#!/usr/bin/env bash
set -euo pipefail

SCHEMA_DIR="${1:-schemas}"
rm -rf "$SCHEMA_DIR"
mkdir -p "$SCHEMA_DIR"

echo "==> Fetching CRDs from cluster..." >&2
kubectl get crds -o json | python3 -c "
import sys, json, os

def write_schema(schema, schema_dir, group, kind, version):
    def strip_descriptions(obj):
        if isinstance(obj, dict):
            return {k: strip_descriptions(v) for k, v in obj.items() if k != 'description'}
        if isinstance(obj, list):
            return [strip_descriptions(i) for i in obj]
        return obj

    schema = strip_descriptions(schema)
    group_dir = os.path.join(schema_dir, group)
    os.makedirs(group_dir, exist_ok=True)
    fname = f'{kind}_{version}.json'.lower()
    with open(os.path.join(group_dir, fname), 'w') as f:
        json.dump(schema, f, indent=2, sort_keys=True)
        f.write('\n')
    print(f'  Generated: {group}/{fname}', file=sys.stderr)

data = json.load(sys.stdin)
for crd in data.get('items', []):
    spec = crd.get('spec', {})
    group = spec.get('group', '')
    kind = spec.get('names', {}).get('kind', '')
    for ver in spec.get('versions', []):
        version = ver.get('name', '')
        openapi = ver.get('schema', {}).get('openAPIV3Schema', {})
        if not openapi:
            continue
        schema = dict(openapi)
        schema['\$schema'] = 'http://json-schema.org/draft-07/schema#'
        schema['type'] = 'object'
        schema.setdefault('properties', {})
        schema['properties'].setdefault('apiVersion', {'type': 'string'})
        schema['properties'].setdefault('kind', {'type': 'string'})
        schema['properties'].setdefault('metadata', {'type': 'object'})
        write_schema(schema, '$SCHEMA_DIR', group, kind, version)
"

total=$(find "$SCHEMA_DIR" -name '*.json' | wc -l)
echo "==> Schema generation complete: $total schemas in $SCHEMA_DIR" >&2