fix: only exclude confirmed jargon from candidate statistics

CHANGELOG.md

@@ -2,6 +2,21 @@
 
 所有重要更改都将记录在此文件中。
 
+## [3.6.4] - 2026-08-29
+
+### 黑话学习
+
+- 修复黑话被收录后长时间无法解释的问题，修复 [#245](https://github.com/NickCharlie/astrbot_plugin_self_learning/issues/245)：候选统计的排除查询未区分确认状态，把 `is_jargon IS NULL` 的待确认词条一并排除，导致候选出现次数始终无法累加到推断阈值。现显式传入 `only_confirmed=True`，仅排除已确认黑话；待确认候选继续参与统计，重复出现时正常累加 count，达到 3 次后触发含义推断。
+- V2 学习批处理路径同步应用相同的确认状态过滤，避免启用 V2 学习时出现同样的问题；`_get_existing_jargon_terms()` 的方法说明同步更新为「返回已确认词条」。
+
+### 测试
+
+- 新增单元测试覆盖：统计查询正确传递 `only_confirmed=True`；旧版学习流程不再排除待确认候选；V2 黑话批处理流程使用相同的确认状态过滤。
+
+### 版本
+
+- 将插件元数据、运行时包、Dashboard 和文档版本统一提升至 `3.6.4`。
+
 ## [3.6.3] - 2026-08-28
 
 ### 集成兼容性

README.md

@@ -6,7 +6,7 @@
 
 让 AstrBot 在群聊中持续采集、学习、审查并注入上下文，使 Bot 逐步具备表达风格、群组黑话、社交关系、长期记忆和人格演化能力。
 
-[![Version](https://img.shields.io/badge/version-3.6.3-blue.svg)](https://github.com/NickCharlie/astrbot_plugin_self_learning)
+[![Version](https://img.shields.io/badge/version-3.6.4-blue.svg)](https://github.com/NickCharlie/astrbot_plugin_self_learning)
 [![License](https://img.shields.io/badge/license-AGPL--3.0-green.svg)](LICENSE)
 [![AstrBot](https://img.shields.io/badge/AstrBot-%3E%3D4.11.4-orange.svg)](https://github.com/Soulter/AstrBot)
 [![Python](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)

README_EN.md

@@ -14,7 +14,7 @@
 
 <br>
 
-[![Version](https://img.shields.io/badge/version-3.6.3-blue.svg)](https://github.com/NickCharlie/astrbot_plugin_self_learning) [![License](https://img.shields.io/badge/license-AGPL--3.0-green.svg)](LICENSE) [![AstrBot](https://img.shields.io/badge/AstrBot-%3E%3D4.11.4-orange.svg)](https://github.com/Soulter/AstrBot) [![Python](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
+[![Version](https://img.shields.io/badge/version-3.6.4-blue.svg)](https://github.com/NickCharlie/astrbot_plugin_self_learning) [![License](https://img.shields.io/badge/license-AGPL--3.0-green.svg)](LICENSE) [![AstrBot](https://img.shields.io/badge/AstrBot-%3E%3D4.11.4-orange.svg)](https://github.com/Soulter/AstrBot) [![Python](https://img.shields.io/badge/python-3.11%2B-blue.svg)](https://www.python.org/)
 
 [Features](#what-we-can-do) · [Quick Start](#quick-start) · [Web UI](#visual-management-interface) · [Community](#community) · [Contributing](CONTRIBUTING.md)
 

__init__.py

@@ -1,5 +1,5 @@
 # AstrBot 自学习插件
-__version__ = "3.6.3"
+__version__ = "3.6.4"
 
