Upload 2 files

app.py

@@ -6,7 +6,6 @@ import os
 import tempfile
 import shutil
 import urllib
-import hashlib
 from huggingface_hub import whoami, HfApi, hf_hub_download, RepoCard
 from huggingface_hub.utils import build_hf_headers, hf_raise_for_status
 from gradio_huggingfacehub_search import HuggingfaceHubSearch
@@ -20,6 +19,35 @@ HF_REPO_PREFIX = os.environ.get("HF_REPO_PREFIX") if os.environ.get("HF_REPO_PRE
 HF_REPO_SUFFIX = os.environ.get("HF_REPO_SUFFIX") if os.environ.get("HF_REPO_SUFFIX") else "" # set your default repo suffix
 HF_USER = os.environ.get("HF_USER") if os.environ.get("HF_USER") else "" # set your username
 REGEX_HF_REPO = r'^[\w_\-\.]+/[\w_\-\.]+$'
+REGEX_HF_PATH = r'^[\w_\-\.]+/[\w_\-\.]+(/?:.+)?$'
+
+def is_valid_reponame(repo_id: str):
+    return re.fullmatch(REGEX_HF_REPO, repo_id)
+
+def is_valid_path(hf_path: str):
+    return re.fullmatch(REGEX_HF_PATH, hf_path)
+
+def extract_src_reponame(source_repo: str):
+    try:
+        if is_valid_reponame(source_repo): target = ""
+        else:
+            source_repo, target = re.findall(r'^(?:http.+\.co/)?(?:datasets)?(?:spaces)?([\w_\-\.]+/[\w_\-\.]+)/?(?:blob/main/)?(?:resolve/main/)?(.+)?$', source_repo)[0]
+            target = urllib.parse.unquote(target.removesuffix("/"))
+        return source_repo, target
+    except Exception as e:
+        print(e)
+        return source_repo, ""
+
+def extract_dst_reponame(dst_repo: str):
+    try:
+        if is_valid_reponame(dst_repo): subfolder = ""
+        else:
+            dst_repo, subfolder = re.findall(r'^([\w_\-\.]+/[\w_\-\.]+)/?(.+)?$', dst_repo)[0]
+            subfolder = subfolder.removesuffix("/")
+        return dst_repo, subfolder
+    except Exception as e:
+        print(e)
+        return dst_repo, ""
 
 def remove_repo_tags(repo_id: str, tags: list[str], repo_type: str, hf_token: str):
     try:
@@ -43,16 +71,11 @@ def duplicate(source_repo, dst_repo, repo_type, private, overwrite, auto_dir, re
     except Exception as e:
         raise gr.Error(f"""Oops, you forgot to login. Please use the loggin button on the top left to migrate your repo {e}""")
     
-    try:
-        if re.fullmatch(REGEX_HF_REPO, source_repo): target = ""
-        else:
-            source_repo, target = re.findall(r'^(?:http.+\.co/)?(?:datasets)?(?:spaces)?([\w_\-\.]+/[\w_\-\.]+)/?(?:blob/main/)?(?:resolve/main/)?(.+)?$', source_repo)[0]
-            target = urllib.parse.unquote(target.removesuffix("/"))
+    if not is_valid_path(dst_repo): raise gr.Error(f"Invalid dst_repo: {dst_repo}")
 
-        if re.fullmatch(REGEX_HF_REPO, dst_repo): subfolder = ""
-        else:
-            dst_repo, subfolder = re.findall(r'^([\w_\-\.]+/[\w_\-\.]+)/?(.+)?$', dst_repo)[0]
-            subfolder = subfolder.removesuffix("/")
+    try:
+        source_repo, target = extract_src_reponame(source_repo)
+        dst_repo, subfolder = extract_dst_reponame(dst_repo)
         if auto_dir: subfolder = source_repo
 
         if not overwrite and api.repo_exists(repo_id=dst_repo, repo_type=repo_type, token=hf_token): raise gr.Error(f"Repo already exists {dst_repo}")
@@ -124,18 +147,14 @@ def duplicate_m2o(source_repos_str, dst_repo, repo_type, private, overwrite, oau
     except Exception as e:
         raise gr.Error(f"""Oops, you forgot to login. Please use the loggin button on the top left to migrate your repo {e}""")
     
