graphiti-core 0.12.0rc3__py3-none-any.whl → 0.12.0rc5__py3-none-any.whl

graphiti_core/prompts/dedupe_nodes.py

@@ -137,8 +137,12 @@ def nodes(context: dict[str, Any]) -> list[Message]:
         <ENTITIES>
         {json.dumps(context['extracted_nodes'], indent=2)}
         </ENTITIES>
+        
+        <EXISTING ENTITIES>
+        {json.dumps(context['existing_nodes'], indent=2)}
+        </EXISTING ENTITIES>
 
-        For each of the above ENTITIES, determine if the entity is a duplicate of any of its duplication candidates.
+        For each of the above ENTITIES, determine if the entity is a duplicate of any of the EXISTING ENTITIES.
 
         Entities should only be considered duplicates if they refer to the *same real-world object or concept*.
 
@@ -152,9 +156,9 @@ def nodes(context: dict[str, Any]) -> list[Message]:
         For each entity, return the id of the entity as id, the name of the entity as name, and the duplicate_idx
         as an integer.
         
-        - If an entity is a duplicate of one of its duplication_candidates, return the idx of the candidate it is a 
+        - If an entity is a duplicate of one of the EXISTING ENTITIES, return the idx of the candidate it is a 
         duplicate of.
-        - If an entity is not a duplicate of one of its duplication candidates, return the -1 as the duplication_idx
+        - If an entity is not a duplicate of one of the EXISTING ENTITIES, return the -1 as the duplication_idx
         """,
         ),
     ]

graphiti_core/prompts/extract_edges.py

@@ -24,8 +24,8 @@ from .models import Message, PromptFunction, PromptVersion
 
 class Edge(BaseModel):
     relation_type: str = Field(..., description='FACT_PREDICATE_IN_SCREAMING_SNAKE_CASE')
-    source_entity_name: str = Field(..., description='The name of the source entity of the fact.')
-    target_entity_name: str = Field(..., description='The name of the target entity of the fact.')
+    source_entity_id: int = Field(..., description='The id of the source entity of the fact.')
+    target_entity_id: int = Field(..., description='The id of the target entity of the fact.')
     fact: str = Field(..., description='')
     valid_at: str | None = Field(
         None,
@@ -77,7 +77,7 @@ def edge(context: dict[str, Any]) -> list[Message]:
 </CURRENT_MESSAGE>
 
 <ENTITIES>
-{context['nodes']}  # Each has: id, label (e.g., Person, Org), name, aliases
+{context['nodes']} 
 </ENTITIES>
 
 <REFERENCE_TIME>
@@ -94,8 +94,9 @@ Only extract facts that:
 - involve two DISTINCT ENTITIES from the ENTITIES list,
 - are clearly stated or unambiguously implied in the CURRENT MESSAGE,
     and can be represented as edges in a knowledge graph.
-- The FACT TYPES provide a list of the most important types of facts, make sure to extract any facts that
-    could be classified into one of the provided fact types
+- The FACT TYPES provide a list of the most important types of facts, make sure to extract facts of these types
+- The FACT TYPES are not an exhaustive list, extract all facts from the message even if they do not fit into one
+    of the FACT TYPES
 
 You may use information from the PREVIOUS MESSAGES only to disambiguate references or support continuity.
 

graphiti_core/search/search_filters.py

@@ -95,12 +95,12 @@ def edge_search_filter_query_constructor(
             and_filter_query = ''
             for j, and_filter in enumerate(and_filters):
                 and_filter_query += and_filter
-                if j != len(and_filter_query) - 1:
+                if j != len(and_filters) - 1:
                     and_filter_query += ' AND '
 
             valid_at_filter += and_filter_query
 
-            if i == len(or_list) - 1:
+            if i == len(filters.valid_at) - 1:
                 valid_at_filter += ')'
             else:
                 valid_at_filter += ' OR '
@@ -120,12 +120,12 @@ def edge_search_filter_query_constructor(
             and_filter_query = ''
             for j, and_filter in enumerate(and_filters):
                 and_filter_query += and_filter
-                if j != len(and_filter_query) - 1:
+                if j != len(and_filters) - 1:
                     and_filter_query += ' AND '
 
