p43F

2026-05-05 17:17:35 +02:00
parent 542b527261
commit ed11e6e9c8
4 changed files with 147 additions and 24 deletions
--- a/config/retriex/search_repair.yaml
+++ b/config/retriex/search_repair.yaml
@@ -19,31 +19,19 @@ parameters:

    requested_accessory_code_fallback_query_templates:
      - '{term} {code}'
-    requested_accessory_code_fallback_terms:
-      - indikatortyp
-      - indikator
-      - indicator
-      - reagenz
-      - reagent
-    requested_accessory_code_context_prefix_terms:
-      - indikatortyp
-      - indikator
-      - indicator
-      - reagenz
-      - reagent
+    # Requested-accessory code terms are resolved from
+    # config/retriex/vocabulary.yaml view search_repair.requested_accessory_code_terms.
+    # Local requested_accessory_code_fallback_terms and
+    # requested_accessory_code_context_prefix_terms lists may still be added here as
+    # explicit project overrides.
    requested_accessory_code_proximity_window: 1600

    specific_model_candidate_patterns:
      - '/\b([A-Za-zÄÖÜäöüß][A-Za-zÄÖÜäöüß®\-]*(?:\s+[A-Za-zÄÖÜäöüß0-9][A-Za-zÄÖÜäöüß0-9®\-]*){0,3}\s+\d{2,5}(?:\s+[A-ZÄÖÜ]{1,8})?)\b/u'
-    model_candidate_exclude_terms:
-      - indikatortyp
-      - indikator
-      - indicator
-      - reagenz
-      - reagent
-      - verfügbarkeit
-      - verfuegbarkeit
-      - shop
+    # Model-candidate exclude terms are resolved from
+    # config/retriex/vocabulary.yaml view search_repair.model_candidate_exclude_terms.
+    # A local model_candidate_exclude_terms list may still be added here as an
+    # explicit project override.

