Folk Data Manipulation Example

1_pull_in_parquet.tcl

set fd [file tempfile py script.py]
puts $fd {

import pandas as pd
file_path = '~/image-label/points.parquet'
df = pd.read_parquet(file_path)
json_data = df.to_json(orient='records')
print(json_data)

}; close $fd

Retract $this claims $this has python output /anything/

On process [list apply {{this py} {
  set output [exec python3 $py]
  Assert $this claims $this has python output $output
  Assert $this claims $this has data $output
  Step
}} $this $py]

When $this has python output /o/ {
  Wish $this is labelled $o
}

Wish $this is outlined white

2_convert_parquet_to_tcl_dict.tcl

package require json

proc cleanJSON {inputStr} {
    # Use the json::json2dict function from the Tcllib JSON package
    set dict [json::json2dict $inputStr]

    return $dict
}

When $this points down with length 0.4 at /target/ {
  When $target has data /data/ {
    set cleanedData [cleanJSON $data]
    Wish $this is labelled $cleanedData
    Claim $this has cleaned data $cleanedData
  }
}

Wish $this is outlined red

3_visualize_target_data.tcl

When $this points down with length 0.4 at /target/ & /target/ has cleaned data /data/ & $this has region /r/ {
  lassign [region top $r] x y
  foreach point $data {
    set pointX [expr {$x + [dict get $point x] * 5}]
    set pointY [expr {$y + [dict get $point y] * 5}]
    
    Display::circle  $pointX $pointY 10 10 white
    Display::text [expr {$pointX + 100}] $pointY 2 [dict get $point label] 0
  }
}

README.md

Folk Data Manipulation

These files are associated with this blog post.

Note

These filesa are named *.tcl to get syntax highlighting in the gists, should be *.folk when copied to your Folk system.

generate_parquet.py

# Step 1: Import necessary libraries
import pandas as pd
import pyarrow as pa
import pyarrow.parquet as pq

# Step 2: Create a DataFrame
df = pd.DataFrame({
    'x': [1, 2, 3, 4, 5, 6, 7, 8, 9, 10],
    'y': [15, 2, 8, 10, 28, 56, 35, 88, 99, 82],
    'label': ['A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J']
})

# Step 3: Convert the DataFrame to PyArrow Table
table = pa.Table.from_pandas(df)

# Step 4: Write the PyArrow Table to Parquet
pq.write_table(table, 'points.parquet')