             invalid_at_filter += and_filter_query
 
-            if i == len(or_list) - 1:
+            if i == len(filters.invalid_at) - 1:
                 invalid_at_filter += ')'
             else:
                 invalid_at_filter += ' OR '
@@ -145,12 +145,12 @@ def edge_search_filter_query_constructor(
             and_filter_query = ''
             for j, and_filter in enumerate(and_filters):
                 and_filter_query += and_filter
-                if j != len(and_filter_query) - 1:
+                if j != len(and_filters) - 1:
                     and_filter_query += ' AND '
 
             created_at_filter += and_filter_query
 
-            if i == len(or_list) - 1:
+            if i == len(filters.created_at) - 1:
                 created_at_filter += ')'
             else:
                 created_at_filter += ' OR '
@@ -170,12 +170,12 @@ def edge_search_filter_query_constructor(
             and_filter_query = ''
             for j, and_filter in enumerate(and_filters):
                 and_filter_query += and_filter
-                if j != len(and_filter_query) - 1:
+                if j != len(and_filters) - 1:
                     and_filter_query += ' AND '
 
             expired_at_filter += and_filter_query
 
-            if i == len(or_list) - 1:
+            if i == len(filters.expired_at) - 1:
                 expired_at_filter += ')'
             else:
                 expired_at_filter += ' OR '

graphiti_core/utils/maintenance/edge_operations.py

@@ -92,8 +92,6 @@ async def extract_edges(
     extract_edges_max_tokens = 16384
     llm_client = clients.llm_client
 
-    node_uuids_by_name_map = {node.name: node.uuid for node in nodes}
-
     edge_types_context = (
         [
             {
@@ -109,7 +107,7 @@ async def extract_edges(
     # Prepare context for LLM
     context = {
         'episode_content': episode.content,
-        'nodes': [node.name for node in nodes],
+        'nodes': [{'id': idx, 'name': node.name} for idx, node in enumerate(nodes)],
         'previous_episodes': [ep.content for ep in previous_episodes],
         'reference_time': episode.valid_at,
         'edge_types': edge_types_context,
@@ -160,14 +158,16 @@ async def extract_edges(
         invalid_at = edge_data.get('invalid_at', None)
         valid_at_datetime = None
         invalid_at_datetime = None
-        source_node_uuid = node_uuids_by_name_map.get(edge_data.get('source_entity_name', ''), '')
-        target_node_uuid = node_uuids_by_name_map.get(edge_data.get('target_entity_name', ''), '')
 
-        if source_node_uuid == '' or target_node_uuid == '':
+        source_node_idx = edge_data.get('source_entity_id', -1)
+        target_node_idx = edge_data.get('target_entity_id', -1)
+        if not (-1 < source_node_idx < len(nodes) and -1 < target_node_idx < len(nodes)):
             logger.warning(
-                f'WARNING: source or target node not filled {edge_data.get("edge_name")}. source_node_uuid: {source_node_uuid} and target_node_uuid: {target_node_uuid} '
+                f'WARNING: source or target node not filled {edge_data.get("edge_name")}. source_node_uuid: {source_node_idx} and target_node_uuid: {target_node_idx} '
             )
             continue
+        source_node_uuid = nodes[source_node_idx].uuid
+        target_node_uuid = nodes[edge_data.get('target_entity_id')].uuid
 
         if valid_at:
             try:

graphiti_core/utils/maintenance/node_operations.py

@@ -29,7 +29,7 @@ from graphiti_core.llm_client import LLMClient
 from graphiti_core.llm_client.config import ModelSize
 from graphiti_core.nodes import EntityNode, EpisodeType, EpisodicNode, create_entity_node_embeddings
 from graphiti_core.prompts import prompt_library
-from graphiti_core.prompts.dedupe_nodes import NodeDuplicate, NodeResolutions
+from graphiti_core.prompts.dedupe_nodes import NodeResolutions
 from graphiti_core.prompts.extract_nodes import (
     ExtractedEntities,
     ExtractedEntity,
@@ -241,7 +241,25 @@ async def resolve_extracted_nodes(
         ]
     )
 
-    existing_nodes_lists: list[list[EntityNode]] = [result.nodes for result in search_results]
+    existing_nodes_dict: dict[str, EntityNode] = {
+        node.uuid: node for result in search_results for node in result.nodes
+    }
+
+    existing_nodes: list[EntityNode] = list(existing_nodes_dict.values())
+
+    existing_nodes_context = (
+        [
+            {
+                **{
+                    'idx': i,
+                    'name': candidate.name,
+                    'entity_types': candidate.labels,
+                },
+                **candidate.attributes,
+            }
+            for i, candidate in enumerate(existing_nodes)
+        ],
+    )
 
     entity_types_dict: dict[str, BaseModel] = entity_types if entity_types is not None else {}
 
