspec.filesystem(
+        "github", org="mwaskom", repo="seaborn-data", username="user", token="token"
+    )
+    with pytest.raises(FileNotFoundError):
+        fs.rm("/this-file-doesnt-exist", branch="master", message="Delete my file")