From c112af0ab0818024be02d2b42f35a996851b2a7c Mon Sep 17 00:00:00 2001
From: Cyber MacGeddon <cybermaggedon@gmail.com>
Date: Wed, 29 Apr 2026 18:00:55 +0100
Subject: [PATCH] align chunker + googleaistudio fixes with release/v2.4

Master had a parallel sibling fix for issue #821 (PR #828) using
self.RecursiveCharacterTextSplitter / self.TokenTextSplitter; release
branches converged on the bare module-level form.  Adopt release/v2.4's
version so downstream branches don't drift further.
---
 .../trustgraph/chunking/recursive/chunker.py   |  4 ++--
 .../trustgraph/chunking/token/chunker.py       |  4 ++--
 .../text_completion/googleaistudio/llm.py      | 18 +++---------------
 3 files changed, 7 insertions(+), 19 deletions(-)

diff --git a/trustgraph-flow/trustgraph/chunking/recursive/chunker.py b/trustgraph-flow/trustgraph/chunking/recursive/chunker.py
index 098e6111..a0052c79 100755
--- a/trustgraph-flow/trustgraph/chunking/recursive/chunker.py
+++ b/trustgraph-flow/trustgraph/chunking/recursive/chunker.py
@@ -58,7 +58,7 @@ class Processor(ChunkingService):
                          2500, 4000, 6400, 10000, 16000]
             )
 
-        self.text_splitter = self.RecursiveCharacterTextSplitter(
+        self.text_splitter = RecursiveCharacterTextSplitter(
             chunk_size=chunk_size,
             chunk_overlap=chunk_overlap,
             length_function=len,
@@ -111,7 +111,7 @@ class Processor(ChunkingService):
             chunk_overlap = int(chunk_overlap)
 
         # Create text splitter with effective parameters
-        text_splitter = self.RecursiveCharacterTextSplitter(
+        text_splitter = RecursiveCharacterTextSplitter(
             chunk_size=chunk_size,
             chunk_overlap=chunk_overlap,
             length_function=len,
diff --git a/trustgraph-flow/trustgraph/chunking/token/chunker.py b/trustgraph-flow/trustgraph/chunking/token/chunker.py
index 3bf907a4..c3935e4b 100755
--- a/trustgraph-flow/trustgraph/chunking/token/chunker.py
+++ b/trustgraph-flow/trustgraph/chunking/token/chunker.py
@@ -56,7 +56,7 @@ class Processor(ChunkingService):
                          2500, 4000, 6400, 10000, 16000]
             )
 
-        self.text_splitter = self.TokenTextSplitter(
+        self.text_splitter = TokenTextSplitter(
             encoding_name="cl100k_base",
             chunk_size=chunk_size,
             chunk_overlap=chunk_overlap,
@@ -108,7 +108,7 @@ class Processor(ChunkingService):
             chunk_overlap = int(chunk_overlap)
 
         # Create text splitter with effective parameters
-        text_splitter = self.TokenTextSplitter(
+        text_splitter = TokenTextSplitter(
             encoding_name="cl100k_base",
             chunk_size=chunk_size,
             chunk_overlap=chunk_overlap,
diff --git a/trustgraph-vertexai/trustgraph/model/text_completion/googleaistudio/llm.py b/trustgraph-vertexai/trustgraph/model/text_completion/googleaistudio/llm.py
index 142fc45c..b01ff410 100644
--- a/trustgraph-vertexai/trustgraph/model/text_completion/googleaistudio/llm.py
+++ b/trustgraph-vertexai/trustgraph/model/text_completion/googleaistudio/llm.py
@@ -43,18 +43,6 @@ class Processor(LlmService):
         temperature = params.get("temperature", default_temperature)
         max_output = params.get("max_output", default_max_output)
 
-        from google import genai                                               
-        from google.genai import types                                          
-        from google.genai.types import HarmCategory, HarmBlockThreshold        
-        from google.genai.errors import ClientError                             
-        from google.api_core.exceptions import ResourceExhausted               
-        self.genai = genai                                                      
-        self.types = types                                                      
-        self.HarmCategory = HarmCategory                                        
-        self.HarmBlockThreshold = HarmBlockThreshold                           
-        self.ClientError = ClientError                                          
-        self.ResourceExhausted = ResourceExhausted                             
-
         if api_key is None:
             raise RuntimeError("Google AI Studio API key not specified")
 
@@ -66,7 +54,7 @@ class Processor(LlmService):
             }
         )
 
-        self.client = self.genai.Client(api_key=api_key, vertexai=False)
+        self.client = genai.Client(api_key=api_key, vertexai=False)
         self.default_model = model
         self.temperature = temperature
         self.max_output = max_output
@@ -74,7 +62,7 @@ class Processor(LlmService):
         # Cache for generation configs per model
         self.generation_configs = {}
 
-        block_level = self.HarmBlockThreshold.BLOCK_ONLY_HIGH
+        block_level = HarmBlockThreshold.BLOCK_ONLY_HIGH
 
         self.safety_settings = [
             types.SafetySetting(                           
@@ -159,7 +147,7 @@ class Processor(LlmService):
 
             return resp
 
-        except self.ResourceExhausted as e:
+        except ResourceExhausted as e:
 
             logger.warning("Rate limit exceeded")