rebaseline07_21Sept25_01 - Errno13 log file permission - HF logout gr.LoginButton

file_handler/file_utils.py

@@ -25,7 +25,7 @@ def create_outputdir(root: Union[str, Path], output_dir_string:str = None) -> Pa
     ## map to img_path. Opt to putting output within same output_md folder rather than individual source folders
     output_dir_string = output_dir_string if output_dir_string else "output_dir"  ##redundant SMY: default to outputdir in config file = "output_md"
     output_dir = Path("data") / output_dir_string  #"output_md"  ##SMY: concatenating output str with src Path
-    output_dir.mkdir(mode=0o2644, parents=True, exist_ok=True)
+    output_dir.mkdir(mode=0o2755, parents=True, exist_ok=True)   #,mode=0o2755
     return output_dir
 
 def find_file(file_name: str) -> Path:  #configparser.ConfigParser:
@@ -134,13 +134,12 @@ def check_create_logfile(filename: str, dir_path: Union[str, Path]="logs") -> Pa
     writable_dir = project_root / dir_path if isinstance(dir_path, str) else Path(dir_path)    
     
     try:
-        
-            writable_dir.mkdir(mode=0o2644, parents=True, exist_ok=True)
+        writable_dir.mkdir(mode=0o2755, parents=True, exist_ok=True)
     except PermissionError: ##[Errno 13] Permission denied: '/home/user/app/logs/app_logging_2025-09-18.log'
         warnings.warn("[Errno 13] Permission denied, possibly insufficient permission or Persistent Storage not enable: attempting chmod 0o2644")
         #writable_dir = Path(tempfile.gettempdir())    # 
-        writable_dir.mkdir(mode=0o2644, parents=True, exist_ok=True)
-        writable_dir.chmod(0o2644)
+        writable_dir.mkdir(mode=0o2755, parents=True, exist_ok=True)
+        writable_dir.chmod(0o2755)
         if not writable_dir.is_dir():
             warnings.warn(f"Working without log files in directory: {writable_dir}")
     
@@ -148,7 +147,7 @@ def check_create_logfile(filename: str, dir_path: Union[str, Path]="logs") -> Pa
     # `mkdir()` with `exist_ok=True` prevents a FileExistsError if the folder exists.
     logs_dir = writable_dir / dir_path  #project_root / dir_path
     if not logs_dir.is_dir():
-        logs_dir.mkdir(mode=0o2644, parents=True, exist_ok=True)
+        logs_dir.mkdir(mode=0o2755, parents=True, exist_ok=True)
     
     # 4. Create log file with a timestamp inside the new logs directory.
     # This ensures a unique log file is created for the day the script runs.
@@ -196,7 +195,7 @@ def check_create_file(filename: str, dir_path: Union[str, Path]="logs") -> Path:
     # `exist_ok=True` prevents an error if the directory already exists.
     dir_path = project_root / dir_path
     if not dir_path.is_dir():
-        dir_path.mkdir(parents=True, exist_ok=True, mode=0o2664)  #, mode=0o2644)
+        dir_path.mkdir(parents=True, exist_ok=True, mode=0o2755)  #, mode=0o2644)
     #dir_path.chmod(0) 
     
     file_path = dir_path / filename  # Concatenate directory and filename to get full path
@@ -344,7 +343,7 @@ def write_markdown(
         #md_path = Path("data\\pdf") / "output_md" / f"{src.stem}" / md_name  ##debug
         md_path = Path("data") / output_dir / f"{src.stem}" / md_name  ##debug
     ##SMY: [resolved] Permission Errno13 - https://stackoverflow.com/a/57454275
-    md_path.parent.mkdir(mode=0o2644, parents=True, exist_ok=True)  ##SMY: create nested md_path if not exists
+    md_path.parent.mkdir(mode=0o2755, parents=True, exist_ok=True)  ##SMY: create nested md_path if not exists
     md_path.parent.chmod(0)
 
     try:

