CSV export feature request.

[Matt-V50]

Thanks for this tool, I also want to export those values into csv file sometimes.
Would you like to add that feature? Or, will you have the relative plan for that?

[kolibril13]

Great idea!
here's already a script that could be used as a starting point.
https://blender.stackexchange.com/questions/318588/how-to-export-a-blender-spreadsheet-information-into-a-csv-and-txt-file
I'll have a look how this could be added to the extension!

[kolibril13]

here's a script that will export the spreadsheet into a data table:

import bpy
import databpy as db
import polars as pl

obj = bpy.data.objects["Cube"]
evaluated_obj = obj.evaluated_get(bpy.context.evaluated_depsgraph_get())
mesh = evaluated_obj.to_mesh()

# Collect all attribute data
attribute_data = {}
for attr in mesh.attributes:
    if attr.name not in {'sharp_face', 'UVMap'} and not attr.name.startswith('.'):
        a = db.named_attribute(evaluated_obj, attr.name)
        attribute_data[attr.name] = a

# Create polars DataFrame
df = pl.DataFrame(attribute_data)

# Sort columns so "position" is first
if "position" in df.columns:
    column_order = ["position"] + [col for col in df.columns if col != "position"]
    df = df.select(column_order)

print(df)

shape: (8, 3)
┌────────────────────┬───────┬───────┐
│ position           ┆ hello ┆ world │
│ ---                ┆ ---   ┆ ---   │
│ array[f64, 3]      ┆ i64   ┆ bool  │
╞════════════════════╪═══════╪═══════╡
│ [-1.0, -1.0, -1.0] ┆ 42    ┆ true  │
│ [-1.0, -1.0, 1.0]  ┆ 42    ┆ true  │
│ [-1.0, 1.0, -1.0]  ┆ 42    ┆ true  │
│ [-1.0, 1.0, 1.0]   ┆ 42    ┆ true  │
│ [1.0, -1.0, -1.0]  ┆ 42    ┆ true  │
│ [1.0, -1.0, 1.0]   ┆ 42    ┆ true  │
│ [1.0, 1.0, -1.0]   ┆ 42    ┆ true  │
│ [1.0, 1.0, 1.0]    ┆ 42    ┆ true  │
└────────────────────┴───────┴───────┘

now is the question:
How do we want to process this data frame further, so that it can be stored as a csv file?
All vectors (e.g. the position [1.0, 1.0, 1.0] ) could to be converted, e.g. in x,y,z columns in the polars dataframe before the export

import tempfile
import os

# Write to temporary file
temp_dir = tempfile.gettempdir()
temp_file = os.path.join(temp_dir, "mesh_attributes.csv")
df.write_csv(temp_file)
print(f"CSV written to: {temp_file}")
print(df)

ComputeError                              Traceback (most recent call last)
/var/folders/0n/w5m51rrn71db4cvkds08wdg80000gn/T/ipykernel_2263/3097064528.py in ?()
      3 
      4 # Write to temporary file
      5 temp_dir = tempfile.gettempdir()
      6 temp_file = os.path.join(temp_dir, "mesh_attributes.csv")
----> 7 df.write_csv(temp_file)
      8 print(f"CSV written to: {temp_file}")
      9 print(df)

~/Library/Application Support/Blender/4.4/extensions/.local/lib/python3.11/site-packages/polars/dataframe/frame.py in ?(***failed resolving arguments***)
   2990         else:
   2991             # Handle empty dict input
   2992             storage_options = None
   2993 
-> 2994         self._df.write_csv(
   2995             file,
   2996             include_bom,
   2997             include_header,

ComputeError: datatype array[f64, 3] cannot be written to CSV

Consider using JSON or a binary format.

[kolibril13]

another idea:

import bpy
import databpy as db
import polars as pl

obj = bpy.data.objects["Cube"]
evaluated_obj = db.evaluate_object(obj)
mesh = evaluated_obj.to_mesh()

# Collect all attribute data
attribute_data = {}
for attr in mesh.attributes:
    if attr.name not in {'sharp_face', 'UVMap'} and not attr.name.startswith('.'):
        a = db.named_attribute(evaluated_obj, attr.name)
        attribute_data[attr.name] = a

