Plotting alignment vertical profile and cant in Bonsai

Rick_BriceRick_Brice
edited August 2025 in General

Using @Jakob 's bonsaiGraph as inspiration, I've created a simple add-in for bonsai to plot alignment vertical profile and cant.
A screenshot showing the add-in, in action, for a railway alignment.

If you want to play around with this, here is the repo. https://github.com/RickBrice/alignmentPlot
There is probably a better way to do this, but you've got to start somewhere. One big problem is the plots are static images. There is no way to edit the alignment, profile, or cant.

I'm open to suggestions to make this more functional.

steverugicarlopavtheoryshawbrunopostleMassimoJohnBedsonBimETSwalpa

Comments

  • steverugisteverugi

    @Rick_Brice
    nice feature that opens to many possibilities
    any chance we can somehow model an alignment/profile in Bonsai based on a csv or similar?

    I noticed the option in the menu but its implementation is a bit unclear for me at the moment.

    Thanks for your efforts, truly appreciated

  • civilx64civilx64
    edited September 2025

    Rick has done a great job of including documentation for create_from_csv.

    IIRC horizontal is currently limited to lines and circular arcs. Vertical is limited to constant grades and parabolic vertical curves. No support for clothoid transitions or cant ATM. PRs welcome ;-)

    steverugi
  • Rick_BriceRick_Brice

    Importing alignment from CSV is straightforward, but it has its limitations.
    Documentation is here:
    https://docs.bonsaibim.org/guides/alignment.html
    The first row in the CSV is a sequence of X, Y, Radius for a horizontal layout. The first and last radii are placeholders so just input zero.
    The second and subsequent rows in the CSV is a sequence of "Distance Along Horizontal to PVI", "PVI Elevation", "Horizontal Length of Parabolic Vertical Curve". The first and last parabolic curve length are placeholders so just input zero.
    Attached is an example input file based on and alignment defined in the US FHWA Bridge Geometry Manual.
    https://rosap.ntl.bts.gov/view/dot/62673
    Horizontal layout is limited to circular curves. Vertical layout is limited to parabolic vertical curves. No spiral transitions. No cant.

  • civilx64civilx64

    @Rick_Brice traditionally most authoring tools have handled this by placing the graphics for vertical alignment in plan view far enough away from the actual project that graphics do not overlap. Oftentimes there is some metadata left behind on the graphics to correlate the Global XY of its placement to the corresponding station/elevation of the profile. It would also need to include information on the corresponding horizontal layout / IfcCompositeCurve as well as the horizontal and vertical scales. Perhaps this could be persisted in Native IFC via GeometricRepresentationSubContext(s) ??

    carlopav
  • steverugisteverugi

    @Rick_Brice

    Attached is an example input file based on and alignment defined in the US FHWA Bridge Geometry Manual.

    I imported the .csv and this is the output (all IfcReferent fall on the same 0,0 point)

    View settings to 1000000m to prevent the cutting

    is it the same at your end?

    thanks

  • Rick_BriceRick_Brice

    @steverugi my apologies - it looks like I have broken something. I now get the same results as you.
    I will work to fix this as soon as possible.

    steverugi
  • Rick_BriceRick_Brice

    @steverugi I got the bug fixed. When reading the CSV value, I assumed the IFC project had plan angle measure in radians. The default for a bonsai project is degree. This made the direction of the start point for each segment incorrect.

    The fix should filter through the daily build process soon.
    In the meantime, change the angle measure to radian before importing the CSV file and you should get the right horizontal alignment.

    steverugi
  • Rick_BriceRick_Brice

    @civilx64

    Perhaps this could be persisted in Native IFC via GeometricRepresentationSubContext(s) ??

    Perhaps - I have some learning to do with representation contexts and subcontexts. They are a bit of a mystery to me at the moment.
    In the IFCOS api, there are two ways to get the geometry of an IfcGradientCurve, One is as a 3D curve and the other is as a 2D curve in "x = Distance Along Horizontal", "y = Elevation" coordinates. This would easily plot on the XY plane, same as the horizontal.

  • steverugisteverugi

    @Rick_Brice

    In the meantime, change the angle measure to radian before importing the CSV file and you should get the right horizontal alignment.

    Looks much better now, thanks for the quick update

    when you have a minute:
    1. Are all IfcReferent points supposed to be on 0,0,0?
    2. by looking at the webpage :


    and your csv:

    If you don't mind, I (and my colleague Jonathan, a road/rail Eng.) would personally see more practical if the csv were transposed vertically, that is, having the first 3 columns with X,Y,R for the horizontal alignment and further columns for the vertical one with D,Z,L or multiple
    like:

    I have no idea about the csv parsing implications so I guess if that is not feasible the format can be changed before being processed (or a "transpose" checkbox added to the import panel )

    Thanks again for your efforts, having alignment implemented in our IFC4x3 projects is a game-changer for infrastructure works, maybe Bonsai won't be a Civil3D competitor anytime soon but many things could be progressively introduced step by step

    ...after all it just turned 6 years old and I have a feeling that the enfant prodige of the AEC industry has a promising future ahead.