@@ -255,23 +273,13 @@ async def resolve_extracted_nodes(
                 next((item for item in node.labels if item != 'Entity'), '')
             ).__doc__
             or 'Default Entity Type',
-            'duplication_candidates': [
-                {
-                    **{
-                        'idx': j,
-                        'name': candidate.name,
-                        'entity_types': candidate.labels,
-                    },
-                    **candidate.attributes,
-                }
-                for j, candidate in enumerate(existing_nodes_lists[i])
-            ],
         }
         for i, node in enumerate(extracted_nodes)
     ]
 
     context = {
         'extracted_nodes': extracted_nodes_context,
+        'existing_nodes': existing_nodes_context,
         'episode_content': episode.content if episode is not None else '',
         'previous_episodes': [ep.content for ep in previous_episodes]
         if previous_episodes is not None
@@ -294,8 +302,8 @@ async def resolve_extracted_nodes(
         extracted_node = extracted_nodes[resolution_id]
 
         resolved_node = (
-            existing_nodes_lists[resolution_id][duplicate_idx]
-            if 0 <= duplicate_idx < len(existing_nodes_lists[resolution_id])
+            existing_nodes[duplicate_idx]
+            if 0 <= duplicate_idx < len(existing_nodes)
             else extracted_node
         )
 
@@ -309,70 +317,6 @@ async def resolve_extracted_nodes(
     return resolved_nodes, uuid_map
 
 
-async def resolve_extracted_node(
-    llm_client: LLMClient,
-    extracted_node: EntityNode,
-    existing_nodes: list[EntityNode],
-    episode: EpisodicNode | None = None,
-    previous_episodes: list[EpisodicNode] | None = None,
-    entity_type: BaseModel | None = None,
-) -> EntityNode:
-    start = time()
-    if len(existing_nodes) == 0:
-        return extracted_node
-
-    # Prepare context for LLM
-    existing_nodes_context = [
-        {
-            **{
-                'id': i,
-                'name': node.name,
-                'entity_types': node.labels,
-            },
-            **node.attributes,
-        }
-        for i, node in enumerate(existing_nodes)
-    ]
-
-    extracted_node_context = {
-        'name': extracted_node.name,
-        'entity_type': entity_type.__name__ if entity_type is not None else 'Entity',  # type: ignore
-    }
-
-    context = {
-        'existing_nodes': existing_nodes_context,
-        'extracted_node': extracted_node_context,
-        'entity_type_description': entity_type.__doc__
-        if entity_type is not None
-        else 'Default Entity Type',
-        'episode_content': episode.content if episode is not None else '',
-        'previous_episodes': [ep.content for ep in previous_episodes]
-        if previous_episodes is not None
-        else [],
-    }
-
-    llm_response = await llm_client.generate_response(
-        prompt_library.dedupe_nodes.node(context),
-        response_model=NodeDuplicate,
-        model_size=ModelSize.small,
-    )
-
-    duplicate_id: int = llm_response.get('duplicate_node_id', -1)
-
-    node = (
-        existing_nodes[duplicate_id] if 0 <= duplicate_id < len(existing_nodes) else extracted_node
-    )
-
-    node.name = llm_response.get('name', '')
-
-    end = time()
-    logger.debug(
-        f'Resolved node: {extracted_node.name} is {node.name}, in {(end - start) * 1000} ms'
-    )
-
-    return node
-
-
 async def extract_attributes_from_nodes(
     clients: GraphitiClients,
     nodes: list[EntityNode],

