From fcf58fa7cf230347ca5be72df54aaa5891444f10 Mon Sep 17 00:00:00 2001
From: Tomas Kracmar <tomas.kracmar@cqre.net>
Date: Mon, 8 Sep 2025 13:16:48 +0200
Subject: [PATCH] Fixing KB name resolution

---
 scripts/podx-tools.sh | 92 ++++++++++++++++---------------------------
 1 file changed, 35 insertions(+), 57 deletions(-)

diff --git a/scripts/podx-tools.sh b/scripts/podx-tools.sh
index ad867b9..6d6e26e 100755
--- a/scripts/podx-tools.sh
+++ b/scripts/podx-tools.sh
@@ -82,6 +82,8 @@ fi
 : "${OPENWEBUI_URL:=http://localhost:3003}"
 : "${OPENWEBUI_URL_HOST:=}"
 : "${OPENWEBUI_API_KEY:=}"
+# Optional: explicit KB id to use for all KB operations (bypasses name resolution)
+: "${OPENWEBUI_KB_ID:=}"
 
 # Resolve a working OpenWebUI base URL (fallback from host.docker.internal -> localhost)
 _owui_url() {
@@ -127,9 +129,14 @@ _kb_create() {
 
 _kb_id_by_name() {
   local kb_name="$1"
+  # If an explicit KB id is provided via env, use it directly
+  if [ -n "${OPENWEBUI_KB_ID:-}" ]; then
+    echo "$OPENWEBUI_KB_ID"
+    return 0
+  fi
   local json; json="$(_owui_get_kb_list)"
-  # Robust KB name resolver: Unicode/space-insensitive, prefers most recently updated
-  printf '%s' "$json" | python3 - "$kb_name" <<'PY' || true
+  local __id=""
+  __id=$(printf '%s' "$json" | python3 - "$kb_name" <<'PY' || true
 import sys, json, unicodedata as ud
 
 def norm(s: str) -> str:
@@ -139,32 +146,49 @@ def norm(s: str) -> str:
 want = norm(sys.argv[1] if len(sys.argv) > 1 else '')
 raw = sys.stdin.read().strip()
 if not raw:
-    sys.exit(0)
+    print('', end=''); raise SystemExit(0)
 try:
     d = json.loads(raw)
 except Exception:
-    sys.exit(0)
+    print('', end=''); raise SystemExit(0)
 
 items = d.get('data') if isinstance(d, dict) and isinstance(d.get('data'), list) else (d if isinstance(d, list) else [])
 
-# If only one KB exists and a name was provided, return it
-if want and isinstance(items, list) and len(items) == 1:
-    print(items[0].get('id',''), end=''); sys.exit(0)
+# If only one KB exists, return it
+if isinstance(items, list) and len(items) == 1:
+    print(items[0].get('id',''), end=''); raise SystemExit(0)
 
-# Prefer exact name match (normalized)
+# Prefer exact normalized name match
 exact = [kb for kb in items if norm(kb.get('name')) == want]
 if exact:
     exact.sort(key=lambda kb: int(kb.get('updated_at') or kb.get('created_at') or 0), reverse=True)
-    print(exact[0].get('id',''), end=''); sys.exit(0)
+    print(exact[0].get('id',''), end=''); raise SystemExit(0)
 
 # Fallback: substring match
 subs = [kb for kb in items if want and want in norm(kb.get('name'))]
 if subs:
     subs.sort(key=lambda kb: int(kb.get('updated_at') or kb.get('created_at') or 0), reverse=True)
-    print(subs[0].get('id',''), end=''); sys.exit(0)
+    print(subs[0].get('id',''), end=''); raise SystemExit(0)
 
 print('', end='')
 PY
+  )
+
+  # If python resolution failed, try jq (if available)
+  if [ -z "${__id:-}" ]; then
+    if command -v jq >/dev/null 2>&1; then
+      __id=$(printf '%s' "$json" | jq -r '
+        if type=="array" and length==1 then .[0].id // empty
+        elif type=="array" then (map(select((.name // "")|ascii_downcase==("'"$kb_name"'"|ascii_downcase))) | .[0].id) // empty
+        elif has("data") and (.data|type=="array") and (.data|length==1) then .data[0].id // empty
+        else empty end')
+    fi
+  fi
+  # Last resort: naive sed/grep — take the first id field
+  if [ -z "${__id:-}" ]; then
+    __id=$(printf '%s' "$json" | sed -n 's/.*"id"[[:space:]]*:[[:space:]]*"\([^"]\+\)".*/\1/p' | head -n1)
+  fi
+  printf '%s' "${__id:-}"
 }
 
 _help() {
@@ -348,53 +372,7 @@ PY
     cat "$tmp_body" | ppjson
     echo "[owui] http_code=$http_code bytes=$bytes"
     json="$(cat "$tmp_body")"; rm -f "$tmp_body"
-    id="$(python3 - "$name" <<'PY'
-import sys, json, unicodedata as ud
-
-def norm(s: str) -> str:
-    s = ud.normalize('NFKC', s or '')
-    return s.strip().casefold()
-
-want = norm(sys.argv[1] if len(sys.argv) > 1 else '')
-raw = sys.stdin.read()
-try:
-    d = json.loads(raw)
-except Exception:
-    print('', end=''); sys.exit(0)
-items = d.get('data') if isinstance(d, dict) and isinstance(d.get('data'), list) else (d if isinstance(d, list) else [])
-
-# Single-KB fallback
-if want and isinstance(items, list) and len(items) == 1:
-    print(items[0].get('id',''), end=''); sys.exit(0)
-
-exact = [kb for kb in items if norm(kb.get('name')) == want]
-subs  = [kb for kb in items if want and want in norm(kb.get('name'))]
-
-cands = exact or subs
-if cands:
-    cands.sort(key=lambda kb: int(kb.get('updated_at') or kb.get('created_at') or 0), reverse=True)
-    print(cands[0].get('id',''), end='')
-else:
-    print('', end='')
-PY
-    <<<"$json")" || id=""
-    # Fallback: if the list contains exactly one KB, use its id
-    if [ -z "$id" ]; then
-      id="$(printf '%s' "$json" | python3 - <<'PY'
-import sys, json
-raw=sys.stdin.read().strip()
-try:
-    d=json.loads(raw)
-except Exception:
-    print('', end=''); raise SystemExit(0)
-items = d.get('data') if isinstance(d, dict) and isinstance(d.get('data'), list) else (d if isinstance(d, list) else [])
-if isinstance(items, list) and len(items)==1:
-    print(items[0].get('id',''), end='')
-else:
-    print('', end='')
-PY
-)" || id=""
-    fi
+    id="$(_kb_id_by_name "$name")"
     if [ -n "$id" ]; then
       echo "[owui] resolved id for \"$name\": $id"
     else