    limits:
      top_product_log_limit: 3
--- a/config/retriex/vocabulary.yaml
+++ b/config/retriex/vocabulary.yaml
@@ -49,6 +49,16 @@ parameters:
      - service set
      - serviceset
      - service-set
+      requested_accessory_code_terms:
+      - indikatortyp
+      - indikator
+      - indicator
+      - reagenz
+      - reagent
+      shop_meta_terms:
+      - verfügbarkeit
+      - verfuegbarkeit
+      - shop
      direct_product_attribute_stop_terms:
      - zeige
      - zeig
@@ -460,6 +470,13 @@ parameters:
        direct_product_attribute_stop_terms:
          include:
          - direct_product_attribute_stop_terms
+        requested_accessory_code_terms:
+          include:
+          - requested_accessory_code_terms
+        model_candidate_exclude_terms:
+          include:
+          - requested_accessory_code_terms
+          - shop_meta_terms
        accessory_candidate_terms:
          add:
          - indikator
--- a/patch_history/RETRIEX_PATCH_43F_SEARCH_REPAIR_ACCESSORY_CODE_VOCABULARY_VIEW_README.md
+++ b/patch_history/RETRIEX_PATCH_43F_SEARCH_REPAIR_ACCESSORY_CODE_VOCABULARY_VIEW_README.md
@@ -0,0 +1,109 @@
+# RetrieX Patch p43F - Search Repair Accessory Code Vocabulary View
+
+## Ziel
+
+p43F setzt auf dem gruen bestaetigten p43E-Stand auf und reduziert weitere doppelte SearchRepair-Begriffslisten ohne fachliche Runtime-Aenderung.
+
+Der Patch zentralisiert die mehrfach gepflegten Zubehoercode-Begriffe fuer SearchRepair in `config/retriex/vocabulary.yaml` und laesst lokale Overrides in `config/retriex/search_repair.yaml` weiterhin zu.
+
+## Geaenderte Dateien
+
+- `config/retriex/vocabulary.yaml`
+- `config/retriex/search_repair.yaml`
+- `src/Config/SearchRepairConfig.php`
+- `patch_history/RETRIEX_PATCH_43F_SEARCH_REPAIR_ACCESSORY_CODE_VOCABULARY_VIEW_README.md`
+
+## Inhalt
+
+### vocabulary.yaml
+
+Neu zentral gepflegte Klassen:
+
+- `requested_accessory_code_terms`
+- `shop_meta_terms`
+
+Neue SearchRepair-Views:
+
+- `search_repair.requested_accessory_code_terms`
+- `search_repair.model_candidate_exclude_terms`
+
+Die View `search_repair.requested_accessory_code_terms` liefert die bisherigen Begriffe fuer:
+
+- `requested_accessory_code_fallback_terms`
+- `requested_accessory_code_context_prefix_terms`
+
+Die View `search_repair.model_candidate_exclude_terms` liefert die bisherigen Modell-Exclude-Begriffe aus Zubehoercode-Begriffen plus Shop-/Verfuegbarkeits-Metabegriffen.
+
+### search_repair.yaml
+
+Lokale Listen wurden entfernt fuer:
+
+- `requested_accessory_code_fallback_terms`
+- `requested_accessory_code_context_prefix_terms`
+- `model_candidate_exclude_terms`
+
+Die YAML-Datei dokumentiert jetzt, dass diese Werte aus `vocabulary.yaml` kommen und lokale Listen weiterhin als explizite Projekt-Overrides wieder hinzugefuegt werden koennen.
+
+### SearchRepairConfig.php
+
+Die drei Accessoren nutzen nun den vorhandenen Mechanismus `configOrVocabularyStringList()`:
+
+- `getRequestedAccessoryCodeFallbackTerms()`
+- `getRequestedAccessoryCodeContextPrefixTerms()`
+- `getModelCandidateExcludeTerms()`
+
+Damit bleibt das bisherige Override-Verhalten erhalten: lokale YAML-Listen haben Vorrang, sonst wird die zentrale Vocabulary-View verwendet.
+
+## Explizit nicht geaendert
+
+- keine neue fachliche Runtime-Logik
+- keine Scoring-Aenderung
+- keine Retrieval-Aenderung
+- keine Prompt-Regel-Aenderung
+- keine Admin-UI
+- keine neuen harten Listen im PHP-Core
+- keine Aenderung an Pattern-Templates oder Regex-Semantik
+
+## Effektive Listen vor/nach p43F
+
+Verglichen wurde p43E-Basis gegen p43F-Arbeitsstand:
+
+- `requested_accessory_code_fallback_terms`: 5 / 5, identisch
+- `requested_accessory_code_context_prefix_terms`: 5 / 5, identisch
+- `model_candidate_exclude_terms`: 8 / 8, identisch
+- `direct_product_attribute_lookup.product_type_terms`: 12 / 12, identisch
+- `direct_product_attribute_lookup.stop_terms`: 40 / 40, identisch
+- `generic_candidate_tokens`: 13 / 13, identisch
+- `accessory_candidate_terms`: 6 / 6, identisch
+- `accessory_or_bundle_terms`: 14 / 14, identisch
+- `specificity_boost_terms`: 7 / 7, identisch
+
+## Lokal ausgefuehrte Checks
+
+```bash
+php -l src/Config/SearchRepairConfig.php
+php -l src/Config/AgentRunnerConfig.php
+python3 YAML parse check fuer geaenderte YAML-Dateien
+python3 effektive SearchRepair-Listen p43E vs p43F vergleichen
+```
+
+Ergebnis: gruen.
+
+## Nicht lokal ausfuehrbar
+
+Die Symfony-Console-Checks konnten in der ChatGPT-Arbeitsumgebung nicht ausgefuehrt werden, weil der entpackte ZIP-Stand kein `vendor/` enthaelt und in dieser Umgebung kein `composer` verfuegbar ist.
+
+Blockierender Fehler:
+
+```text
+Dependencies are missing. Try running "composer install".
+```
+
+Bitte nach dem Einspielen im Projekt ausfuehren:
+
+```bash
+bin/console mto:agent:config:validate
+bin/console mto:agent:regression:test
+bin/console mto:agent:config:audit-source --details
+bin/console mto:agent:config:audit-patterns --details
+```
--- a/src/Config/SearchRepairConfig.php
+++ b/src/Config/SearchRepairConfig.php
@@ -93,13 +93,19 @@ final class SearchRepairConfig
    /** @return string[] */
    public function getRequestedAccessoryCodeFallbackTerms(): array
    {
-        return $this->requiredStringList('requested_accessory_code_fallback_terms');
+        return $this->configOrVocabularyStringList(
+            'requested_accessory_code_fallback_terms',
+            'search_repair.requested_accessory_code_terms'
+        );
    }

    /** @return string[] */
    public function getRequestedAccessoryCodeContextPrefixTerms(): array
    {
-        return $this->requiredStringList('requested_accessory_code_context_prefix_terms');
+        return $this->configOrVocabularyStringList(
+            'requested_accessory_code_context_prefix_terms',
+            'search_repair.requested_accessory_code_terms'
+        );
    }

    public function getRequestedAccessoryCodeProximityWindow(): int
@@ -116,7 +122,10 @@ final class SearchRepairConfig
    /** @return string[] */
    public function getModelCandidateExcludeTerms(): array
    {
-        return $this->requiredStringList('model_candidate_exclude_terms');
+        return $this->configOrVocabularyStringList(
+            'model_candidate_exclude_terms',
+            'search_repair.model_candidate_exclude_terms'
+        );
    }

    public function getTopProductLogLimit(): int