nucliadb-models 6.5.0.post4426__py3-none-any.whl → 6.5.0.post4484__py3-none-any.whl

nucliadb_models/graph/requests.py

@@ -116,7 +116,7 @@ class GraphFilterExpression(BaseModel, extra="forbid"):
 
 
 class BaseGraphSearchRequest(BaseModel):
-    top_k: int = Field(default=50, title="Number of results to retrieve")
+    top_k: int = Field(default=50, le=500, title="Number of results to retrieve")
     filter_expression: Optional[GraphFilterExpression] = Field(
         default=None,
         title="Filter resource by an expression",

nucliadb_models/search.py

@@ -2035,6 +2035,39 @@ def validate_facets(facets):
     return facets
 
 
+class TextBlockAugmentationType(str, Enum):
+    NEIGHBOURING_PARAGRAPHS = "neighbouring_paragraphs"
+    CONVERSATION = "conversation"
+    HIERARCHY = "hierarchy"
+    FULL_RESOURCE = "full_resource"
+    FIELD_EXTENSION = "field_extension"
+    METADATA_EXTENSION = "metadata_extension"
+
+
+class AugmentedTextBlock(BaseModel):
+    id: str = Field(
+        description="The id of the augmented text bloc. It can be a paragraph id or a field id."
+    )
+    text: str = Field(
+        description="The text of the augmented text block. It may include additional metadata to enrich the context"
+    )
+    parent: Optional[str] = Field(
+        default=None, description="The parent text block that was augmented for."
+    )
+    augmentation_type: TextBlockAugmentationType = Field(description="Type of augmentation.")
+
+
+class AugmentedContext(BaseModel):
+    paragraphs: dict[str, AugmentedTextBlock] = Field(
+        default={},
+        description="Paragraphs added to the context as a result of using the `rag_strategies` parameter, typically the neighbouring_paragraphs or the conversation strategies",
+    )
+    fields: dict[str, AugmentedTextBlock] = Field(
+        default={},
+        description="Field extracted texts added to the context as a result of using the `rag_strategies` parameter, typically the hierarcy or full_resource strategies.",
+    )
+
+
 class AskTokens(BaseModel):
     input: int = Field(
         title="Input tokens",
@@ -2132,6 +2165,13 @@ class SyncAskResponse(BaseModel):
         title="Citations",
         description="The citations of the answer. List of references to the resources used to generate the answer.",
     )
+    augmented_context: Optional[AugmentedContext] = Field(
+        default=None,
+        description=(
+            "Augmented text blocks that were sent to the LLM as part of the RAG strategies "
+            "applied on the retrieval results in the request."
+        ),
+    )
     prompt_context: Optional[list[str]] = Field(
         default=None,
         title="Prompt context",
@@ -2194,6 +2234,16 @@ class MetadataAskResponseItem(BaseModel):
     timings: AskTimings
 
 
+class AugmentedContextResponseItem(BaseModel):
+    type: Literal["augmented_context"] = "augmented_context"
+    augmented: AugmentedContext = Field(
+        description=(
+            "Augmented text blocks that were sent to the LLM as part of the RAG strategies "
+            "applied on the retrieval results in the request."
+        )
+    )
+
+
 class CitationsAskResponseItem(BaseModel):
     type: Literal["citations"] = "citations"
     citations: dict[str, Any]
@@ -2226,6 +2276,7 @@ AskResponseItemType = Union[
     AnswerAskResponseItem,
     JSONAskResponseItem,
     MetadataAskResponseItem,
+    AugmentedContextResponseItem,
     CitationsAskResponseItem,
     StatusAskResponseItem,
     ErrorAskResponseItem,
@@ -2260,3 +2311,19 @@ def parse_rephrase_prompt(item: AskRequest) -> Optional[str]:
 
 # We need this to avoid issues with pydantic and generic types defined in another module
 FindRequest.model_rebuild()
+
+
+class CatalogFacetsPrefix(BaseModel):
+    prefix: str = Field(pattern="^((/[^/]+)*)$")
+    depth: Optional[int] = Field(default=None, ge=0)
+
+    @model_validator(mode="before")
+    @classmethod
+    def build_from_string(cls, data: Any) -> Any:
+        if isinstance(data, str):
+            data = {"prefix": data}
+        return data
+
+
+class CatalogFacetsRequest(BaseModel):
+    prefixes: list[CatalogFacetsPrefix] = Field(default=[])