{graphiti_core-0.12.0rc3.dist-info → graphiti_core-0.12.0rc5.dist-info}/METADATA

@@ -1,6 +1,6 @@
 Metadata-Version: 2.3
 Name: graphiti-core
-Version: 0.12.0rc3
+Version: 0.12.0rc5
 Summary: A temporal graph building library
 License: Apache-2.0
 Author: Paul Paliychuk
@@ -22,7 +22,7 @@ Requires-Dist: groq (>=0.2.0) ; extra == "groq"
 Requires-Dist: neo4j (>=5.23.0)
 Requires-Dist: numpy (>=1.0.0)
 Requires-Dist: openai (>=1.53.0)
-Requires-Dist: pydantic (>=2.8.2)
+Requires-Dist: pydantic (>=2.11.5)
 Requires-Dist: python-dotenv (>=1.0.1)
 Requires-Dist: tenacity (>=9.0.0)
 Project-URL: Homepage, https://help.getzep.com/graphiti/graphiti/overview
@@ -236,7 +236,7 @@ Graphiti supports Azure OpenAI for both LLM inference and embeddings. To use Azu
 ```python
 from openai import AsyncAzureOpenAI
 from graphiti_core import Graphiti
-from graphiti_core.llm_client import OpenAIClient
+from graphiti_core.llm_client import LLMConfig, OpenAIClient
 from graphiti_core.embedder.openai import OpenAIEmbedder, OpenAIEmbedderConfig
 from graphiti_core.cross_encoder.openai_reranker_client import OpenAIRerankerClient
 
@@ -252,12 +252,19 @@ azure_openai_client = AsyncAzureOpenAI(
     azure_endpoint=azure_endpoint
 )
 
+# Create LLM Config with your Azure deployed model names
+azure_llm_config = LLMConfig(
+    small_model="gpt-4.1-nano",
+    model="gpt-4.1-mini",
+)
+
 # Initialize Graphiti with Azure OpenAI clients
 graphiti = Graphiti(
     "bolt://localhost:7687",
     "neo4j",
     "password",
     llm_client=OpenAIClient(
+        llm_config=azure_llm_config,
         client=azure_openai_client
     ),
     embedder=OpenAIEmbedder(
@@ -268,6 +275,7 @@ graphiti = Graphiti(
     ),
     # Optional: Configure the OpenAI cross encoder with Azure OpenAI
     cross_encoder=OpenAIRerankerClient(
+        llm_config=azure_llm_config,
         client=azure_openai_client
     )
 )

{graphiti_core-0.12.0rc3.dist-info → graphiti_core-0.12.0rc5.dist-info}/RECORD

@@ -31,10 +31,10 @@ graphiti_core/models/nodes/node_db_queries.py,sha256=AQgRGVO-GgFWfLq1G6k8s86WItw
 graphiti_core/nodes.py,sha256=U19DZ0MIi8GfEsx8D-Jgl8c2SGXO8QovVQpYy6FmUpo,18542
 graphiti_core/prompts/__init__.py,sha256=EA-x9xUki9l8wnu2l8ek_oNf75-do5tq5hVq7Zbv8Kw,101
 graphiti_core/prompts/dedupe_edges.py,sha256=AFVC1EQ0TvNkSp0G7QZmIh3YpGg9FVXo1_sT3TlRqA8,5473
-graphiti_core/prompts/dedupe_nodes.py,sha256=hUdlEUaYUJGLeX6Usy_hfF7fVkaZW-Qhuq5hYrgQ2ZM,7298
+graphiti_core/prompts/dedupe_nodes.py,sha256=OIhMkKexRpQQ0dEr4NW_WE1ta7wLO3RibJA7Ge41uDg,7407
 graphiti_core/prompts/eval.py,sha256=gnBQTmwsCl3Qvwpcm7aieVszzo6y1sMCUT8jQiKTvvE,5317
 graphiti_core/prompts/extract_edge_dates.py,sha256=3Drs3CmvP0gJN5BidWSxrNvLet3HPoTybU3BUIAoc0Y,4218
-graphiti_core/prompts/extract_edges.py,sha256=i7fXBVZ_FH_sAP413T8D02yylIEIia7scaTOuc3dkwY,6497
+graphiti_core/prompts/extract_edges.py,sha256=9NdxAKyXHiFOSuyAzzxRM38BmqtynGEbtmMUr3VTrtM,6513
 graphiti_core/prompts/extract_nodes.py,sha256=EYuX99P8ly7pSOBz87ZA9fJF8V6g6epbVj5Cq0YM8h8,9624
