harden history find tokens and shops earch

config/retriex/agent.yaml

@@ -13,6 +13,7 @@ parameters:
       check_internet_sources: 'Ich prüfe auf Internetquellen...'
       retrieve_knowledge: 'Ich hole relevante Daten aus meinem RAG-Wissen...'
       optimize_search: 'Ich optimiere die Recherche...'
+      no_concrete_shop_query: 'Ich habe keine konkrete Shop-Suchanfrage erkannt. Bitte nenne das Produkt, Zubehör oder die Artikelnummer.'
       fetch_search_data_template: 'Ich rufe Recherchedaten ab (type: %s)'
       analyze_all_information: 'Ich analysiere alle Informationen...'
       thinking_while_streaming: 'Denke nach...'
@@ -53,11 +54,14 @@ parameters:
         - '- Maximum 6 search terms, preferably fewer.'
         - '- Remove filler words, polite phrases, and irrelevant words.'
         - '- Preserve product names, brands, model numbers, and compound terms exactly if they are relevant.'
+        - '- Preserve the language of the CURRENT USER INPUT for generic product/search terms; do not translate German search terms into English.'
+        - '- For German user input, output German shop terms, for example "freies Chlor Messung" instead of "free chlorine measurement".'
+        - '- Preserve domain terms from the current user input or resolved context in their original language.'
         - '- Numbers that belong to a product name or model must be preserved (e.g. Indikator 300, Testomat 808, Testomat 2000).'
         - '- Separate terms using spaces only.'
         - '- If a relevant product name is present, it must be placed at the beginning of the final search query.'
         - '- Try to always identify all products mentioned in the user input text, even in long prompts.'
-        - '- Look for terms such as Testomat, Horiba, Tritromat, or words like indicator.'
+        - '- Look for terms such as Testomat, Horiba, Tritromat, or words like indicator/Indikator.'
         - '- If the current user input is vague or referential, use the recent conversation context only as support.'
         - '- Do not output words that only describe conversation flow, such as "same", "again", "also", or "like above".'
       conversation_context_rules:
@@ -67,3 +71,121 @@ parameters:
         - '- Do not revive older products unless the current user input clearly refers to them.'
         - '- If the current input starts a new topic, ignore older product context.'
         - '- Prefer the most recent product reference over older ones.'
+      context_anchor_enrichment:
+        enabled: true
+        max_query_terms: 2
+        template: '{anchor} {query}'
+        trigger_terms:
+          - indikator
+          - indikatortyp
+          - indicator
+          - reagenz
+          - reagenzsatz
+          - reagent
+          - zubehör
+          - zubehor
+          - accessory
+        anchor_patterns:
+          - '/\b(?:indikator(?:typ)?|indicator(?:\s+type)?|reagenz(?:satz|typ)?|reagent(?:\s+set|\s+type)?|typ|type)\s+[A-Za-zÄÖÜäöüß]{0,8}\s*\d{1,5}(?:\s*[A-ZÄÖÜ]{1,4})?(?:\s*%)?\b/iu'
+      meta_query_guard:
+        enabled: true
+        context_fallback_enabled: true
+        context_fallback_question_limit: 12
+        context_fallback_history_budget_chars: 20000
+        context_fallback_use_full_history: true
+        context_fallback_max_terms: 6
+        context_fallback_filter_terms:
+          - mit
+          - welche
+          - welcher
+          - welches
+          - welchem
+          - welchen
+          - was
+          - wie
+          - wo
+          - kann
+          - koennen
+          - können
+          - konnte
+          - könnte
+          - ich
+          - wir
+          - man
+          - nutzen
+          - benutzen
+          - verwenden
+          - verwende
+          - nehmen
+          - zur
+          - zum
+          - für
+          - fuer
+          - messen
+          - gemessen
+          - messung
+        meta_only_terms:
+          - shop
+          - shopsuche
+          - shop-suche
+          - suche
+          - suchen
+          - such
+          - finde
+          - find
+          - zeige
+          - zeig
+          - bitte
+          - mal
+          - im
+          - in
+          - nach
+          - den
+          - die
+          - das
+          - der
+          - dem
+      language_preservation:
+        enabled: true
+        language_markers:
+          de:
+            - ' ä '
+            - ' ö '
+            - ' ü '
+            - ' ß '
+            - ' der '
+            - ' die '
+            - ' das '
+            - ' ein '
+            - ' eine '
+            - ' einer '
+            - ' einen '
+            - ' welchem '
+            - ' welchen '
+            - ' welche '
+            - ' welcher '
+            - ' kann '
+            - ' nutzen '
+            - ' zur '
+            - ' für '
+            - ' fuer '
+            - ' messung '
+            - ' indikator '
+            - ' reagenz '
+            - ' chlor '
+        translation_replacements:
+          de:
+            free chlorine: 'freies chlor'
+            free chlor: 'freies chlor'
+            total chlorine: 'gesamtchlor'
+            chlorine measurement: 'chlor messung'
+            water hardness: 'wasserhärte'
+            measurement: 'messung'
+            measuring: 'messung'
+            chlorine: 'chlor'
+            indicator: 'indikator'
+            indicators: 'indikatoren'
+            reagent: 'reagenz'
+            reagents: 'reagenzien'
+            accessory: 'zubehör'
+            accessories: 'zubehör'