 # Ensure parent namespace packages ("data", "data.plugins") are
 # durably registered in sys.modules.  AstrBot loads plugins via

docs/README.md

@@ -6,7 +6,7 @@ AstrBot 自主学习插件的实现文档和使用文档。
 
 - 插件名: `astrbot_plugin_self_learning`
 - 展示名: `self-learning`
-- 当前元数据版本: `3.6.3`
+- 当前元数据版本: `3.6.4`
 - 最低 AstrBot 版本: `4.11.4`
 - 主要入口: `main.py`
 - 配置入口: `_conf_schema.json`, `config.py`

metadata.yaml

@@ -2,7 +2,7 @@ name: "astrbot_plugin_self_learning"
 author: "NickMo, EterUltimate"
 display_name: "self-learning"
 description: "SELF LEARNING 自主学习插件 — 让 AI 聊天机器人自主学习对话风格、理解群组黑话、管理社交关系与好感度、自适应人格演化，像真人一样自然对话。(使用前必须手动备份人格数据)"
-version: "3.6.3"
+version: "3.6.4"
 repo: "https://github.com/NickCharlie/astrbot_plugin_self_learning"
 tags:
   - "自学习"

services/core_learning/v2_learning_integration.py

@@ -734,6 +734,7 @@ async def _jargon_batch(group_id: str) -> None:
                         rows = await db.get_recent_jargon_list(
                             chat_id=group_id,
                             limit=1000,
+                            only_confirmed=True,
                         )
                         exclude_terms = {
                             str(row.get("content") or "").strip()

services/learning/jargon_learning.py

@@ -164,13 +164,14 @@ async def mine_jargon(self, group_id: str) -> None:
             )
 
     async def _get_existing_jargon_terms(self, group_id: str) -> Set[str]:
-        """Return existing DB jargon/candidate terms for statistical pre-filter exclusion."""
+        """Return confirmed DB jargon terms for statistical pre-filter exclusion."""
         if not self._db_manager or not hasattr(self._db_manager, "get_recent_jargon_list"):
             return set()
         try:
             rows = await self._db_manager.get_recent_jargon_list(
                 chat_id=group_id,
                 limit=1000,
+                only_confirmed=True,
             )
         except Exception as exc:
             logger.debug(

