refactor: refactor typo corrector

.gitignore

@@ -9,3 +9,6 @@ manifest.ini
 *.nvda-addon
 .sconsign.dblite
 /[0-9]*.[0-9]*.[0-9]*.json
+venv
+.env
+.python-version

addon/globalPlugins/WordBridge/configManager.py

@@ -12,7 +12,7 @@
 	"Google": _("Google"),
 	"OpenAI": _("OpenAI"),
 	"OpenRouter": _("OpenRouter"),
-	"claude-3-5-haiku-20241022": _("claude-3-5-haiku"),
+	"claude-haiku-4-5-20251001": _("claude-4-5-haiku"),
 	"claude-3-7-sonnet-20250219": _("claude-3.7-sonnet"),
 	"claude-sonnet-4-20250514": _("claude-4-sonnet"),
 	"deepseek-v3": _("deepseek-v3"),

addon/globalPlugins/WordBridge/corrector_config/Anthropic-00002-claude-4-5-haiku.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "claude-haiku-4-5-20251001",
+		"provider": "Anthropic",
+		"llm_access_method": "personal_api_key",
+		"require_secret_key": false,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": true,
+			"no_explanation": false
+		}
+	}
+}

addon/globalPlugins/WordBridge/corrector_config/Google-00002-gemini-2.5-pro-preview.json → addon/globalPlugins/WordBridge/corrector_config/Google-00001-gemini-2.5-flash.json

@@ -1,6 +1,6 @@
 {
 	"model": {
-		"model_name": "gemini-2.5-pro-preview-06-05",
+		"model_name": "gemini-2.5-flash",
 		"provider": "Google",
 		"llm_access_method": "personal_api_key",
 		"require_secret_key": false,
@@ -13,4 +13,4 @@
 			"no_explanation": false
 		}
 	}
-}
+}

addon/globalPlugins/WordBridge/corrector_config/Google-00002-gemini-2.5-pro.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "gemini-2.5-pro",
+		"provider": "Google",
+		"llm_access_method": "personal_api_key",
+		"require_secret_key": false,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": true,
+			"no_explanation": false
+		}
+	}
+}

addon/globalPlugins/WordBridge/lib/llm/__init__.py

@@ -0,0 +1,5 @@
+from .cost_calculator import CostCalculator
+
+__all__ = [
+    'CostCalculator',
+]

addon/globalPlugins/WordBridge/lib/llm/cost_calculator.py

@@ -0,0 +1,44 @@
+"""
+Cost Calculator for tracking LLM usage and costs.
+
+This module provides utilities for:
+- Tracking token usage from API responses
+- Calculating costs based on model pricing
+"""
+
+from decimal import Decimal
+from collections import defaultdict
+
+class CostCalculator:
+	def __init__(self, model_entry: dict):
+		self._model_entry = model_entry
+		self._pricing = model_entry.get("pricing", {})
+		self._usage_key = model_entry.get("usage_key")
+
+	def get_total_usage(self, response_history: list) -> dict:
+		total_usage = defaultdict(int)
+		if not self._usage_key:
+			return total_usage
+
+		for response in response_history:
+			if isinstance(response, dict) and self._usage_key in response:
+				for usage_type in self._pricing:
+					if usage_type == "base_unit":
+						continue
+					try:
+						total_usage[usage_type] += response[self._usage_key][usage_type]
+					except KeyError:
+						pass
+
+		return dict(total_usage)
+
+	def get_total_cost(self, response_history: list) -> Decimal:
+		cost = Decimal("0")
+		usages = self.get_total_usage(response_history)
+		for key, value in usages.items():
+			cost += (
+				Decimal(str(self._pricing[key]))
+				* Decimal(str(value))
+				/ Decimal(str(self._pricing["base_unit"]))
+			)
+		return cost

addon/globalPlugins/WordBridge/lib/typo_corrector.py

@@ -4,6 +4,7 @@
 from decimal import Decimal
 from threading import Thread
 from typing import Any, Dict, List, Tuple
+from pathlib import Path
 
 import json
 import logging
@@ -18,6 +19,7 @@
 from .utils import get_char_pinyin, has_chinese, has_simplified_chinese_char, has_traditional_chinese_char
 from .utils import PUNCTUATION, SEPERATOR, is_chinese_character, strings_diff, text_segmentation
 from .utils import find_correction_errors, review_correction_errors, get_segments_to_recorrect
+from .llm import CostCalculator
 
