feat(retrieval): add dynamic temporal retriever

.gitignore

@@ -11,6 +11,7 @@ build/**/logs
 configurations/*
 !configurations/configuration.default.json
 !configurations/configuration.local.json
+!configurations/temporal_domains/
 !configurations/configuration.test*.json
 data/
 tmp/

configurations/temporal_domains/bundestag.json

@@ -0,0 +1,115 @@
+{
+  "name": "bundestag",
+  "metadata_schema": {
+    "temporal_field": "legislature_period",
+    "current_period": 21,
+    "historical_period": 20
+  },
+  "temporal_keywords": {
+    "current": {
+      "en": [
+        "current",
+        "recent",
+        "latest",
+        "today",
+        "now",
+        "this year",
+        "present",
+        "nowadays"
+      ],
+      "de": [
+        "aktuell",
+        "jetzt",
+        "neueste",
+        "derzeitig",
+        "gegenwärtig",
+        "dieses jahr",
+        "momentan",
+        "derzeit",
+        "heute"
+      ]
+    },
+    "historical": {
+      "en": [
+        "previous",
+        "past",
+        "former",
+        "old",
+        "historical",
+        "before",
+        "earlier",
+        "last year"
+      ],
+      "de": [
+        "vorherig",
+        "vergangen",
+        "früher",
+        "alt",
+        "historisch",
+        "letztes jahr"
+      ]
+    }
+  },
+  "period_identifiers": {
+    "20": {
+      "names": [
+        "20. Wahlperiode",
+        "Wahlperiode 20",
+        "WP20",
+        "20th legislative period",
+        "20th legislature",
+        "period 20"
+      ],
+      "years": [2021, 2022, 2023, 2024],
+      "temporal_type": "historical"
+    },
+    "21": {
+      "names": [
+        "21. Wahlperiode",
+        "Wahlperiode 21",
+        "WP21",
+        "21st legislative period",
+        "21st legislature",
+        "period 21"
+      ],
+      "years": [2025],
+      "temporal_type": "current"
+    }
+  },
+  "query_expansion": {
+    "temporal_current": {
+      "de": "21. Wahlperiode 2025 aktuelle Bundesregierung neueste",
+      "en": "21st legislature 2025 current government latest 21. Wahlperiode"
+    },
+    "temporal_historical": {
+      "de": "20. Wahlperiode 2021 2022 2023 2024 frühere Bundesregierung vergangene",
+      "en": "20th legislature 2021 2022 2023 2024 former government previous 20. Wahlperiode"
+    },
+    "entity_terms": {
+      "de": "Fraktionen Bundestag Bundestagsfraktionen parlamentarische Gruppen Wahlperiode",
+      "en": "fractions Bundestag parliamentary groups legislature period"
+    }
+  },
+  "language_detection": {
+    "de": [
+      "wer",
+      "was",
+      "welche",
+      "alle",
+      "der",
+      "die",
+      "das",
+      "ist",
+      "sind"
+    ],
+    "en": [
+      "who",
+      "what",
+      "which",
+      "all",
+      "the",
+      "is",
+      "are"
+    ]
+  }
+}

configurations/temporal_domains/generic.json

@@ -0,0 +1,21 @@
+{
+  "name": "generic",
+  "metadata_schema": {
+    "temporal_field": "period",
+    "current_period": null,
+    "historical_period": null
+  },
+  "temporal_keywords": {
+    "current": {
+      "en": []
+    },
+    "historical": {
+      "en": []
+    }
+  },
+  "period_identifiers": {},
+  "query_expansion": {},
+  "language_detection": {
+    "en": ["the", "is", "are", "a", "an"]
+  }
+}

src/augmentation/bootstrap/configuration/components/retriever_configuration.py

@@ -12,6 +12,7 @@ class RetrieverName(str, Enum):
 
     BASIC = "basic"
     AUTO = "auto"
+    DYNAMIC_TEMPORAL = "dynamic_temporal"
 
 
 class RetrieverConfiguration(BaseConfiguration):

src/augmentation/bootstrap/configuration/configuration.py

@@ -1,4 +1,6 @@
-from typing import Any
+import json
+from pathlib import Path
+from typing import Any, Optional, Union
 
 from pydantic import Field, ValidationInfo, field_validator
 
@@ -11,6 +13,9 @@
 from augmentation.bootstrap.configuration.langfuse_configuration import (
     LangfuseConfiguration,
 )
+from augmentation.bootstrap.configuration.temporal_domain_config import (
+    TemporalDomainConfiguration,
+)
 from core.base_configuration import BaseConfiguration
 from embedding.bootstrap.configuration.configuration import (
     EmbeddingConfiguration,
@@ -22,7 +27,8 @@ class _AugmentationConfiguration(BaseConfiguration):
     Internal configuration class for augmentation process settings.
 
     This class defines the structure for augmentation configuration including
-    Langfuse monitoring, Chainlit UI, and Chat Engine components.
+    Langfuse monitoring, Chainlit UI, Chat Engine components, and optional
+    temporal domain configuration.
     """
 
     langfuse: LangfuseConfiguration = Field(
@@ -34,6 +40,76 @@ class _AugmentationConfiguration(BaseConfiguration):
     chat_engine: Any = Field(
         ..., description="Configuration of the Chat Engine."
     )
+    temporal_domain: Optional[Union[str, TemporalDomainConfiguration]] = Field(
+        default=None,
+        description="Temporal domain configuration. Can be a file path (str) or inline config. "
+        "If not provided, system runs in generic mode without temporal filtering.",
+    )
+
+    @field_validator("temporal_domain")
+    @classmethod
+    def _validate_temporal_domain(
+        cls, value: Optional[Union[str, TemporalDomainConfiguration]]
+    ) -> Optional[TemporalDomainConfiguration]:
+        """
+        Validates and loads temporal domain configuration.
+
+        If value is a string, treats it as a file path and loads the configuration
+        from that file. If value is already a TemporalDomainConfiguration, returns it.
+        If value is None, returns None (generic mode).
+
+        Args:
+            value: Temporal domain config (file path, config object, or None)
+
+        Returns:
+            Loaded TemporalDomainConfiguration or None
+
+        Raises:
+            ValueError: If file path is invalid or file content is malformed
+        """
+        if value is None:
+            return None
+
+        if isinstance(value, TemporalDomainConfiguration):
+            return value
+
+        if isinstance(value, str):
+            # Treat as file path
+            file_path = Path(value)
+
+            # If relative path, resolve relative to configurations directory
+            if not file_path.is_absolute():
+                config_dir = (
+                    Path(__file__).parent.parent.parent.parent.parent
+                    / "configurations"
+                )
+                file_path = config_dir / value
+
+            if not file_path.exists():
+                raise ValueError(
+                    f"Temporal domain configuration file not found: {file_path}"
+                )
+
+            try:
+                with open(file_path, "r") as f:
+                    config_data = json.load(f)
+                return TemporalDomainConfiguration(**config_data)
+            except json.JSONDecodeError as e:
+                raise ValueError(
+                    f"Invalid JSON in temporal domain config file {file_path}: {e}"
+                )
+            except Exception as e:
+                raise ValueError(
+                    f"Failed to load temporal domain config from {file_path}: {e}"
+                )
+
+        if isinstance(value, dict):
+            # Treat as inline configuration
+            return TemporalDomainConfiguration(**value)
+
+        raise ValueError(
+            f"temporal_domain must be a file path (str), dict, or TemporalDomainConfiguration, got {type(value)}"
+        )
 
     @field_validator("chat_engine")
     @classmethod