# Create polars DataFrame
df = pl.DataFrame(attribute_data)

# Sort columns so "position" is first
if "position" in df.columns:
    column_order = ["position"] + [col for col in df.columns if col != "position"]
    df = df.select(column_order)

print(df)

shape: (8, 4)
┌────────────────────┬───────┬──────────┬─────────────────────────────────┐
│ position           ┆ hello ┆ world!!! ┆ advanced                        │
│ ---                ┆ ---   ┆ ---      ┆ ---                             │
│ array[f64, 3]      ┆ i64   ┆ bool     ┆ array[f64, 4]                   │
╞════════════════════╪═══════╪══════════╪═════════════════════════════════╡
│ [-1.0, -1.0, -1.0] ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [-1.0, -1.0, 1.0]  ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [-1.0, 1.0, -1.0]  ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [-1.0, 1.0, 1.0]   ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [1.0, -1.0, -1.0]  ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [1.0, -1.0, 1.0]   ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [1.0, 1.0, -1.0]   ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
│ [1.0, 1.0, 1.0]    ┆ 42    ┆ true     ┆ [0.992774, 0.11234, … 0.016647… │
└────────────────────┴───────┴──────────┴─────────────────────────────────┘

# Check dtypes and expand array columns
dtypes = df.dtypes
print("Column dtypes:", dtypes)

# Get columns that are arrays and expand them
array_columns = []
expanded_df = df.clone()

for i, dtype in enumerate(dtypes):
    col_name = df.columns[i]
    if str(dtype).startswith('Array'):
        array_columns.append(col_name)
        print(f"Array column found: {col_name} with dtype {dtype}")
        
        # Get the array length from the first non-null value
        first_array = expanded_df.select(pl.col(col_name)).item(0, 0)
        array_length = len(first_array)
        
        # Expand array into separate columns with indexed names
        for j in range(array_length):
            expanded_df = expanded_df.with_columns([
                pl.col(col_name).arr.get(j).alias(f"{col_name}{j+1}")
            ])
        # Drop the original array column
        expanded_df = expanded_df.drop(col_name)

# Reorder columns to place position columns first
position_columns = [col for col in expanded_df.columns if col.startswith('position')]
other_columns = [col for col in expanded_df.columns if not col.startswith('position')]
column_order = position_columns + other_columns
expanded_df = expanded_df.select(column_order)

print("Array columns:", array_columns)
print("DataFrame after expansion:")
print(expanded_df)

Column dtypes: [Array(Float64, shape=(3,)), Int64, Boolean, Array(Float64, shape=(4,))]
Array column found: position with dtype Array(Float64, shape=(3,))
Array column found: advanced with dtype Array(Float64, shape=(4,))
Array columns: ['position', 'advanced']
DataFrame after expansion:
shape: (8, 9)
┌───────────┬───────────┬───────────┬───────┬───┬───────────┬───────────┬───────────┬───────────┐
│ position1 ┆ position2 ┆ position3 ┆ hello ┆ … ┆ advanced1 ┆ advanced2 ┆ advanced3 ┆ advanced4 │
│ ---       ┆ ---       ┆ ---       ┆ ---   ┆   ┆ ---       ┆ ---       ┆ ---       ┆ ---       │
│ f64       ┆ f64       ┆ f64       ┆ i64   ┆   ┆ f64       ┆ f64       ┆ f64       ┆ f64       │
╞═══════════╪═══════════╪═══════════╪═══════╪═══╪═══════════╪═══════════╪═══════════╪═══════════╡
│ -1.0      ┆ -1.0      ┆ -1.0      ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ -1.0      ┆ -1.0      ┆ 1.0       ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ -1.0      ┆ 1.0       ┆ -1.0      ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ -1.0      ┆ 1.0       ┆ 1.0       ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ 1.0       ┆ -1.0      ┆ -1.0      ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ 1.0       ┆ -1.0      ┆ 1.0       ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ 1.0       ┆ 1.0       ┆ -1.0      ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
│ 1.0       ┆ 1.0       ┆ 1.0       ┆ 42    ┆ … ┆ 0.992774  ┆ 0.11234   ┆ 0.038769  ┆ 0.016647  │
└───────────┴───────────┴───────────┴───────┴───┴───────────┴───────────┴───────────┴───────────┘

import tempfile
import os

# Write to temporary file
temp_dir = tempfile.gettempdir()
temp_file = os.path.join(temp_dir, "mesh_attributes.csv")
expanded_df.write_csv(temp_file)
print(f"CSV written to: {temp_file}")
print(expanded_df)

[kolibril13]

@Matt-V50 , feel free to have a look at the UI at #28

[Matt-V50]

Wow, looks great, that's exactly what I want. Can't wait to use it. Thanks for your work.