fix(pb-init): rewrite ensure_field with pure sh/awk, fix HTTP status parsing

The previous ensure_fields helper used python3 which is not available in
alpine:3.19, causing exit 127 in the cloud init container.

Replaced with ensure_field (per-field, no python/jq) that uses only
busybox sh, wget, sed, and awk. Also fixed the HTTP status grep pattern
in create_collection and ensure_field to match only the response header
line (^  HTTP/) instead of the wget error line, eliminating the spurious
'unexpected status server' warnings.

scripts/pb-init.sh

@@ -47,7 +47,7 @@ create_collection() {
     --header="Content-Type: application/json" \
     --header="Authorization: Bearer $TOKEN" \
     --post-data="$BODY" \
-    "$PB_URL/api/collections" 2>&1 | grep "HTTP/" | tail -1 | awk '{print $2}')
+    "$PB_URL/api/collections" 2>&1 | grep "^  HTTP/" | awk '{print $2}')
   case "$STATUS" in
     200|201) log "created collection: $NAME" ;;
     400|422) log "collection already exists (skipped): $NAME" ;;
@@ -55,56 +55,49 @@ create_collection() {
   esac
 }
 
-# Patch a collection by name: adds any fields listed in FIELDS_JSON that are
+# ensure_field COLLECTION FIELD_NAME FIELD_TYPE
-# missing from the current schema.  This is idempotent and safe to re-run.
+#
-ensure_fields() {
+# Checks whether FIELD_NAME exists in COLLECTION's schema.  If it is missing,
-  NAME="$1"
+# sends a PATCH with the full current fields list plus the new field appended.
-  FIELDS_JSON="$2"   # JSON array of field objects to guarantee exist
+# Uses only busybox sh + wget + sed/awk — no python/jq required.
+ensure_field() {
+  COLL="$1"
+  FIELD_NAME="$2"
+  FIELD_TYPE="$3"
 
-  # Fetch current schema
+  SCHEMA=$(wget -qO- \
-  CURRENT=$(wget -qO- \
     --header="Authorization: Bearer $TOKEN" \
-    "$PB_URL/api/collections/$NAME" 2>/dev/null)
+    "$PB_URL/api/collections/$COLL" 2>/dev/null)
 
-  COLLECTION_ID=$(echo "$CURRENT" | sed 's/.*"id":"\([^"]*\)".*/\1/')
+  # Check if the field already exists (look for "name":"<FIELD_NAME>" in the fields array)
-  if [ -z "$COLLECTION_ID" ] || [ "$COLLECTION_ID" = "$CURRENT" ]; then
+  if echo "$SCHEMA" | grep -q "\"name\":\"$FIELD_NAME\""; then
-    log "WARNING: could not get id for collection $NAME — skipping ensure_fields"
+    log "field $COLL.$FIELD_NAME already exists — skipping"
     return
   fi
 
-  # Build merged fields list via a small python snippet
+  COLLECTION_ID=$(echo "$SCHEMA" | sed 's/.*"id":"\([^"]*\)".*/\1/')
-  MERGED=$(python3 - "$CURRENT" "$FIELDS_JSON" <<'PYEOF'
+  if [ -z "$COLLECTION_ID" ] || [ "$COLLECTION_ID" = "$SCHEMA" ]; then
-import sys, json
+    log "WARNING: could not get id for collection $COLL — skipping ensure_field"
-current = json.loads(sys.argv[1])
-desired = json.loads(sys.argv[2])
-existing_names = {f["name"] for f in current.get("fields", [])}
-all_fields = list(current.get("fields", []))
-added = []
-for f in desired:
-    if f["name"] not in existing_names:
-        all_fields.append(f)
-        added.append(f["name"])
-print(json.dumps({"fields": all_fields, "_added": added}))
-PYEOF
-)
-
-  ADDED=$(echo "$MERGED" | python3 -c 'import sys,json; d=json.load(sys.stdin); print(",".join(d.get("_added",[])))' 2>/dev/null)
-  if [ -z "$ADDED" ]; then
-    log "collection $NAME schema up-to-date"
     return
   fi
 
-  PATCH_BODY=$(echo "$MERGED" | python3 -c 'import sys,json; d=json.load(sys.stdin); d.pop("_added",None); print(json.dumps(d))' 2>/dev/null)
+  # Extract current fields array (everything between the outermost [ ] of "fields":[...])
+  # and append the new field object before the closing bracket.
+  CURRENT_FIELDS=$(echo "$SCHEMA" | sed 's/.*"fields":\(\[.*\]\).*/\1/')
+  # Strip the trailing ] and append the new field
+  TRIMMED=$(echo "$CURRENT_FIELDS" | sed 's/]$//')
+  NEW_FIELDS="${TRIMMED},{\"name\":\"${FIELD_NAME}\",\"type\":\"${FIELD_TYPE}\"}]"
+  PATCH_BODY="{\"fields\":${NEW_FIELDS}}"
 
   STATUS=$(wget -qSO- \
     --header="Content-Type: application/json" \
     --header="Authorization: Bearer $TOKEN" \
     --body-data="$PATCH_BODY" \
     --method=PATCH \
-    "$PB_URL/api/collections/$COLLECTION_ID" 2>&1 | grep "HTTP/" | tail -1 | awk '{print $2}')
+    "$PB_URL/api/collections/$COLLECTION_ID" 2>&1 | grep "^  HTTP/" | awk '{print $2}')
   case "$STATUS" in
-    200|201) log "patched collection $NAME — added fields: $ADDED" ;;
+    200|201) log "patched $COLL — added field: $FIELD_NAME ($FIELD_TYPE)" ;;
-    *)       log "WARNING: patch returned $STATUS for collection $NAME" ;;
+    *)       log "WARNING: patch returned $STATUS when adding $FIELD_NAME to $COLL" ;;
   esac
 }
 
@@ -191,8 +184,6 @@ create_collection "app_users" '{
 # ─── 5. Schema migrations (idempotent field additions) ───────────────────────
 # Ensures fields added after initial deploy are present in existing instances.
 
-ensure_fields "progress" '[
+ensure_field "progress" "user_id" "text"
-  {"name": "user_id", "type": "text"}
-]'
 
 log "all collections ready"