docs: unify dataframe documentation (#2)

kosiew · kosiew · commit 820f5ae0f78f · 2025-06-20T18:45:16.000+08:00
diff --git a/docs/source/api/dataframe.rst b/docs/source/api/dataframe.rst
@@ -150,238 +150,3 @@ To materialize the results of your DataFrame operations:
     # Count rows
     count = df.count()
 
-HTML Rendering in Jupyter
--------------------------
-
-When working in Jupyter notebooks or other environments that support rich HTML display, 
-DataFusion DataFrames automatically render as nicely formatted HTML tables. This functionality
-is provided by the ``_repr_html_`` method, which is automatically called by Jupyter.
-
-Basic HTML Rendering
-~~~~~~~~~~~~~~~~~~~~
-
-In a Jupyter environment, simply displaying a DataFrame object will trigger HTML rendering:
-
-.. code-block:: python
-
-    # Will display as HTML table in Jupyter
-    df
-
-    # Explicit display also uses HTML rendering
-    display(df)
-
-HTML Rendering Customization
-----------------------------
-
-DataFusion provides extensive customization options for HTML table rendering through the
-``datafusion.html_formatter`` module.
-
-Configuring the HTML Formatter
-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
-
-You can customize how DataFrames are rendered by configuring the formatter:
-
-.. code-block:: python
-
-    from datafusion.html_formatter import configure_formatter
-    
-    configure_formatter(
-        max_cell_length=30,              # Maximum length of cell content before truncation
-        max_width=800,                   # Maximum width of table in pixels
-        max_height=400,                  # Maximum height of table in pixels
-        max_memory_bytes=2 * 1024 * 1024,# Maximum memory used for rendering (2MB)
-        min_rows_display=10,             # Minimum rows to display
-        repr_rows=20,                    # Number of rows to display in representation
-        enable_cell_expansion=True,      # Allow cells to be expandable on click
-        custom_css=None,                 # Custom CSS to apply
-        show_truncation_message=True,    # Show message when data is truncated
-        style_provider=None,             # Custom style provider class
-        use_shared_styles=True           # Share styles across tables to reduce duplication
-    )
-
-Custom Style Providers
-~~~~~~~~~~~~~~~~~~~~~~
-
-For advanced styling needs, you can create a custom style provider class:
-
-.. code-block:: python
-
-    from datafusion.html_formatter import configure_formatter
-    
-    class CustomStyleProvider:
-        def get_cell_style(self) -> str:
-            return "background-color: #f5f5f5; color: #333; padding: 8px; border: 1px solid #ddd;"
-    
-        def get_header_style(self) -> str:
-            return "background-color: #4285f4; color: white; font-weight: bold; padding: 10px;"
-    
-    # Apply custom styling
-    configure_formatter(style_provider=CustomStyleProvider())
-
-Custom Type Formatters
-~~~~~~~~~~~~~~~~~~~~~~
-
-You can register custom formatters for specific data types:
-
-.. code-block:: python
-
-    from datafusion.html_formatter import get_formatter
-    
-    formatter = get_formatter()
-    
-    # Format integers with color based on value
-    def format_int(value):
-        return f'<span style="color: {"red" if value > 100 else "blue"}">{value}</span>'
-    
-    formatter.register_formatter(int, format_int)
-    
-    # Format date values
-    def format_date(value):
-        return f'<span class="date-value">{value.isoformat()}</span>'
-    
-    formatter.register_formatter(datetime.date, format_date)
-
-Custom Cell Builders
-~~~~~~~~~~~~~~~~~~~~
-
-For complete control over cell rendering:
-
-.. code-block:: python
-
-    formatter = get_formatter()
-    
-    def custom_cell_builder(value, row, col, table_id):
-        try:
-            num_value = float(value)
-            if num_value > 0:  # Positive values get green
-                return f'<td style="background-color: #d9f0d3">{value}</td>'
-            if num_value < 0:  # Negative values get red
-                return f'<td style="background-color: #f0d3d3">{value}</td>'
-        except (ValueError, TypeError):
-            pass
-        
-        # Default styling for non-numeric or zero values
-        return f'<td style="border: 1px solid #ddd">{value}</td>'
-    
-    formatter.set_custom_cell_builder(custom_cell_builder)
-
-Custom Header Builders
-~~~~~~~~~~~~~~~~~~~~~~
-
-Similarly, you can customize the rendering of table headers:
-
-.. code-block:: python
-
-    def custom_header_builder(field):
-        tooltip = f"Type: {field.type}"
-        return f'<th style="background-color: #333; color: white" title="{tooltip}">{field.name}</th>'
-    
-    formatter.set_custom_header_builder(custom_header_builder)
-
-Managing Formatter State
------------------------~
-
-The HTML formatter maintains global state that can be managed:
-
-.. code-block:: python
-
