Add missing field declarations to dataclass child classes

Co-authored-by: ChenZiHong-Gavin <58508660+ChenZiHong-Gavin@users.noreply.github.com>

Author: Copilot

graphgen/models/generator/aggregated_generator.py

@@ -1,7 +1,7 @@
 from dataclasses import dataclass
 from typing import Any
 
-from graphgen.bases import BaseGenerator
+from graphgen.bases import BaseGenerator, BaseLLMClient
 from graphgen.templates import AGGREGATED_GENERATION_PROMPT
 from graphgen.utils import compute_content_hash, detect_main_language, logger
 
@@ -15,6 +15,8 @@ class AggregatedGenerator(BaseGenerator):
     2. question generation: Generate relevant questions based on the rephrased text.
     """
 
+    llm_client: BaseLLMClient = None
+
     @staticmethod
     def build_prompt(
         batch: tuple[list[tuple[str, dict]], list[tuple[Any, Any, dict]]]

graphgen/models/generator/atomic_generator.py

@@ -1,13 +1,15 @@
 from dataclasses import dataclass
 from typing import Any
 
-from graphgen.bases import BaseGenerator
+from graphgen.bases import BaseGenerator, BaseLLMClient
 from graphgen.templates import ATOMIC_GENERATION_PROMPT
 from graphgen.utils import compute_content_hash, detect_main_language, logger
 
 
 @dataclass
 class AtomicGenerator(BaseGenerator):
+    llm_client: BaseLLMClient = None
+
     @staticmethod
     def build_prompt(
         batch: tuple[list[tuple[str, dict]], list[tuple[Any, Any, dict]]]

graphgen/models/generator/cot_generator.py

@@ -1,13 +1,15 @@
 from dataclasses import dataclass
 from typing import Any
 
-from graphgen.bases import BaseGenerator
+from graphgen.bases import BaseGenerator, BaseLLMClient
 from graphgen.templates import COT_GENERATION_PROMPT
 from graphgen.utils import compute_content_hash, detect_main_language, logger
 
 
 @dataclass
 class CoTGenerator(BaseGenerator):
+    llm_client: BaseLLMClient = None
+
     @staticmethod
     def build_prompt(
         batch: tuple[list[tuple[str, dict]], list[tuple[Any, Any, dict]]]

graphgen/models/generator/multi_hop_generator.py

@@ -1,13 +1,15 @@
 from dataclasses import dataclass
 from typing import Any
 
-from graphgen.bases import BaseGenerator
+from graphgen.bases import BaseGenerator, BaseLLMClient
 from graphgen.templates import MULTI_HOP_GENERATION_PROMPT
 from graphgen.utils import compute_content_hash, detect_main_language, logger
 
 
 @dataclass
 class MultiHopGenerator(BaseGenerator):
+    llm_client: BaseLLMClient = None
+
     @staticmethod
     def build_prompt(
         batch: tuple[list[tuple[str, dict]], list[tuple[Any, Any, dict]]]

graphgen/models/tokenizer/hf_tokenizer.py

@@ -8,6 +8,8 @@
 
 @dataclass
 class HFTokenizer(BaseTokenizer):
+    model_name: str = "cl100k_base"
+
     def __post_init__(self):
         self.enc = AutoTokenizer.from_pretrained(self.model_name)
 

graphgen/models/tokenizer/tiktoken_tokenizer.py

@@ -8,6 +8,8 @@
 
 @dataclass
 class TiktokenTokenizer(BaseTokenizer):
+    model_name: str = "cl100k_base"
+
     def __post_init__(self):
         self.enc = tiktoken.get_encoding(self.model_name)