 import chinese_converter
 
@@ -54,124 +56,6 @@ class BaseTypoCorrector():
 		"google": GoogleProvider,
 		"openrouter": OpenrouterProvider,
 	}
-	MODEL = {
-		"claude-3-5-haiku-20241022": {
-			"usage_key": "usage",
-			"input_tokens": "0.8",
-			"cache_creation_input_tokens": "1",
-			"cache_read_input_tokens": "0.08",
-			"output_tokens": "4",
-			"base_unit": "1000000"
-		},
-		"claude-3-7-sonnet-20250219": {
-			"usage_key": "usage",
-			"input_tokens": "3",
-			"cache_creation_input_tokens": "3.75",
-			"cache_read_input_tokens": "0.3",
-			"output_tokens": "15",
-			"base_unit": "1000000"
-		},
-		"claude-sonnet-4-20250514": {
-			"usage_key": "usage",
-			"input_tokens": "3",
-			"cache_creation_input_tokens": "3.75",
-			"cache_read_input_tokens": "0.3",
-			"output_tokens": "15",
-			"base_unit": "1000000"
-		},
-		"deepseek-v3": {},
-		"deepseek-chat": {
-			"usage_key": "usage",
-			"prompt_cache_hit_tokens": "0.07",
-			"prompt_cache_miss_tokens": "0.27",
-			"completion_tokens": "1.1",
-			"base_unit": "1000000"
-		},
-		"deepseek-reasoner": {
-			"usage_key": "usage",
-			"prompt_cache_hit_tokens": "0.14",
-			"prompt_cache_miss_tokens": "0.55",
-			"completion_tokens": "2.19",
-			"base_unit": "1000000"
-		},
-		"deepseek/deepseek-chat:free": {},
-		"deepseek/deepseek-chat-v3-0324:free": {},
-		"deepseek/deepseek-r1-0528:free": {},
-		"deepseek/deepseek-r1-0528-qwen3-8b:free": {},
-		"gemini-2.5-flash-preview-05-20": {
-			"usage_key": "usageMetadata",
-			"promptTokenCount": "0.15",
-			"candidatesTokenCount": "0.6",
-			"base_unit": "1000000"
-		},
-		"gemini-2.5-pro-preview-06-05": {
-			"usage_key": "usageMetadata",
-			"promptTokenCount": "1.25",
-			"candidatesTokenCount": "10",
-			"base_unit": "1000000"
-		},
-		"gpt-4o-2024-08-06": {
-			"usage_key": "usage",
-			"prompt_tokens": "2.5",
-			"completion_tokens": "10",
-			"base_unit": "1000000"
-		},
-		"gpt-4o-mini-2024-07-18": {
-			"usage_key": "usage",
-			"prompt_tokens": "0.15",
-			"completion_tokens": "0.6",
-			"base_unit": "1000000"
-		},
-		"gpt-4.1-2025-04-14": {
-			"usage_key": "usage",
-			"prompt_tokens": "2",
-			"completion_tokens": "8",
-			"base_unit": "1000000"
-		},
-		"gpt-4.1-mini-2025-04-14": {
-			"usage_key": "usage",
-			"prompt_tokens": "0.4",
-			"completion_tokens": "1.6",
-			"base_unit": "1000000"
-		},
-		"gpt-4.1-nano-2025-04-14": {
-			"usage_key": "usage",
-			"prompt_tokens": "0.1",
-			"completion_tokens": "0.4",
-			"base_unit": "1000000"
-		},
-		"o4-mini-2025-04-16": {
-			"usage_key": "usage",
-			"prompt_tokens": "1.1",
-			"completion_tokens": "4.4",
-			"base_unit": "1000000"
-		},
-		"gpt-5-chat-latest": {
-			"usage_key": "usage",
-			"prompt_tokens": "1.25",
-			"completion_tokens": "10",
-			"base_unit": "1000000",
-		},
-		"gpt-5": {
-			"usage_key": "usage",
-			"prompt_tokens": "1.25",
-			"completion_tokens": "10",
-			"base_unit": "1000000",
-		},
-		"gpt-5-mini": {
-			"usage_key": "usage",
-			"prompt_tokens": "0.25",
-			"completion_tokens": "2",
-			"base_unit": "1000000",
-		},
-		"gpt-5-nano": {
-			"usage_key": "usage",
-			"prompt_tokens": "0.05",
-			"completion_tokens": "0.4",
-			"base_unit": "1000000",
-		},
-		"ernie-4.0-turbo-8k": {},
-	}
 
 	def __init__(
 		self,
@@ -210,6 +94,16 @@ def __init__(
 		self.question_string = ""
 		self.answer_string = ""
 