tests/unit/test_jargon_learning.py

@@ -0,0 +1,183 @@
+from types import SimpleNamespace
+from unittest.mock import AsyncMock, Mock
+
+import pytest
+
+from self_learning_EterU.services.core_learning.v2_learning_integration import (
+    V2LearningIntegration,
+)
+from self_learning_EterU.services.learning.jargon_learning import JargonLearningModule
+
+
+@pytest.mark.unit
+@pytest.mark.asyncio
+async def test_jargon_exclusion_only_loads_confirmed_terms():
+    pending = {"content": "待确认", "is_jargon": None}
+    confirmed = {"content": "已确认", "is_jargon": True}
+    calls = []
+
+    async def get_recent_jargon_list(**kwargs):
+        calls.append(kwargs)
+        if kwargs.get("only_confirmed") is True:
+            return [confirmed]
+        return [pending, confirmed]
+
+    module = JargonLearningModule(
+        config=SimpleNamespace(enable_jargon_learning=True),
+        message_collector=None,
+        jargon_miner_manager=None,
+        jargon_statistical_filter=None,
+        db_manager=SimpleNamespace(get_recent_jargon_list=get_recent_jargon_list),
+    )
+
+    terms = await module._get_existing_jargon_terms("group-a")
+
+    assert terms == {"已确认"}
+    assert calls == [
+        {
+            "chat_id": "group-a",
+            "limit": 1000,
+            "only_confirmed": True,
+        }
+    ]
+
+
+@pytest.mark.unit
+@pytest.mark.asyncio
+async def test_mine_jargon_keeps_pending_statistical_candidates_eligible():
+    pending = {"content": "待确认", "is_jargon": None}
+    confirmed = {"content": "已确认", "is_jargon": True}
+    query_calls = []
+    candidate = {
+        "term": "待确认",
+        "context_examples": ["待确认上下文"],
+    }
+
+    async def get_recent_jargon_list(**kwargs):
+        query_calls.append(kwargs)
+        if kwargs.get("only_confirmed") is True:
+            return [confirmed]
+        return [pending, confirmed]
+
+    def get_jargon_candidates(group_id, *, top_k, exclude_terms):
+        assert group_id == "group-a"
+        assert top_k == 20
+        return [] if "待确认" in exclude_terms else [candidate]
+
+    statistical_filter = SimpleNamespace(
+        get_jargon_candidates=Mock(side_effect=get_jargon_candidates),
+    )
+    miner = SimpleNamespace(
+        should_trigger=Mock(return_value=True),
+        run_once=AsyncMock(),
+    )
+    db = SimpleNamespace(
+        get_recent_raw_messages=AsyncMock(
+            return_value=[
+                {"sender_id": "user-a", "message": "这是一条聊天上下文消息"}
+                for _ in range(10)
+            ]
+        ),
+        get_recent_jargon_list=get_recent_jargon_list,
+    )
+    module = JargonLearningModule(
+        config=SimpleNamespace(enable_jargon_learning=True),
+        message_collector=SimpleNamespace(
+            get_statistics=AsyncMock(return_value={"raw_messages": 10})
+        ),
+        jargon_miner_manager=SimpleNamespace(
+            get_or_create_miner=Mock(return_value=miner)
+        ),
+        jargon_statistical_filter=statistical_filter,
+        db_manager=db,
+    )
+
+    await module.mine_jargon("group-a")
+
+    assert query_calls == [
+        {
+            "chat_id": "group-a",
+            "limit": 1000,
+            "only_confirmed": True,
+        }
+    ]
+    statistical_filter.get_jargon_candidates.assert_called_once_with(
+        "group-a",
+        top_k=20,
+        exclude_terms={"已确认"},
+    )
+    miner.run_once.assert_awaited_once()
+    assert miner.run_once.await_args.kwargs["statistical_candidates"] == [candidate]
+
+
+@pytest.mark.unit
+@pytest.mark.asyncio
+async def test_v2_jargon_batch_only_excludes_confirmed_terms():
+    pending = {"content": "待确认", "is_jargon": None}
+    confirmed = {"content": "已确认", "is_jargon": True}
+    query_calls = []
+
+    async def get_recent_jargon_list(**kwargs):
+        query_calls.append(kwargs)
+        if kwargs.get("only_confirmed") is True:
+            return [confirmed]
+        return [pending, confirmed]
+
+    candidate = {"term": "待确认"}
+
+    def get_jargon_candidates(group_id, *, top_k, exclude_terms):
+        assert group_id == "group-a"
+        assert top_k == 20
+        return [] if "待确认" in exclude_terms else [candidate]
+
+    jargon_filter = SimpleNamespace(
+        get_jargon_candidates=Mock(side_effect=get_jargon_candidates),
+    )
+    db = SimpleNamespace(
+        get_recent_jargon_list=get_recent_jargon_list,
+        save_or_update_jargon=AsyncMock(),
+    )
+    llm = SimpleNamespace(
+        generate_response=AsyncMock(return_value="测试释义"),
+    )
+    integration = V2LearningIntegration.__new__(V2LearningIntegration)
+    integration._jargon_filter = jargon_filter
+    integration._llm = llm
+    integration._db = db
+    integration._knowledge_manager = None
+    integration._memory_manager = None
+    integration._exemplar_library = None
+    integration._social_analyzer = None
+    integration._ingestion_buffer = {}
+
+    from self_learning_EterU.services.quality import TieredLearningTrigger
+
+    integration._trigger = TieredLearningTrigger()
+    integration._register_trigger_operations()
+
+    assert await integration._trigger.force_tier2("jargon", "group-a") is True
+
+    assert query_calls == [
+        {
+            "chat_id": "group-a",
+            "limit": 1000,
+            "only_confirmed": True,
+        }
+    ]
+    jargon_filter.get_jargon_candidates.assert_called_once_with(
+        "group-a",
+        top_k=20,
+        exclude_terms={"已确认"},
+    )
+    llm.generate_response.assert_awaited_once()
+    db.save_or_update_jargon.assert_awaited_once_with(
+        "group-a",
+        "待确认",
+        {
+            "meaning": "测试释义",
+            "raw_content": "[]",
+            "is_jargon": True,
+            "count": 1,
+            "is_complete": True,
+        },
+    )

web_src/package.json

@@ -1,6 +1,6 @@
 {
   "name": "astrbot-self-learning-dashboard",
-  "version": "3.6.3",
+  "version": "3.6.4",
   "description": "Solid.js source for the self-learning plugin dashboard",
   "type": "module",
   "scripts": {