crd-extractor.sh

#!/usr/bin/env bash

# Check if python3 is installed
if ! command -v python3 &> /dev/null; then
    printf "python3 is required for this utility, and is not installed on your machine"
    printf "please visit https://www.python.org/downloads/ to install it"
    exit 1
fi
# Check if kubectl is installed
if ! command -v kubectl &> /dev/null; then
    printf "kubectl is required for this utility, and is not installed on your machine"
    printf "please visit https://kubernetes.io/docs/tasks/tools/#kubectl to install it"
    exit 1
fi

# Check if the pyyaml module is installed
if ! echo 'import yaml' | python3 &> /dev/null; then
    printf "the python3 module 'yaml' is required, and is not installed on your machine.\n"

    while true; do
        read -p "Do you wish to install this program? (y/n) " yn
        case $yn in
            [Yy] ) pip3 install pyyaml; break;;
            "" ) pip3 install pyyaml; break;;
            [Nn] ) echo "Exiting..."; exit;;
            * ) echo "Please answer 'y' (yes) or 'n' (no).";;
        esac
    done
fi

# Create temp folder for CRDs
TMP_CRD_DIR=$HOME/.datree/crds
mkdir -p $TMP_CRD_DIR

# Create final schemas directory
SCHEMAS_DIR=$HOME/.datree/crdSchemas
mkdir -p $SCHEMAS_DIR
cd $SCHEMAS_DIR

# Create array to store CRD kinds and groups
ORGANIZE_BY_GROUP=true
declare -A CRD_GROUPS 2>/dev/null
if [ $? -ne 0 ]; then
    # Array creation failed, signal to skip organization by group
    ORGANIZE_BY_GROUP=false
fi

# Extract CRDs from cluster
NUM_OF_CRDS=0
while read -r crd 
do
    resourceKind=${crd%% *}
    kubectl get crds "$resourceKind" -o yaml > "$TMP_CRD_DIR/"$resourceKind".yaml" 2>&1 
    
    # Get singular name from crd
    singularNameValue=$(grep singular: "$TMP_CRD_DIR/"$resourceKind".yaml" -m 1)
    singularName=${singularNameValue##* }
    
    # Get group
    resourceGroup=${resourceKind#*.}

    # Save name and group for later directory organization
    CRD_GROUPS["$singularName"]="$resourceGroup"
    
    let NUM_OF_CRDS++
done < <(kubectl get crds 2>&1 | sed -n '/NAME/,$p' | tail -n +2)

# If no CRDs exist in the cluster, exit
if [ $NUM_OF_CRDS == 0 ]; then
    printf "No CRDs found in the cluster, exiting...\n"
    exit 0
fi

# Download converter script
curl https://raw.githubusercontent.com/yannh/kubeconform/master/scripts/openapi2jsonschema.py --output $TMP_CRD_DIR/openapi2jsonschema.py 2>/dev/null

# Convert crds to jsonSchema
python3 $TMP_CRD_DIR/openapi2jsonschema.py $TMP_CRD_DIR/*.yaml
conversionResult=$?

# Copy and rename files to support kubeval
rm -rf $SCHEMAS_DIR/master-standalone
mkdir -p $SCHEMAS_DIR/master-standalone
cp $SCHEMAS_DIR/*.json $SCHEMAS_DIR/master-standalone
find $SCHEMAS_DIR/master-standalone -name '*json' -exec bash -c ' mv -f $0 ${0/\_/-stable-}' {} \;

# Organize schemas by group
if [ $ORGANIZE_BY_GROUP == true ]; then
    for schema in $SCHEMAS_DIR/*.json
    do
    crdFileName=$(basename $schema .json)
    crdKind=${crdFileName%%_*}
    crdGroup=${CRD_GROUPS[$crdKind]}
    mkdir -p $crdGroup
    mv $schema ./$crdGroup
    done
fi

CYAN='\033[0;36m'
GREEN='\033[0;32m'
NC='\033[0m' # No Color

if [ $conversionResult == 0 ]; then
    printf "${GREEN}Successfully converted $NUM_OF_CRDS CRDs to JSON schema${NC}\n"
    
    printf "\nTo validate a CR using various tools, run the relevant command:\n"
    printf "\n- ${CYAN}datree:${NC}\n\$ datree test /path/to/file\n"
    printf "\n- ${CYAN}kubeconform:${NC}\n\$ kubeconform -summary -output json -schema-location default -schema-location '$HOME/.datree/crdSchemas/{{ .ResourceKind }}_{{ .ResourceAPIVersion }}.json' /path/to/file\n"
    printf "\n- ${CYAN}kubeval:${NC}\n\$ kubeval --additional-schema-locations file:\"$HOME/.datree/crdSchemas\" /path/to/file\n\n"
fi

rm -rf $TMP_CRD_DIR