Cleanup adapter

Author: YoanSallami

coverage-badge.svg

@@ -447,7 +447,7 @@ The modular architecture enables teams to iteratively refine their extraction pi
 
 By separating concerns across different stages, Synalinks allows for specialized optimization at each level of the extraction hierarchy. Entity extraction can leverage lightweight, fast models, while complex reasoning tasks can utilize more powerful, specialized models. This granular control over computational resources leads to more efficient systems that deliver better results at lower costs.
 
-Whether you're building a simple entity extraction system or a comprehensive knowledge discovery platform, Synalinks provides the flexibility, control, and scalability needed to transform unstructured data into actionable insights. 
+Whether you're building a simple entity extraction system or a comprehensive knowledge discovery platform, Synalinks provides the flexibility, control, and scalability needed to transform unstructured data into actionable insights.
 
 In an era where structured data is the new oil, Synalinks provides the refinery that transforms raw information into structured knowledge, enabling organizations to unlock the full potential of their data assets through intelligent, adaptive extraction pipelines.
 

docs/Code Examples/Advanced/Knowledge Extraction.md

@@ -102,28 +102,6 @@ def __init__(
                 self.query("CALL db.awaitIndexes(300)"),
             )
 
-    def sanitize_label(self, label):
-        """Sanitize labels for Neo4j compatibility"""
-        if not label:
-            return "Entity"
-        # Remove special characters and replace spaces with underscores
-        sanitized = re.sub(r"[^a-zA-Z0-9_]", "_", str(label))
-        # Ensure it starts with a letter or underscore
-        if sanitized and not sanitized[0].isalpha() and sanitized[0] != "_":
-            sanitized = "_" + sanitized
-        return sanitized
-
-    def sanitize_properties(self, properties):
-        """Sanitize property names and values for Neo4j compatibility"""
-        if not isinstance(properties, dict):
-            return {}
-
-        sanitized = {}
-        for key, value in properties.items():
-            # Sanitize key names
-            sanitized[self.sanitize(key)] = value
-        return sanitized
-
     async def query(self, query: str, params: Dict[str, Any] = None):
         driver = neo4j.GraphDatabase.driver(
             self.index_name, auth=(self.username, self.password)

synalinks/src/knowledge_bases/database_adapters/neo4j_adapter.py

@@ -117,7 +117,7 @@ async def similarity_search(
         self,
         similarity_search,
         k=10,
-        threshold=0.7,
+        threshold=0.8,
     ):
         """Perform similarity search to find entities similar to the given text.
 
@@ -130,7 +130,7 @@ async def similarity_search(
                 Defaults to 10.
             threshold (float): Minimum similarity score for results.
                 Entities with similarity below this threshold are excluded.
-                Should be between 0.0 and 1.0. Defaults to 0.7.
+                Should be between 0.0 and 1.0 (Defaults to 0.8).
         """
         return await self.adapter.similarity_search(
             similarity_search,
@@ -142,7 +142,7 @@ async def triplet_search(
         self,
         triplet_search,
         k=10,
-        threshold=0.7,
+        threshold=0.8,
     ):
         """Search for triplets in the knowledge graph.
 
@@ -155,7 +155,7 @@ async def triplet_search(
                 (Defaults to 10).
             threshold (float, optional): Minimum similarity score for triplet matches.
                 Triplets with similarity below this threshold are excluded.
-                Should be between 0.0 and 1.0. Defaults to 0.7.
+                Should be between 0.0 and 1.0. (Defaults to 0.8).
         """
         return await self.adapter.triplet_search(
             triplet_search,