 graphiti_core/prompts/invalidate_edges.py,sha256=yfpcs_pyctnoM77ULPZXEtKW0oHr1MeLsJzC5yrE-o4,3547
 graphiti_core/prompts/lib.py,sha256=DCyHePM4_q-CptTpEXGO_dBv9k7xDtclEaB1dGu7EcI,4092
@@ -46,7 +46,7 @@ graphiti_core/search/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hS
 graphiti_core/search/search.py,sha256=XCEYz4-I341eWiZ-czeFlH5hdbHTTLymhHiD153p6DQ,15122
 graphiti_core/search/search_config.py,sha256=VvKg6AB_RPhoe56DBBXHRBXHThAVJ_OLFCyq_yKof-A,3765
 graphiti_core/search/search_config_recipes.py,sha256=4GquRphHhJlpXQhAZOySYnCzBWYoTwxlJj44eTOavZQ,7443
-graphiti_core/search/search_filters.py,sha256=hT35bKsFyPwocB-8EH2vh6jNev_AXvtdq-l2egdoZbk,6444
+graphiti_core/search/search_filters.py,sha256=jG30nMWX03xoT9ohgyHNu_Xes8GwjIF2eTv6QaiWMqw,6466
 graphiti_core/search/search_helpers.py,sha256=G5Ceaq5Pfgx0Weelqgeylp_pUHwiBnINaUYsDbURJbE,2636
 graphiti_core/search/search_utils.py,sha256=AimBkRgvSFHqAkt1vraTVj_bVAp3JKrR6JUMpoZa8RI,34469
 graphiti_core/utils/__init__.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
@@ -54,13 +54,13 @@ graphiti_core/utils/bulk_utils.py,sha256=WFCfo_OrFD2bpm13Vkex4A1YLVHX4pjCm5acZ1C
 graphiti_core/utils/datetime_utils.py,sha256=Ti-2tnrDFRzBsbfblzsHybsM3jaDLP4-VT2t0VhpIzU,1357
 graphiti_core/utils/maintenance/__init__.py,sha256=vW4H1KyapTl-OOz578uZABYcpND4wPx3Vt6aAPaXh78,301
 graphiti_core/utils/maintenance/community_operations.py,sha256=TF-4eHuvMe_jMqvWg3swxK80zLLtOR0t1pmUUQlNulM,10067
-graphiti_core/utils/maintenance/edge_operations.py,sha256=9_vC3piLUlGM-C30Z4DsN6UWQoxbabsSlJYD7z1zsr4,19222
+graphiti_core/utils/maintenance/edge_operations.py,sha256=5Sq6HgP_s3rutbWov8wK3buEQEL8MDszVosztfmYmcI,19273
 graphiti_core/utils/maintenance/graph_data_operations.py,sha256=BIJKc8tbvU4IjWxLgeotw57b1eE3Iw8YtV74j6eo4RQ,7493
-graphiti_core/utils/maintenance/node_operations.py,sha256=xuXKY0aoe_Idl9Edtb8FxSqoCa45M043nCMraJuAcW8,16606
+graphiti_core/utils/maintenance/node_operations.py,sha256=uChQtyIgBjl9L5nxlO9gYMAeJ8PqU08a63cmiGt36Ss,14796
 graphiti_core/utils/maintenance/temporal_operations.py,sha256=mJkw9xLB4W2BsLfC5POr0r-PHWL9SIfNj_l_xu0B5ug,3410
 graphiti_core/utils/maintenance/utils.py,sha256=47DEQpj8HBSa-_TImW-5JCeuQeRkm5NMpJWZG3hSuFU,0
 graphiti_core/utils/ontology_utils/entity_types_utils.py,sha256=QJX5cG0GSSNF_Mm_yrldr69wjVAbN_MxLhOSznz85Hk,1279
-graphiti_core-0.12.0rc3.dist-info/LICENSE,sha256=KCUwCyDXuVEgmDWkozHyniRyWjnWUWjkuDHfU6o3JlA,11325
-graphiti_core-0.12.0rc3.dist-info/METADATA,sha256=UpVXSKHsEuMupOCGVD6jAuknoTyuhMB9pQEfv0okJ3g,15301
-graphiti_core-0.12.0rc3.dist-info/WHEEL,sha256=fGIA9gx4Qxk2KDKeNJCbOEwSrmLtjWCwzBz351GyrPQ,88
-graphiti_core-0.12.0rc3.dist-info/RECORD,,
+graphiti_core-0.12.0rc5.dist-info/LICENSE,sha256=KCUwCyDXuVEgmDWkozHyniRyWjnWUWjkuDHfU6o3JlA,11325
+graphiti_core-0.12.0rc5.dist-info/METADATA,sha256=K7Hw4oCdcN-1ofwSojPFlBi19P7CILK-IyN4eatafmw,15535
+graphiti_core-0.12.0rc5.dist-info/WHEEL,sha256=fGIA9gx4Qxk2KDKeNJCbOEwSrmLtjWCwzBz351GyrPQ,88
+graphiti_core-0.12.0rc5.dist-info/RECORD,,