From 993b0016bdf56de156dc1081656842753ae79dcd Mon Sep 17 00:00:00 2001
From: Tomaz Bratanic <bratanic.tomaz@gmail.com>
Date: Fri, 13 Dec 2024 20:06:57 +0100
Subject: [PATCH] Deduplicate nodes for gliner graph transformer (#26)

---
 .../graph_transformers/gliner.py              | 24 ++++++++++---------
 1 file changed, 13 insertions(+), 11 deletions(-)

diff --git a/libs/experimental/langchain_experimental/graph_transformers/gliner.py b/libs/experimental/langchain_experimental/graph_transformers/gliner.py
index 85566e2..a7745df 100644
--- a/libs/experimental/langchain_experimental/graph_transformers/gliner.py
+++ b/libs/experimental/langchain_experimental/graph_transformers/gliner.py
@@ -54,21 +54,21 @@ def __init__(
             import gliner_spacy  # type: ignore # noqa: F401
         except ImportError:
             raise ImportError(
-                "Could not import relik python package. "
+                "Could not import gliner-spacy python package. "
                 "Please install it with `pip install gliner-spacy`."
             )
         try:
             import spacy  # type: ignore
         except ImportError:
             raise ImportError(
-                "Could not import relik python package. "
+                "Could not import spacy python package. "
                 "Please install it with `pip install spacy`."
             )
         try:
             import glirel  # type: ignore # noqa: F401
         except ImportError:
             raise ImportError(
-                "Could not import relik python package. "
+                "Could not import gliner python package. "
                 "Please install it with `pip install gliner`."
             )
 
@@ -101,15 +101,17 @@ def process_document(self, document: Document) -> GraphDocument:
                 [(document.page_content, self.allowed_relationships)], as_tuples=True
             )
         )
-        # Convert nodes
-        nodes = []
-        for node in docs[0][0].ents:
-            nodes.append(
-                Node(
-                    id=node.text,
-                    type=node.label_,
-                )
+        # Deduplicate nodes
+        deduplicated_nodes = {(node.text, node.label_) for node in docs[0][0].ents}
+
+        # Step 2: Convert back to Node objects
+        nodes = [
+            Node(
+                id=node_text,
+                type=node_label,
             )
+            for node_text, node_label in deduplicated_nodes
+        ]
         # Convert relationships
         relationships = []
         relations = docs[0][0]._.relations