Naive approach to finding unused files, classes, and functions

find-unused.sh

#!/bin/bash

if [ -z "$1" ]; then
  echo "Usage: $0 <dir>"
  exit 1
fi

DIR="$(realpath $1)"

find_unused_classes()
{
  echo "Finding Suspected Unused Classes"

  LOG_LIST=class.log
  LOG_UNUSED=class.unused.log

  if [ ! -f "$LOG_LIST" ]; then
    grep -IiP '^\s{0,}class\s{1,}[a-zA-Z0-9]+' $DIR -r --exclude-dir=vendor \
      --include="*.php" --include="*.ctp" \
      > $LOG_LIST
  fi

  truncate -s0 "$LOG_UNUSED"

  while read LINE; do
    CLASS=$(echo "$LINE" | sed -re 's/^[^:]+:\s{0,}class\s{1,}([^ ]+)(\s{1,}.*)?$/\1/')
    FILE=$(echo "$LINE" | sed -e 's_\(.*\):.*_\1_')
    REFERENCES=$(grep -m 1 -Ii "$CLASS" $DIR -r --include="*.php" --include="*.ctp" | grep -iP ':\s{0,}class\s{1,}[a-zA-Z0-9]+' -v | head -n 1)

    # If grep returns with no matches, output the file as potentially unused.
    if [ -z "$REFERENCES" ]; then
      echo "$CLASS in $FILE" >> "$LOG_UNUSED"
    fi
  done < $LOG_LIST

}

find_unused_files()
{
  echo "Finding Suspected Unused Files"

  LOG_LIST=files.log
  LOG_UNUSED=files.unused.log

  if [ ! -f "$LOG_LIST" ]; then
    find $BASE_DIR -type f -regextype awk \
      -regex '.*\.(php|ctp)' \
      -not -iregex '.*/(vendor|\.hg|\.git)/.*' \
      | sort > $LOG_LIST
  fi

  truncate -s0 "$LOG_UNUSED"

  while read LINE; do
    BASE=$(basename "$LINE")
    FILE=$(echo "$BASE" | sed -r 's/([^.]+).*/\1/')

    REFERENCES=$(grep -m 1 -Ii "$FILE" $DIR -r --include="*.php" --include="*.ctp" --exclude-dir=vendor --exclude="$BASE")

    # If grep returns with no matches, output the file as potentially unused.
    if [ -z "$REFERENCES" ]; then
      echo "$LINE" >> "$LOG_UNUSED"
    fi

  done < $LOG_LIST
}

find_unused_functions()
{
  echo "Finding Suspected Unused Functions"

  LOG_LIST=functions.log
  LOG_UNUSED=functions.unused.log

  if [ ! -f "$LOG_LIST" ]; then
      grep -IiP '\s{0,}function\s{1,}[a-zA-Z0-9]+' $DIR -r --exclude-dir=vendor \
      --include="*.php" --include="*.ctp" \
      > $LOG_LIST
  fi

  truncate -s0 "$LOG_UNUSED"

  while read LINE; do
    FUNCTION=$(echo "$LINE" | sed -re 's/.*\s{0,}function\s{1,}([^ (]+).*([ (])?/\1/')
    FILE=$(echo "$LINE" | sed -re 's/^([^:]+):.*/\1/')
    REFERENCES=$(grep -Ii "$FUNCTION" $DIR -r --exclude-dir=vendor --include="*.php" --include="*.ctp" | grep -iP 'function\s{1,}[a-zA-Z0-9]+' -v | head -n 1)

    # If grep returns with no matches, output the file as potentially unused.
    if [ -z "$REFERENCES" ]; then
      echo "$FUNCTION in $FILE" >> "$LOG_UNUSED"
    fi
  done < $LOG_LIST
}

main()
{
  find_unused_classes
  find_unused_files
  find_unused_functions
}

main

This is a cleaned up version of a script I wrote and used frequently at the first few jobs I had in the tech industry. Since PHP 7, there have been multiple static analysis tools released that perform a better job of identifying unused features.