+		config_key = f"{model}&{provider}"
+		model_entry = self._load_model_config(config_key)
+		self._cost_calculator = CostCalculator(model_entry)
+
+	def _load_model_config(self, config_key: str) -> dict:
+		config_path = Path(__file__).parent.parent / "setting" / "llm_models.json"
+		with open(config_path, "r", encoding="utf8") as f:
+			config = json.load(f)
+		return config.get(config_key, {})
+
 	def correct_text(self, text: str, batch_mode: bool = True, fake_corrected_text: str = None) -> Tuple:
 		"""
 		Analyze typos of text using self.segment_corrector. It also analyzes the difference between the original
@@ -342,43 +236,10 @@ def correct_segment_batch(self, input_text_list: list, previous_results_list: li
 		return output_text_list
 
 	def get_total_usage(self) -> Dict:
-		"""
-		Get the total usage of OpenAI model (in tokens)
-
-		Returns:
-			The total usage of OpenAI model (in tokens)
-		"""
-		usage_key = self.MODEL[self.model].get("usage_key")
-		total_usage = defaultdict(int)
-		if not usage_key:
-			return total_usage
-
-		for response in self.response_history:
-			if isinstance(response, dict) and usage_key in response:
-				for usage_type in set(self.MODEL[self.model].keys()):
-					if usage_type == "base_unit" or usage_type == "usage_key":
-						continue
-					try:
-						total_usage[usage_type] += response[usage_key][usage_type]
-					except KeyError:
-						pass
-
-		return total_usage
-
-	def get_total_cost(self) -> int:
-		"""
-		Get the total cost of provider model (in USD)
-
-		Returns:
-			The total cost of provider model (in USDs)
-		"""
-		price_info = self.MODEL[self.model]
-		cost = Decimal("0")
-		usages = self.get_total_usage()
-		for key, value in usages.items():
-			cost += Decimal(price_info[key]) * Decimal(str(value)) / Decimal(price_info["base_unit"])
+		return self._cost_calculator.get_total_usage(self.response_history)
 
-		return cost
+	def get_total_cost(self) -> Decimal:
+		return self._cost_calculator.get_total_cost(self.response_history)
 
 	def _correct_segment_task(
 		self,

addon/globalPlugins/WordBridge/corrector_config/Anthropic-00002-claude-3-5-haiku.json → addon/globalPlugins/WordBridge/setting/corrector/Anthropic-00001-claude-4-sonnet.json

@@ -1,6 +1,6 @@
 {
 	"model": {
-		"model_name": "claude-3-5-haiku-20241022",
+		"model_name": "claude-sonnet-4-20250514",
 		"provider": "Anthropic",
 		"llm_access_method": "personal_api_key",
 		"require_secret_key": false,

addon/globalPlugins/WordBridge/setting/corrector/Anthropic-00002-claude-4-5-haiku.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "claude-haiku-4-5-20251001",
+		"provider": "Anthropic",
+		"llm_access_method": "personal_api_key",
+		"require_secret_key": false,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": true,
+			"no_explanation": false
+		}
+	}
+}

addon/globalPlugins/WordBridge/setting/corrector/Baidu-00001-deepseek-v3.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "deepseek-v3",
+		"provider": "Baidu",
+		"llm_access_method": "personal_api_key",
+		"require_secret_key": true,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": false,
+			"no_explanation": true
+		}
+	}
+}

addon/globalPlugins/WordBridge/setting/corrector/Baidu-00002-ernie-4.0-turbo-8k.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "ernie-4.0-turbo-8k",
+		"provider": "Baidu",
+		"llm_access_method": "personal_api_key",
+		"require_secret_key": true,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": false,
+			"no_explanation": true
+		}
+	}
+}

addon/globalPlugins/WordBridge/setting/corrector/Coseeing-00001-deepseek-chat-v3-0324.json

@@ -0,0 +1,16 @@
+{
+	"model": {
+		"model_name": "deepseek/deepseek-chat-v3-0324:free",
+		"provider": "OpenRouter",
+		"llm_access_method": "coseeing_relay",
+		"require_secret_key": true,
+		"template_name": {
+			"standard": "Standard_v1.json",
+			"lite": "Lite_v1.json"
+		},
+		"optional_guidance_enable": {
+			"keep_non_chinese_char": false,
+			"no_explanation": true
+		}
+	}
+}