From 9be69ace5c125c3c41244e1adca2cc375a62d50c Mon Sep 17 00:00:00 2001
From: Lars <Lars@stommer.de>
Date: Fri, 22 May 2026 09:59:56 +0200
Subject: [PATCH] Enhance exercise_ai module with skill input sanitization and
 version update

- Introduced a new constant `_MAX_SANITIZE_SKILL_INPUT_ROWS` to limit the number of skill entries processed, improving performance and preventing issues with excessively long skill arrays.
- Updated the `_extract_json_array` and `_sanitize_skill_entries` functions to enforce this limit, ensuring that only a maximum of 250 skill entries are handled and that processing stops after 5 valid entries.
- Incremented the application version to 0.8.155 and updated the changelog to reflect these changes, including a note on the improvements made to the AI endpoint for skill arrays.
---
 backend/exercise_ai.py | 18 +++++++++++++++---
 backend/version.py     | 11 +++++++++--
 2 files changed, 24 insertions(+), 5 deletions(-)

diff --git a/backend/exercise_ai.py b/backend/exercise_ai.py
index 255ad4c..eb48e2f 100644
--- a/backend/exercise_ai.py
+++ b/backend/exercise_ai.py
@@ -34,6 +34,7 @@ _TOKEN_FIND = re.compile(r"[a-zäöüß0-9]+", re.IGNORECASE)
 _MAX_PLAIN_FIELD = 28_000
 _MAX_SKILLS_CATALOG_LINES = 240
 _MAX_SUMMARY_CHARS = 220
+_MAX_SANITIZE_SKILL_INPUT_ROWS = 250
 
 _FALLBACK_RETRIEVAL_CONFIG: Dict[str, Any] = {
     "version": 1,
@@ -514,23 +515,34 @@ def _extract_json_array(text: str) -> Any:
         end = s.rfind("]")
         if end > 0:
             s = s[: end + 1]
-            return json.loads(s)
+            parsed = json.loads(s)
+            if isinstance(parsed, list) and len(parsed) > _MAX_SANITIZE_SKILL_INPUT_ROWS:
+                parsed = parsed[:_MAX_SANITIZE_SKILL_INPUT_ROWS]
+            return parsed
     if s.startswith("{"):
         obj = json.loads(s)
         if isinstance(obj, dict):
             for k in ("skills", "items", "data"):
                 v = obj.get(k)
                 if isinstance(v, list):
+                    if len(v) > _MAX_SANITIZE_SKILL_INPUT_ROWS:
+                        return v[:_MAX_SANITIZE_SKILL_INPUT_ROWS]
                     return v
         raise ValueError("JSON-Objekt ohne Skills-Liste")
-    return json.loads(s)
+    parsed_end = json.loads(s)
+    if isinstance(parsed_end, list) and len(parsed_end) > _MAX_SANITIZE_SKILL_INPUT_ROWS:
+        return parsed_end[:_MAX_SANITIZE_SKILL_INPUT_ROWS]
+    return parsed_end
 
 
 def _sanitize_skill_entries(cur, rows: Any) -> List[Dict[str, Any]]:
     if not isinstance(rows, list):
         return []
     out: List[Dict[str, Any]] = []
-    for raw in rows:
+    cap = rows[:_MAX_SANITIZE_SKILL_INPUT_ROWS]
+    for raw in cap:
+        if len(out) >= 5:
+            break
         if not isinstance(raw, dict):
             continue
         sid = raw.get("skill_id")
diff --git a/backend/version.py b/backend/version.py
index eb40771..4407d29 100644
--- a/backend/version.py
+++ b/backend/version.py
@@ -1,6 +1,6 @@
 # Shinkan Jinkendo Version Information
 
-APP_VERSION = "0.8.154"
+APP_VERSION = "0.8.155"
 BUILD_DATE = "2026-05-29"
 DB_SCHEMA_VERSION = "20260529068"
 
@@ -23,7 +23,7 @@ MODULE_VERSIONS = {
     "skills": "0.1.1",  # DB 065 karate_relevance + relevance_level; CRUD unterstützt Felder
     "skill_profiles": "1.0.0",  # Phase 3: gewichtetes Fähigkeiten-Profil + skill-discovery/suggestions
     "methods": "0.1.0",
-    "exercises": "2.30.0",  # Migration 068 ai_skill_retrieval_profiles; suggest focus_areas_context; exercise_ai Kontext-Katalog + Gewichtungen
+    "exercises": "2.30.1",  # exercise_ai: Skills-JSON max. 250 Eintraege, Sanitize bricht nach 5 gueltigen ab (Performance)
     "training_units": "0.4.0",  # POST .../publish-to-framework: Ablauf aus geplanter Einheit → Rahmen-Slot-Blueprint
     "training_programs": "0.1.0",
     "planning": "0.15.0",  # Vorlagen: Strukturvorschau, Bearbeiten inkl. Split-Sessions + Beschreibung
@@ -38,6 +38,13 @@ MODULE_VERSIONS = {
 }
 
 CHANGELOG = [
+    {
+        "version": "0.8.155",
+        "date": "2026-05-29",
+        "changes": [
+            "exercise_ai: Fix haengender KI-Endpunkt bei sehr langen Skill-Arrays vom Modell (Cap + frueher Abbruch nach 5 gueltigen Zeilen)",
+        ],
+    },
     {
         "version": "0.8.154",
         "date": "2026-05-29",