{nucliadb_models-6.5.0.post4426.dist-info → nucliadb_models-6.5.0.post4484.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: nucliadb_models
-Version: 6.5.0.post4426
+Version: 6.5.0.post4484
 Author-email: Nuclia <nucliadb@nuclia.com>
 License-Expression: Apache-2.0
 Project-URL: Homepage, https://nuclia.com

{nucliadb_models-6.5.0.post4426.dist-info → nucliadb_models-6.5.0.post4484.dist-info}/RECORD

@@ -16,7 +16,7 @@ nucliadb_models/notifications.py,sha256=3w1HeX9F8nuA7WupHdpXIksX7d0bHO9ofmemYp4R
 nucliadb_models/processing.py,sha256=nhKuHQjqCdb9zJVkYGPTLub23tK9e_lwL5OCDVymZjY,719
 nucliadb_models/py.typed,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 nucliadb_models/resource.py,sha256=FkhwmbjoqQBGyKQIqa40WK2Vq3wtryZeIVWmJ4b-84g,9003
-nucliadb_models/search.py,sha256=govZKhri5fohFcSJwUmAvh7AMQeAcm3psA3KcPcCT0c,86700
+nucliadb_models/search.py,sha256=W2dRUUst4ZqWitbCrvR1OV9rptYHXzODMI0seeTUzbs,89147
 nucliadb_models/security.py,sha256=opxaDLfvk3aU0sjesK0jGrYLx5h4YCwlKKN0moYs_ig,1150
 nucliadb_models/synonyms.py,sha256=afbaVqSQSxGLwi2PusVaLSRpkOtA5AZmWOKd1f4nl2E,690
 nucliadb_models/text.py,sha256=kY2ub7AaGm-4vNaLX3Ju2VvRw-eKZ2LRdM9z7XCNaG0,2898
@@ -27,12 +27,12 @@ nucliadb_models/vectorsets.py,sha256=XAgg9DfdfLYpfLh9OepJ_KPH0_RqRQNpVZJr74UnNh0
 nucliadb_models/writer.py,sha256=diwrarp6DxjSUoRmdEljZb68z_ghNvpOgPUGZeKg328,8220
 nucliadb_models/agents/ingestion.py,sha256=W9cJ0dQT_1vPcjeJ4_Fjb8DylnhQ6qqZrY4v8x1RqUs,3093
 nucliadb_models/graph/__init__.py,sha256=X538kZPZnndmQeEtnzzPv1hYVGUTDe9U1O7UmAqqxXU,645
-nucliadb_models/graph/requests.py,sha256=KwdP9xSDu1Td2O-7cRXBv_YFF6DsqAlD6M0-4LKAD60,6506
+nucliadb_models/graph/requests.py,sha256=LgudW-Wg1UW13_CN57k-ek9M2h5zeTWqb13ktdorRaE,6514
 nucliadb_models/graph/responses.py,sha256=Sdq8OgFAL1YT-1lJyLLrkqcScvj7YTEqAUwQ-kFAk9M,1399
 nucliadb_models/internal/__init__.py,sha256=zG33bUz1rHFPtvqQPWn4rDwBJt3FJodGuQYD45quiQg,583
 nucliadb_models/internal/predict.py,sha256=Pnx6MmLfK65eExe1XnVxqmSlvMwdowewwks9BOEoqMw,2029
 nucliadb_models/internal/shards.py,sha256=__y1OZtWGiNcPQEWfSFOj8yw458WGi7mM4vZe0K-L1Y,1691
-nucliadb_models-6.5.0.post4426.dist-info/METADATA,sha256=btEK0Gfuuo6fh2Ob-WkI2REdxyZBmnoxYYAglvSIvZI,776
-nucliadb_models-6.5.0.post4426.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
-nucliadb_models-6.5.0.post4426.dist-info/top_level.txt,sha256=UrY1I8oeovIRwkXLYplssTrxQdUjhSEFDFbnwaIV3tA,16
-nucliadb_models-6.5.0.post4426.dist-info/RECORD,,
+nucliadb_models-6.5.0.post4484.dist-info/METADATA,sha256=wLaNzylvsESbAZAeYHJqH6zgEuGSp0gsV5xV3ineLmA,776
+nucliadb_models-6.5.0.post4484.dist-info/WHEEL,sha256=_zCd3N1l69ArxyTb8rzEoP9TpbYXkqRFSNOD5OuxnTs,91
+nucliadb_models-6.5.0.post4484.dist-info/top_level.txt,sha256=UrY1I8oeovIRwkXLYplssTrxQdUjhSEFDFbnwaIV3tA,16
+nucliadb_models-6.5.0.post4484.dist-info/RECORD,,