Using grep to extract all BPMN cam-form-variables within HTML files

27.10.2023 — Camunda, BPMN, Bash — 1 min read

I came across a problem when documenting some process: I needed to provide a data dictionary for all of the variables I've used.

To solve this issue without taking to much time, I've developed a bash to extract them.

It is very simple but might be handy:

#!/bin/bash

# Store the original working directory
original_directory=$(pwd)
output=$original_directory/output/everything.txt

# Clear the output file
cat /dev/null > "$output"

while read line; do
    # Extract the value of the "res" variable from the 'formKey' pattern in the line
    res=$(echo "$line" | grep -oP '(?<=formKey="embedded:app:forms/).*')

    # Use the full path to the "forms" directory
    forms_directory="$original_directory/src/main/webapp/forms/"

    # Change to the "forms" directory
    cd "$forms_directory" || (echo "Error: Directory '$forms_directory' not found." && exit 1)

    # Use the "find" command to locate files that match the specified pattern
    find . -type f -name "$res" -printf '%P\n' | while read file; do
        echo "Reading file: $file"
        cat "$file" | grep -oP '(?<=cam-variable-name=").*?(?=")' | grep -v 'readonly' | while read variables; do
            # Create the output directory if it doesn't exist
            mkdir -p "$original_directory/output"
            chmod 777 "$original_directory/output"
            
            # Create the output file and append data to it
            touch "$output"
            if ! grep -q ".*$variables.*" "$output"; then
                echo "$variables" >> "$output" || (echo "Error: Failed to write to the '$output' file." && exit 1)
            fi            
            # Return to the original working directory
            cd "$original_directory" || (echo "Error: Failed to return to the original directory." && exit 1)
        done
    done

    # Restore the working directory for the next iteration
    cd "$original_directory" || (echo "Error: Failed to return to the original directory." && exit 1)

done < forms.txt

Also on Github