# Data conversion code is as follows:

import vtk

def convert_by_line_id(input_path, output_path):
    points = vtk.vtkPoints()
    lines = vtk.vtkCellArray()

    line_id_array = vtk.vtkIntArray()
    line_id_array.SetName("LineID")

    # Store point IDs by line ID
    line_map = {}

    with open(input_path, 'r', encoding='utf-8') as f:
        for line_num, raw_line in enumerate(f):
            s = raw_line.strip()

            # Skip comments
            if s.startswith('#'):
                continue

            # Skip empty lines (empty lines do not affect segmentation now)
            if not s:
                continue

            parts = s.split()
            if len(parts) != 5:
                print(f"Skipping invalid format line {line_num+1}")
                continue

            try:
                x = float(parts[0])
                y = float(parts[1])
                z = float(parts[2])
                # parts[3] is unused and discarded
                line_id = int(parts[4])
            except:
                print(f"Parsing failed at line {line_num+1}, skipping")
                continue

            # Insert point
            # print(x)
            # print(y)
            # print(z)
            pt_id = points.InsertNextPoint(x, y, z)
            line_id_array.InsertNextValue(line_id)

            # Add to the point list of the corresponding line
            if line_id not in line_map:
                line_map[line_id] = []
            line_map[line_id].append(pt_id)

    # Generate VTK PolyLine for all lines
    for lid in sorted(line_map.keys()):
        pt_ids = line_map[lid]
        if len(pt_ids) < 2:
            continue

        poly_line = vtk.vtkPolyLine()
        poly_line.GetPointIds().SetNumberOfIds(len(pt_ids))
        for i, pid in enumerate(pt_ids):
            poly_line.GetPointIds().SetId(i, pid)

        lines.InsertNextCell(poly_line)

    # Build polydata
    polydata = vtk.vtkPolyData()
    polydata.SetPoints(points)
    polydata.SetLines(lines)
    polydata.GetPointData().AddArray(line_id_array)

    # Write VTP file
    writer = vtk.vtkXMLPolyDataWriter()
    writer.SetFileName(output_path)
    writer.SetInputData(polydata)
    writer.Write()

    print(f"Completed! Identified {len(line_map)} lines → {output_path}")


if __name__ == "__main__":
    # ========== Modify your file names here ==========
    INPUT  = "E:\BK1-Test2.dat"  # Replace with your raw data file name
    OUTPUT = "E:\BK1-Test2.vtp"  # Output VTP file name
    convert_by_line_id(INPUT, OUTPUT)