Refactor and enhance agent tasks and tools

- Updated task definitions in notebooks to use more descriptive field names (e.g., 'query' to 'question').
- Changed the default task in domain tuning notebook from "sentiment" to "plan_execute".
- Improved agent behavior optimization by refining prompts and adding explanations for model choices.
- Enhanced search functionality in tools to provide better ticket search results and counts.
- Updated calculations for plan quality and self-correct accuracy to align with new output structures.
- Added MIPROv2 optimization step to improve agent responses based on vague prompts.
- Adjusted dataset for search agent to include more complex queries and answers.
- Updated kernel specifications across notebooks to use Python 3.13.12.

Author: abossard

notebooks/03_domain_tuning.ipynb

@@ -83,7 +83,7 @@
    "outputs": [],
    "source": [
     "# Ändere den Task hier (z.B. \"sentiment\", \"math_word\", \"fact_verification\")\n",
-    "TASK = \"sentiment\"\n",
+    "TASK = \"plan_execute\"\n",
     "\n",
     "task = get_task(TASK)\n",
     "print(f\"📋 {task.name}\")\n",
@@ -102,73 +102,6 @@
    "cell_type": "markdown",
    "id": "6",
    "metadata": {},
-   "source": [
-    "## 3. Jetzt mit deinen echten Daten\n",
-    "\n",
-    "Generische Aufgaben wie Sentiment oder Mathe sind nett — aber der **echte Wert** zeigt sich bei **deinen eigenen Daten**.\n",
-    "\n",
-    "Wir laden echte Support-Tickets aus dem Projekt. Jedes Ticket hat:\n",
-    "- **Zusammenfassung** — was ist das Problem?\n",
-    "- **Kategorie** — Network, Software, Hardware, ...\n",
-    "- **Priorität** — Critical, High, Medium, Low\n",
-    "- **Team** — welches Team kümmert sich?\n",
-    "\n",
-    "Das Modell soll neue Tickets automatisch klassifizieren.\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "7",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "task = get_task(\"ticket_routing\")\n",
-    "examples = task.load_examples()\n",
-    "\n",
-    "print(f\"📊 {len(examples)} echte Tickets geladen\\n\")\n",
-    "for ex in examples[:5]:\n",
-    "    print(f\"  📋 {str(ex.summary)[:80]}...\")\n",
-    "    print(f\"     → Kategorie: {ex.category} | Priorität: {ex.priority} | Team: {ex.assigned_group}\")\n",
-    "    print()\n"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "8",
-   "metadata": {},
-   "source": [
-    "## 4. Generischer Prompt → wie gut ist er?\n",
-    "\n",
-    "Erst testen wir ohne Tuning — nur ein generischer Prompt: \"Klassifiziere dieses Ticket.\"\n",
-    "\n",
-    "Die Metrik bewertet drei Felder gleichzeitig:\n",
-    "- Priorität korrekt? → **40%** Gewicht (am wichtigsten!)\n",
-    "- Kategorie korrekt? → **35%** Gewicht\n",
-    "- Richtiges Team? → **25%** Gewicht\n"
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": null,
-   "id": "9",
-   "metadata": {},
-   "outputs": [],
-   "source": [
-    "print(f\"⏳ Teste generischen Prompt auf {MODEL}...\\n\")\n",
-    "\n",
-    "baseline = run_baseline(\"ticket_routing\", max_eval=10)\n",
-    "\n",
-    "display_score(\"Generischer Prompt (Baseline)\", baseline.score)\n",
-    "display_results_table(baseline.individual_scores[:5])\n",
-    "\n",
-    "print(f\"\\n👆 {baseline.score:.0%} — das Modell kennt eure Teams und Kategorien nicht!\")\n"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "id": "10",
-   "metadata": {},
    "source": [
     "## 5. Domain-Tuning → der Unterschied\n",
     "\n",
@@ -178,7 +111,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "11",
+   "id": "7",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -192,7 +125,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "12",
+   "id": "8",
    "metadata": {},
    "source": [
     "## 6. Was du gerade gesehen hast\n",
@@ -214,7 +147,7 @@
   {
    "cell_type": "code",
    "execution_count": null,
-   "id": "13",
+   "id": "9",
    "metadata": {},
    "outputs": [],
    "source": [
@@ -226,7 +159,7 @@
   },
   {
    "cell_type": "markdown",
-   "id": "14",
+   "id": "10",
    "metadata": {},
    "source": [
     "## ⏭️ Weiter geht's!\n",
@@ -239,13 +172,21 @@
  ],
  "metadata": {
   "kernelspec": {
-   "display_name": "Python 3",
+   "display_name": ".venv (3.13.12)",
    "language": "python",
    "name": "python3"
   },
   "language_info": {
+   "codemirror_mode": {
+    "name": "ipython",
+    "version": 3
+   },
+   "file_extension": ".py",
+   "mimetype": "text/x-python",
    "name": "python",
-   "version": "3.11.0"
+   "nbconvert_exporter": "python",
+   "pygments_lexer": "ipython3",
+   "version": "3.13.12"
   }
  },
  "nbformat": 4,