llm/hf_client.py

@@ -99,8 +99,11 @@ class HFChatClient:
                 logger.warning("hf_login_failed", extra={"error": str(exc)})
                 # Silent fallback; client will still work if token is passed directly
                 #pass
-        '''        
-        login_huggingface(self.token) if not is_login_huggingface() else logger.log(level=20, msg=f"logged in to HF Hub already") ## attempt login if not already logged in. NB: HF CLI login prompt would not display in Process Worker.
+        '''
+        
+        login_huggingface(self.token) if not is_login_huggingface() else logger.log(level=20, msg=f"You are logged in to HF Hub already") ## attempt login if not already logged in. NB: HF CLI login prompt would not display in Process Worker.
+        ##SMY: TODO: Mapped with openai_client.py
+        #self.islogged_in = is_login_huggingface()
 
     @staticmethod
     def _normalise_history(history: list, system_message: str, latest_user_message: str) -> list[dict]:

llm/llm_login.py

@@ -1,4 +1,4 @@
-from huggingface_hub import login, logout
+from huggingface_hub import HfApi, login, logout, get_token
 import os
 import traceback
 from time import sleep
@@ -30,12 +30,17 @@ def login_huggingface(token: Optional[str] = None):
     token = token
     # Privacy-first login: try interactive CLI first; fallback to provided/env token only if needed
     try:
-        login()
-        sleep(5)  ##SMY pause for login. Helpful: pool async opex 
-        logger.info("✔️ hf_login already", extra={"mode": "cli"})
+        if HfApi.whoami():
+            logger.info("✔️ hf_login already", extra={"mode": "cli"})
+            #return True
+        else:
+            login()
+            sleep(5)  ##SMY pause for login. Helpful: pool async opex 
+            logger.info("✔️ hf_login already", extra={"mode": "cli"})
+            #return True
     except Exception as exc:
         # Respect common env var names; prefer explicit token arg when provided
-        fallback_token = token or os.getenv("HF_TOKEN") or os.getenv("HUGGINGFACEHUB_API_TOKEN")
+        fallback_token = token or os.getenv("HF_TOKEN") or os.getenv("HUGGINGFACEHUB_API_TOKEN") or get_token()
         if fallback_token:
             try:
                 login(token=fallback_token)

llm/openai_client.py

@@ -45,11 +45,12 @@ class OpenAIChatClient:
             #self.token = api_token if api_token else None   ##debug
             self.token = openai_api_key_env if openai_api_key_env else api_token  #dotenv.get_key(".env", "OPENAI_API_KEY")
             #self.token = token or os.getenv("HF_TOKEN") or os.getenv("HUGGINGFACEHUB_API_TOKEN")  ## not preferred
-            login_huggingface(self.token) if not is_login_huggingface() else logger.log(level=20, msg=f"logged in to HF Hub already") ## attempt login if not already logged in. NB: HF CLI login prompt would not display in Process Worker.
+            login_huggingface(self.token) if not is_login_huggingface() else logger.log(level=20, msg=f"You are logged in to HF Hub already") ## attempt login if not already logged in. NB: HF CLI login prompt would not display in Process Worker.
             #self.fake_token = api_token or "a1b2c3" #or os.getenv("HF_TOKEN") or os.getenv("HUGGINGFACEHUB_API_TOKEN")
             self.openai_api_key = self.token  #self.fake_token
             self.temperature = temperature
             self.top_p = top_p
+            self.islogged_in = is_login_huggingface()
 
             logger.log(level=2, msg="initialised OpenAIChatClient:", extra={"base_url": self.base_url, "openai_api_key": self.openai_api_key})
             

ui/gradio_ui.py

@@ -248,39 +248,7 @@ def convert_pdfs_to_md(file: gr.File | None, folder: str | None) -> dict:
     # Gradio expects a dict of {filename: content}
     return results
 