    cheers

  • steverugisteverugi
    edited September 2025

    @Rick_Brice
    here a quick script (test = OK) to rearrange the .csv from horizontal to vertical with column headers, multiple vertical profiles affect the numbering of column D, Z, L (credit to Grok)

    def transpose_alignment_csv(input_file, output_file):
    import csv
    # Read the input CSV file
    with open(input_file, 'r') as f:
        reader = csv.reader(f)
        data = list(reader)
    # Process horizontal alignment (first row)
    horizontal = data[0]
    horizontal_triplets = [horizontal[i:i+3] for i in range(0, len(horizontal), 3)]
    # Process vertical alignment (second row and beyond)
    vertical_triplets = []
    for row in data[1:]:
        triplets = [row[i:i+3] for i in range(0, len(row), 3)]
        vertical_triplets.append(triplets)
    # Create transposed data
    max_rows = max(len(horizontal_triplets), max(len(triplets) for triplets in vertical_triplets))
    transposed_data = []
    # Headers
    headers = ['X', 'Y', 'R']
    for i in range(len(vertical_triplets)):
        headers.extend([f'D{i+1}', f'Z{i+1}', f'L{i+1}'])
    transposed_data.append(headers)
     # Data rows
    for i in range(max_rows):
        row = []
        # Add horizontal alignment data
        if i < len(horizontal_triplets):
            row.extend(horizontal_triplets[i])
        else:
            row.extend(['', '', ''])
     # Add vertical alignment data for each profile
        for triplets in vertical_triplets:
            if i < len(triplets):
                row.extend(triplets[i])
            else:
                row.extend(['', '', ''])
        transposed_data.append(row)
    # Write to output CSV
    with open(output_file, 'w', newline='') as f:
        writer = csv.writer(f)
        writer.writerows(transposed_data)

    output:

  • steverugisteverugi
    edited September 2025

    @Rick_Brice
    here the "untransposed" version (Test = OK) from column back to by-row arrangement

    def untranspose_alignment_csv(input_file, output_file):
    import csv
    # Read the input CSV file
    with open(input_file, 'r') as f:
        reader = csv.reader(f)
        data = list(reader)
     # Extract headers and data
    headers = data[0]
    data_rows = data[1:]
    # Identify horizontal and vertical alignment columns
    horizontal_data = []
    vertical_data = []
    # Extract horizontal alignment (X, Y, R)
    for row in data_rows:
        xyr = row[:3]  # First three columns are X, Y, R
        if any(xyr):  # Only include non-empty rows
            horizontal_data.extend(xyr)
    # Extract vertical alignments (D, Z, L sets)
    num_vertical_profiles = (len(headers) - 3) // 3
    for i in range(num_vertical_profiles):
        profile = []
        d_col = 3 + i * 3
        z_col = d_col + 1
        l_col = d_col + 2
        for row in data_rows:
            dzl = row[d_col:d_col+3]  # Extract D, Z, L for this profile
            if any(dzl):  # Only include non-empty rows
                profile.extend(dzl)
        vertical_data.append(profile)
    # Create output data
    output_data = [horizontal_data] + vertical_data
    # Write to output CSV
    with open(output_file, 'w', newline='') as f:
        writer = csv.writer(f)
        writer.writerows(output_data)

    output of the previous .csv

    thanks again for everything
    cheers

    PS the import panel might need to factor if units are in feet or meter?

  • Rick_BriceRick_Brice

    Thanks for the suggestion. I'll take at look at the details. I'm not married to the CSV format.

    Are all IfcReferent points supposed to be on 0,0,0?

    They are not supposed to be at (0,0,0). I don't know what is going wrong. After the alignment is created from CSV and the IFC objects are mapped to blender objects, I dump the linear placement information and the fallback coordinates, and they are not at (0,0). I may need to enlist the help of advanced developers to help me figure that out.

    steverugi
  • valbeletsvalbelets

    Thanks for implementing the graph vis into it and for continuous work on IfcAlignment @Rick_Brice!

    One idea I had was to try and implement graph-based alignment components (vert profile, cant, etc) into the Graph Editor:
    https://docs.blender.org/manual/en/latest/editors/graph_editor/introduction.html#main-region
    Some of the graphs would be read-only (terrain), but some could potentially be modified using most of the curve editing functionality available there. I bet current frame location can be synchronised with some element moving along the alignment (Vertical plane for example).
    Frames == Chainages.
    It can also visualise parameters change along the alignment (road widenings etc).

    Happy to go into a 🐇 hole or two to explore it further. Any pointers would be much appreciated.

  • Rick_BriceRick_Brice

    @valbelets I was thinking the same thing but really have no idea where to begin.
    I can give you some pointers about the alignment geometry and segments. It's probably better to answer specific questions when you have them.

  • civilx64civilx64

    You could load the vertical layout from the business logic and then use the center circle to edit the PVI station and elevation, with the other two circles corresponding to the beginning and end of the vertical curve. After the user finishes editing the business logic, the representation is re-calculated from it. The UI would maybe look something like this:

    theoryshawwalpa
Sign In or Register to comment.