public DataTable ExtractTable( IFormat format )
{
PathSeriesFormat pathSeriesFormat = format as PathSeriesFormat;
PathTableFormat pathTableFormat = format as PathTableFormat;
if ( pathSeriesFormat != null )
{
var htmlSettings = new HtmlExtractionSettings();
htmlSettings.ExtractLinkUrl = pathSeriesFormat.ExtractLinkUrl;
var result = Content.ExtractTable( HtmlPath.Parse( pathSeriesFormat.Path ), pathSeriesFormat.ToExtractionSettings(), htmlSettings );
if ( !result.Success )
{
throw new Exception( "Failed to extract table from document: " + result.FailureReason );
}
return pathSeriesFormat.ToFormattedTable( result.Value );
}
else if ( pathTableFormat != null )
{
var result = Content.ExtractTable( HtmlPath.Parse( pathTableFormat.Path ), true );
if ( !result.Success )
{
throw new Exception( "Failed to extract table from document: " + result.FailureReason );
}
return pathTableFormat.ToFormattedTable( result.Value );
}
else if ( format is PathSingleValueFormat )
{
var f = (PathSingleValueFormat)format;
var str = Content.GetTextByPath( HtmlPath.Parse( f.Path ) );
var value = f.ValueFormat.Convert( str );
// XXX: this is really ugly - i have to create a table just to satisfy the interface :(
return CreateTableForScalar( f.ValueFormat.Type, value );
}
else
{
throw new NotSupportedException( "Format not supported for Html documents: " + format.GetType() );
}
}
/// <summary>
/// Extracts a cell or a series of the html table the given path is pointing to.
/// If the path points to the table element itself instead of a cell the whole table will be
/// extracted.
/// The series to be extracted is always arranged in a column (independed of the original layout
/// in the html table). The first column contains the values, the second the series header (if
/// any defined). The series name is stored in the ColumnName of the first column.
/// <remarks>
/// Returns null if table not found by path. Currently we cannot handle thead
/// and tfoot.
/// </remarks>
/// </summary>
/// <param name="path">points to a cell or the body of a table (pointers to TR elements are invalid)</param>
/// <param name="doc">the HTML document</param>
/// <param name="htmlSettings">the HTML settings used to configure the extraction process</param>
/// <param name="tableSettings">the table specific configuration</param>
public static FallibleActionResult <DataTable> ExtractTable(this IHtmlDocument doc, HtmlPath path, TableExtractionSettings tableSettings, HtmlExtractionSettings htmlSettings)
{
if (!path.PointsToTable && !path.PointsToTableCell)
{
throw new InvalidExpressionException("Path neither points to table nor to cell");
}
FallibleActionResult <DataTable> result = ExtractTable(doc, path, !htmlSettings.ExtractLinkUrl);
if (!result.Success)
{
// pass throu failure result
return(result);
}
// path points to whole table => return whole table
if (path.PointsToTable)
{
return(result);
}
// get the x,y position of the cell the path is pointing to
Point cellCoords = path.GetTableCellPosition();
if (cellCoords.X < 0 || cellCoords.Y < 0)
{
throw new InvalidExpressionException("Path expression corrupt: cell position in table could not be calculated");
}
// get the value of the raw cell. extract the link url if configured.
Func <object, object> GetValue = e =>
{
if (htmlSettings.ExtractLinkUrl)
{
return(((IHtmlElement)e).FirstLinkOrInnerText());
}
else
{
return(e);
}
};
var t = result.Value.ExtractSeries(cellCoords, GetValue, tableSettings);
if (t == null)
{
return(FallibleActionResult <DataTable> .CreateFailureResult("Could not extract series specified"));
}
return(FallibleActionResult <DataTable> .CreateSuccessResult(t));
}
