feat: add basic version of repo cleaner

.env/lib/python3.10/site-packages/_distutils_hack/override.py

@@ -0,0 +1 @@
+__import__('_distutils_hack').do_override()

.gitignore

@@ -0,0 +1 @@
+.env

README.md

@@ -1,6 +1,6 @@
 ---
-title: Hub Concent Cleaner
-emoji: 📚
+title: Hub Repo Cleaner
+emoji: 🧹 🧼 🗑️
 colorFrom: purple
 colorTo: green
 sdk: gradio
@@ -8,6 +8,12 @@ sdk_version: 4.44.0
 app_file: app.py
 pinned: false
 license: apache-2.0
+hf_oauth: true
+hf_oauth_scopes:
+  - read-repos
+  - write-repos
+  - manage-repos
+
 ---
 
 Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,128 @@
+import gradio as gr
+from gradio.oauth import OAuthToken
+from huggingface_hub import HfApi, whoami
+from huggingface_hub.hf_api import DatasetInfo, ModelInfo, SpaceInfo
+
+
+def get_token(oauth_token: OAuthToken):
+    return oauth_token.token
+
+
+def list_orgs(oauth_token: OAuthToken = None):
+    if oauth_token is None:
+        return []
+    data = whoami(oauth_token.token)
+    if data["auth"]["type"] == "oauth":
+        organisations = [data["name"]] + [org["name"] for org in data["orgs"]]
+    else:
+        organisations = [
+            entry["entity"]["name"]
+            for entry in data["auth"]["accessToken"]["fineGrained"]["scoped"]
+            if "repo.write" in entry["permissions"]
+        ]
+        organisations = [org for org in organisations if org != data["name"]]
+        organisations = [data["name"]] + organisations
+    return organisations
+
+
+def get_org_dropdown(oauth_token: OAuthToken = None):
+    orgs = list_orgs(oauth_token)
+    return gr.Dropdown(
+        label="Organization",
+        choices=orgs,
+        value=orgs[0] if orgs else None,
+        allow_custom_value=True,
+    )
+
+
+def get_resource_type_dropdown(
+    org_name: str, resource_type: str, oauth_token: OAuthToken = None
+):
+    if oauth_token is None:
+        return []
+
+    hf_client = HfApi(token=oauth_token.token)
+
+    if resource_type == "model":
+        resources = hf_client.list_models(author=org_name)
+    elif resource_type == "dataset":
+        resources = hf_client.list_datasets(author=org_name)
+    elif resource_type == "space":
+        resources = hf_client.list_spaces(author=org_name)
+    else:
+        resources = []
+
+    resources = [resource.id for resource in resources]
+
+    return gr.Dropdown(
+        label="Resource",
+        choices=resources,
+        value=resources[0] if resources else None,
+    )
+
+
+def get_resource(resource_id: str, resource_type: str, oauth_token: OAuthToken = None):
+    if oauth_token is None:
+        return []
+
+    hf_client = HfApi(token=oauth_token.token)
+
+    resource: ModelInfo | DatasetInfo | SpaceInfo = hf_client.repo_info(
+        resource_id, repo_type=resource_type
+    )
+    return str(resource)
+
+
+def delete_resource(
+    resource_id: str, resource_type: str, oauth_token: OAuthToken = None
+):
+    if oauth_token is None:
+        return []
+
+    hf_client = HfApi(token=oauth_token.token)
+
+    hf_client.delete_repo(resource_id, repo_type=resource_type, missing_ok=True)
+
+    return gr.Markdown(f"Deleted {resource_id}", visible=True)
+
+
+with gr.Blocks() as demo:
+    gr.Markdown("# 🧹 🧼 🗑️ Hub Repo Cleaner 🗑️ 🧼 🧹")
+    gr.LoginButton()
+    org_name = gr.Dropdown(label="Organization", choices=list_orgs())
+    resource_type = gr.Dropdown(
+        label="Resource Type",
+        choices=["dataset", "model", "space"],
+        value="dataset",
+    )
+    resources = gr.Dropdown(label="Resources")
+    resource_info = gr.Code("Resource Info", language="json")
+
+    org_name.change(
+        get_resource_type_dropdown,
+        inputs=[org_name, resource_type],
+        outputs=[resources],
+    )
+    resource_type.change(
+        get_resource_type_dropdown,
+        inputs=[org_name, resource_type],
+        outputs=[resources],
+    )
+    resources.change(
+        get_resource,
+        inputs=[resources, resource_type],
+        outputs=[resource_info],
+    )
+
+    markdown = gr.Markdown("message", label="Message", visible=False)
+    button = gr.Button("Delete")
+    button.click(
+        delete_resource, inputs=[resources, resource_type], outputs=[markdown]
+    ).then(
+        fn=get_resource_type_dropdown,
+        inputs=[org_name, resource_type],
+        outputs=[resources],
+    )
+    demo.load(get_org_dropdown, outputs=[org_name])
+
+demo.launch()

requirements.txt

@@ -0,0 +1 @@
+gradio[oauth]