+    if not is_valid_path(dst_repo): raise gr.Error(f"Invalid dst_repo: {dst_repo}")
+
     try:
-        if re.fullmatch(REGEX_HF_REPO, dst_repo): subfolder_prefix = ""
-        else:
-            dst_repo, subfolder_prefix = re.findall(r'^([\w_\-\.]+/[\w_\-\.]+)/?(.+)?$', dst_repo)[0]
-            subfolder_prefix = subfolder.removesuffix("/")
+        dst_repo, subfolder_prefix = extract_dst_reponame(dst_repo)
         if not overwrite and api.repo_exists(repo_id=dst_repo, repo_type=repo_type, token=hf_token): raise gr.Error(f"Repo already exists {dst_repo}")
         source_repos = parse_repos(source_repos_str)
         for source_repo in source_repos:
-            if re.fullmatch(REGEX_HF_REPO, source_repo): target = ""
-            else:
-                source_repo, target = re.findall(r'^(?:http.+\.co/)?(?:datasets)?(?:spaces)?([\w_\-\.]+/[\w_\-\.]+)/?(?:blob/main/)?(?:resolve/main/)?(.+)?$', source_repo)[0]
-                target = urllib.parse.unquote(target.removesuffix("/"))
+            source_repo, target = extract_src_reponame(source_repo)
 
             subfolder = subfolder_prefix + "/" + source_repo if subfolder_prefix else source_repo
 
@@ -184,9 +203,9 @@ def duplicate_m2m(source_repos_str, hf_user, repo_type, private, overwrite, remo
         source_repos = parse_repos(source_repos_str)
         repo_url_result = 'Find your repo '
         for source_repo in source_repos:
-            if not re.fullmatch(REGEX_HF_REPO, source_repo) or not api.repo_exists(repo_id=source_repo, repo_type=repo_type, token=hf_token): continue
+            if not is_valid_reponame(source_repo) or not api.repo_exists(repo_id=source_repo, repo_type=repo_type, token=hf_token): continue
             dst_repo = hf_user + "/" + repo_prefix + source_repo.split("/")[-1] + repo_suffix
-            if not re.fullmatch(REGEX_HF_REPO, dst_repo): continue
+            if not is_valid_reponame(dst_repo): continue
             if not overwrite and api.repo_exists(repo_id=dst_repo, repo_type=repo_type, token=hf_token):
                 gr.Info(f"Repo already exists {dst_repo}")
                 continue
@@ -241,9 +260,9 @@ with gr.Blocks(css=css) as demo:
                         repo_type = gr.Dropdown(label="repo_type", choices=REPO_TYPES, value="model")
                         with gr.Row():
                             is_private = gr.Checkbox(label="Make new repo private?", value=True)
-                            is_overwrite = gr.Checkbox(label="Overwrite existing repo?", value=True)
-                            is_subdir = gr.Checkbox(label="Create subdirectories automatically?", value=True)
-                            is_remtag = gr.Checkbox(label="Remove NFAA tag?", value=True)
+                            is_overwrite = gr.Checkbox(label="Overwrite existing repo?", value=False)
+                            is_subdir = gr.Checkbox(label="Create subdirectories automatically?", value=False)
+                            is_remtag = gr.Checkbox(label="Remove NFAA tag?", info="To avoid Inference API bug", value=False)
                     with gr.Row():
                         submit_button = gr.Button("Submit", variant="primary")
                         clear_button = gr.Button("Clear", variant="secondary")
@@ -291,7 +310,7 @@ with gr.Blocks(css=css) as demo:
                         with gr.Row():
                             m2m_is_private = gr.Checkbox(label="Make new repo private?", value=True)
                             m2m_is_overwrite = gr.Checkbox(label="Overwrite existing repo?", value=False)
-                            m2m_is_remtag = gr.Checkbox(label="Remove NFAA tag?", value=True)
+                            m2m_is_remtag = gr.Checkbox(label="Remove NFAA tag?", info="To avoid Inference API bug", value=True)
                     with gr.Row():
                         m2m_submit_button = gr.Button("Submit", variant="primary")
                         m2m_clear_button = gr.Button("Clear", variant="secondary")