-## SMY: to refactor and moved to logic file. Currently unused
-def convert_htmls_to_md(file: gr.File | None, folder: str | None) -> dict:
-    """
-    Gradio callback for HTML → Markdown.
-    Accepts either a single file or a folder path (recursively).
-    Returns a dictionary of filename → Markdown string.
-    """
-    if not file and not folder:
-        return {"error": "Please provide a HTML file or a folder."}
-
-    html_paths = []
-
-    # Single file
-    if file:
-        html_path = Path(file.name)
-        html_paths.append(html_path)
-
-    # Folder (recursively)
-    if folder:
-        try:
-            html_paths.extend(collect_html_paths(folder))
-        except Exception as exc:
-            logger.exception("Folder traversal failed.")
-            return {"error": str(exc)}
-
-    if not html_paths:
-        return {"error": "No HTML files found."}
-
-    results = html2md_converter.batch_convert(html_paths)
-    # Gradio expects a dict of {filename: content}
-    return results
-
-## SMY: to refactor and moved to logic file
+## SMY: to be implemented AND to refactor and moved to logic file
 def convert_md_to_pdf(file: gr.File | None, folder: str | None) -> list[gr.File]:
     """
     Gradio callback for Markdown → PDF.
@@ -314,6 +282,42 @@ def convert_md_to_pdf(file: gr.File | None, folder: str | None) -> list[gr.File]
     # Convert to Gradio File objects
     gr_files = [gr.File(path=str(p)) for p in pdf_files]
     return gr_files
+
+
+## SMY: to refactor and moved to logic file. Currently unused
+'''
+def convert_htmls_to_md(file: gr.File | None, folder: str | None) -> dict:
+    """
+    Gradio callback for HTML → Markdown.
+    Accepts either a single file or a folder path (recursively).
+    Returns a dictionary of filename → Markdown string.
+    """
+    if not file and not folder:
+        return {"error": "Please provide a HTML file or a folder."}
+
+    html_paths = []
+
+    # Single file
+    if file:
+        html_path = Path(file.name)
+        html_paths.append(html_path)
+
+    # Folder (recursively)
+    if folder:
+        try:
+            html_paths.extend(collect_html_paths(folder))
+        except Exception as exc:
+            logger.exception("Folder traversal failed.")
+            return {"error": str(exc)}
+
+    if not html_paths:
+        return {"error": "No HTML files found."}
+
+    results = html2md_converter.batch_convert(html_paths)
+    # Gradio expects a dict of {filename: content}
+    return results
+'''
+
 ##====================
 
 def build_interface() -> gr.Blocks:
@@ -530,19 +534,25 @@ def build_interface() -> gr.Blocks:
         max_retries_sl.change(update_state_stored_value, inputs=max_retries_sl, outputs=state_max_retries)
 
 
-        with gr.Accordion("🤗 HuggingFace Logout", open=False):
+        with gr.Accordion("🤗 HuggingFace Client Logout", open=True):  #, open=False):
             # Logout controls
             def do_logout():
                 #ok = docextractor.client.logout()
                 ok = docconverter.client.logout()
                 # Reset token textbox on successful logout
-                msg = "✅ Logged out of Hugging Face and cleared tokens." if ok else "⚠️ Logout failed."
-                return gr.update(value=""), gr.update(visible=True, value=msg)
+                msg = "✅ Logged out of HuggingFace and cleared tokens. Remember to log out of HuggingFace completely." if ok else "⚠️ Logout failed."
+                return gr.update(value=""), gr.update(visible=True, value=msg), gr.update(value="Sign in to HuggingFace 🤗")
+            
+            def custom_do_logout():
+                return gr.update(value="Sign in to HuggingFace 🤗")
             
             logout_status = gr.Markdown(visible=False)
-            logout_btn = gr.Button("Logout from Hugging Face", variant="stop")
+            with gr.Row:
+                hf_login_logout_btn = gr.LoginButton(value="Sign in to HuggingFace 🤗", logout_value="Logout of HF: ({})", variant="huggingface")
+                logout_btn = gr.Button("Logout from session and Hugging Face (inference) Client", variant="stop", )
 
-            logout_btn.click(fn=do_logout, inputs=None, outputs=[api_token_tb, logout_status]) 
+            hf_login_logout_btn.click(fn=custom_do_logout, inputs=None, outputs=hf_login_logout_btn)
+            logout_btn.click(fn=do_logout, inputs=None, outputs=[api_token_tb, logout_status, hf_login_logout_btn]) 
 
         
         # The gr.State component to hold the accumulated list of files