UnicodeDecodeError: 'cp950' codec can't decode byte 0xe2 in position 2072: illegal multibyte sequence

[Adamchanadam]

Environment : Windows 11 (Conda) with Python 3.11.8 praisonAI 0.0.17

After installation ( : pip install praisonai

I ran the Initialise : praisonai --init create a movie script about dog in moon

it return error code as below , I think it's UTF-8 issue :

Traceback (most recent call last):
  File "<frozen runpy>", line 198, in _run_module_as_main
  File "<frozen runpy>", line 88, in _run_code
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Scripts\praisonai.exe\__main__.py", line 4, in <module>
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\praisonai\__init__.py", line 1, in <module>
    from .cli import PraisonAI
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\praisonai\cli.py", line 11, in <module>
    import gradio as gr
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\__init__.py", line 3, in <module>
    import gradio._simple_templates
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\_simple_templates\__init__.py", line 1, in <module>
    from .simpledropdown import SimpleDropdown
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\_simple_templates\simpledropdown.py", line 6, in <module>
    from gradio.components.base import FormComponent
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\components\__init__.py", line 40, in <module>
    from gradio.components.multimodal_textbox import MultimodalTextbox
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\components\multimodal_textbox.py", line 34, in <module>
    class MultimodalTextbox(FormComponent):
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\component_meta.py", line 198, in __new__
    create_or_modify_pyi(component_class, name, events)
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\site-packages\gradio\component_meta.py", line 92, in create_or_modify_pyi
    source_code = source_file.read_text()
                  ^^^^^^^^^^^^^^^^^^^^^^^
  File "C:\Users\adam\anaconda3\envs\adamlab2_env\Lib\pathlib.py", line 1059, in read_text
    return f.read()
           ^^^^^^^^
UnicodeDecodeError: 'cp950' codec can't decode byte 0xe2 in position 2072: illegal multibyte sequence

[Adamchanadam]

The problem is solved by adding the encoding="utf-8" into the create_or_modify_pyi() function in .\gradio\component_meta.py.

Four modifications are needed:

1 x Change source_file.read_text() to source_file.read_text(encoding="utf-8") once. 3 x Change pyi_file.read_text() to pyi_file.read_text(encoding='utf-8') three times.

def create_or_modify_pyi(
    component_class: type, class_name: str, events: list[str | EventListener]
):
    source_file = Path(inspect.getfile(component_class))

    source_code = source_file.read_text(encoding="utf-8") # <---- add utf-8 here

    current_impl, lineno = extract_class_source_code(source_code, class_name)

    if not (current_impl and lineno):
        raise ValueError("Couldn't find class source code")

    new_interface = create_pyi(current_impl, events)

    pyi_file = source_file.with_suffix(".pyi")
    if not pyi_file.exists():
        last_empty_line_before_class = -1
        lines = source_code.splitlines()
        for i, line in enumerate(lines):
            if line in ["", " "]:
                last_empty_line_before_class = i
            if i >= lineno:
                break
        lines = (
            lines[:last_empty_line_before_class]
            + ["from gradio.events import Dependency"]
            + lines[last_empty_line_before_class:]
        )
        with no_raise_exception():
            pyi_file.write_text("\n".join(lines))
    current_interface, _ = extract_class_source_code(pyi_file.read_text(encoding='utf-8'), class_name) # <---- add utf-8 here
    if not current_interface:
        with no_raise_exception():
            with open(str(pyi_file), mode="a") as f:
                f.write(new_interface)
    else:
        contents = pyi_file.read_text(encoding='utf-8') # <---- add utf-8 here
        contents = contents.replace(current_interface, new_interface.strip())
        current_contents = pyi_file.read_text(encoding='utf-8') # <---- add utf-8 here
        if current_contents != contents:
            with no_raise_exception():
                pyi_file.write_text(contents)

[MervinPraison]

@claude review this request and propose a fix

[claude[bot]]

Claude finished @MervinPraison's task —— View job

---

Analysis: Unicode Decode Error in Gradio Import

Todo List:

• [x] Read repository guidelines (CLAUDE.md) - No CLAUDE.md found
• [x] Review the error and proposed solution
• [x] Analyze if this is a PraisonAI or upstream gradio issue
• [x] Propose appropriate fix strategy
• [x] Provide recommendations for resolution

Issue Analysis:

Root Cause: This is a third-party dependency issue in Gradio's component_meta.py file. The problem occurs when pathlib.read_text() defaults to the system encoding (cp950 on Chinese Windows) instead of UTF-8, causing a decode error when reading UTF-8 encoded Python files.

User's Solution Review: ✅ Technically Correct The proposed fix to add encoding="utf-8" to the 4 read_text() calls is the correct solution. This explicitly specifies UTF-8 encoding instead of relying on system defaults.

Recommended Fix Strategy:

Option 1: Upstream Fix (Recommended)

• This is a Gradio library bug that should be reported upstream
• The issue affects Windows users with non-UTF-8 system encodings
• Gradio should handle UTF-8 encoding explicitly in their component_meta.py

Option 2: PraisonAI Workaround Since this is a third-party issue blocking PraisonAI users, consider adding a post-installation patch:

# In praisonai/setup/post_install.py
def patch_gradio_encoding():
    """Patch gradio component_meta.py for Windows encoding issues."""
    try:
        import gradio.component_meta
        import inspect
        import os
        
        # Get gradio component_meta.py path
        gradio_file = inspect.getfile(gradio.component_meta)
        
        # Read and patch the file if needed
        with open(gradio_file, 'r', encoding='utf-8') as f:
            content = f.read()
        
        if 'read_text()' in content and 'encoding=' not in content:
            # Apply the encoding fixes
            content = content.replace(
                'source_file.read_text()', 
                'source_file.read_text(encoding="utf-8")'
            )
            content = content.replace(
                'pyi_file.read_text()', 
                'pyi_file.read_text(encoding="utf-8")'
            )
            
            with open(gradio_file, 'w', encoding='utf-8') as f:
                f.write(content)
                
            logging.info("Applied gradio encoding patch for Windows compatibility")
            
    except Exception as e:
        logging.warning(f"Could not apply gradio encoding patch: {e}")

Final Recommendations:

1. Document the workaround in PraisonAI's troubleshooting docs for Windows users
2. Report this upstream to Gradio's GitHub repository
3. Consider implementing the post-install patch as a temporary solution
4. Pin gradio version until the upstream fix is available
5. Add environment variable option to set PYTHONIOENCODING=utf-8 in setup scripts

The user's analysis and solution are spot-on. This is a Windows-specific issue that affects any Python package importing Gradio on systems with non-UTF-8 default encodings.

---

[MervinPraison]

@Adamchanadam Can you please test it with the latest version after upgrading and let me know if this is still an issue. Thanks for you contribution.