-    from datafusion.html_formatter import reset_formatter, reset_styles_loaded_state, get_formatter
-    
-    # Reset the formatter to default settings
-    reset_formatter()
-    
-    # Reset only the styles loaded state (useful when styles were loaded but need reloading)
-    reset_styles_loaded_state()
-    
-    # Get the current formatter instance to make changes
-    formatter = get_formatter()
-
-Advanced Example: Dashboard-Style Formatting
-------------------------------------------~~
-
-This example shows how to create a dashboard-like styling for your DataFrames:
-
-.. code-block:: python
-
-    from datafusion.html_formatter import configure_formatter, get_formatter
-    
-    # Define custom CSS
-    custom_css = """
-    .datafusion-table {
-        font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
-        border-collapse: collapse;
-        width: 100%;
-        box-shadow: 0 2px 3px rgba(0,0,0,0.1);
-    }
-    .datafusion-table th {
-        position: sticky;
-        top: 0;
-        z-index: 10;
-    }
-    .datafusion-table tr:hover td {
-        background-color: #f1f7fa !important;
-    }
-    .datafusion-table .numeric-positive {
-        color: #0a7c00;
-    }
-    .datafusion-table .numeric-negative {
-        color: #d13438;
-    }
-    """
-    
-    class DashboardStyleProvider:
-        def get_cell_style(self) -> str:
-            return "padding: 8px 12px; border-bottom: 1px solid #e0e0e0;"
-        
-        def get_header_style(self) -> str:
-            return ("background-color: #0078d4; color: white; font-weight: 600; "
-                    "padding: 12px; text-align: left; border-bottom: 2px solid #005a9e;")
-    
-    # Apply configuration
-    configure_formatter(
-        max_height=500,
-        enable_cell_expansion=True,
-        custom_css=custom_css,
-        style_provider=DashboardStyleProvider(),
-        max_cell_length=50
-    )
-    
-    # Add custom formatters for numbers
-    formatter = get_formatter()
-    
-    def format_number(value):
-        try:
-            num = float(value)
-            cls = "numeric-positive" if num > 0 else "numeric-negative" if num < 0 else ""
-            return f'<span class="{cls}">{value:,}</span>' if cls else f'{value:,}'
-        except (ValueError, TypeError):
-            return str(value)
-    
-    formatter.register_formatter(int, format_number)
-    formatter.register_formatter(float, format_number)
-
-Best Practices
---------------
-
-1. **Memory Management**: For large datasets, use ``max_memory_bytes`` to limit memory usage.
-
-2. **Responsive Design**: Set reasonable ``max_width`` and ``max_height`` values to ensure tables display well on different screens.
-
-3. **Style Optimization**: Use ``use_shared_styles=True`` to avoid duplicate style definitions when displaying multiple tables.
-
-4. **Reset When Needed**: Call ``reset_formatter()`` when you want to start fresh with default settings.
-
-5. **Cell Expansion**: Use ``enable_cell_expansion=True`` when cells might contain longer content that users may want to see in full.
-
-Additional Resources
---------------------
-
-* :doc:`../user-guide/dataframe` - Complete guide to using DataFrames
-* :doc:`../user-guide/io/index` - I/O Guide for reading data from various sources
-* :doc:`../user-guide/data-sources` - Comprehensive data sources guide
-* :ref:`io_csv` - CSV file reading
-* :ref:`io_parquet` - Parquet file reading  
-* :ref:`io_json` - JSON file reading
-* :ref:`io_avro` - Avro file reading
-* :ref:`io_custom_table_provider` - Custom table providers
-* `API Reference <https://arrow.apache.org/datafusion-python/api/index.html>`_ - Full API reference
diff --git a/docs/source/api/index.rst b/docs/source/api/index.rst
@@ -23,5 +23,3 @@ This section provides detailed API documentation for the DataFusion Python libra
 
 .. toctree::
    :maxdepth: 2
-   
-   dataframe
diff --git a/docs/source/index.rst b/docs/source/index.rst
@@ -72,7 +72,7 @@ Example
    user-guide/introduction
    user-guide/basics
    user-guide/data-sources
-   user-guide/dataframe
+   user-guide/dataframe/index
    user-guide/common-operations/index
    user-guide/io/index
    user-guide/configuration
@@ -88,10 +88,3 @@ Example
    contributor-guide/introduction
    contributor-guide/ffi
 
-.. _toc.api:
-.. toctree::
-   :hidden:
-   :maxdepth: 1
-   :caption: API
-   
-   api/index
diff --git a/docs/source/user-guide/dataframe/index.rst b/docs/source/user-guide/dataframe/index.rst
@@ -215,4 +215,9 @@ You can control how much data is displayed and how much memory is used for rende
         repr_rows=20                       # Show 20 rows in __repr__ output
     )
 
-These parameters help balance comprehensive data display against performance considerations.
+These parameters help balance comprehensive data display against performance considerations.
+
+.. toctree::
+   :maxdepth: 1
+
+   rendering
diff --git a/docs/source/user-guide/dataframe/rendering.rst b/docs/source/user-guide/dataframe/rendering.rst
