myssia-classifier 0.0.1__tar.gz

myssia_classifier-0.0.1/.gitignore

@@ -0,0 +1,113 @@
+# ===== 安全红线:私有数据零上传 =====
+# 本仓库与 .trellis/ 都会公开;任何凭据、抓取数据、密钥文件不得入库。
+# MYIA 凭据只走 env:/keychain: 引用,配置文件出现明文凭据即拒跑(spec 铁律)。
+
+# ---- 密钥与凭据文件(防误拷入仓库)----
+*.pem
+*.key
+*.p12
+*.pfx
+secrets.*
+# 例外:核心源码模块不叫"secrets 文件",必须入库(fresh clone / wheel 缺它即崩)
+!src/myssia/secrets.py
+.env
+.env.*
+.envrc
+
+# ---- 本地私有路径索引(真实姓名/内网拓扑,仅本机)----
+LOCAL-NOTES.md
+
+# ---- MYIA 运行时数据(抓取的情报内容属私有数据)----
+myssia.db
+# 改名过渡期旧缺省名落下的运行时库(桌面 serve 仓库 cwd 直跑),同零容忍口径
+/myia.db
+*.db-journal
+*.db-wal
+*.db-shm
+data/
+# 例外:随包分发的程序数据(prompt/关键词表)不是抓取数据,必须入库与进 wheel
+# (classify 词表已随 10-01-v10-classifier-pypi 拆去 myssia-classifier 包,例外同步删;
+#  模块改名后路径须跟 src/myssia,否则 hatchling 按 gitignore 吞数据文件进不了 wheel)
+!src/myssia/enrich/data/
+!myssia-classifier/myssia_classifier/data/
+# credhunter 指纹库数据(20 发现查询包 + 25 验证规格)同上:程序数据非抓取
+# 产物,fresh clone/CI 缺它即三 lane 启动全坏(10-03-aipocket-fusion R3)。
+!plugins/myssia-credhunter/credhunter/data/
+out/
+output/
+exports/
+backups/
+/logs/
+*.log
+# 场景插件运行时产物(osint 类插件的侦察情报:邮箱/密钥泄漏指纹等)
+# 与抓取数据同级零容忍:插件安装根可在仓库树内(repo 内 compose 副本原地跑、
+# MYIA_PLUGIN_DIR 指向仓库),loot/ 落进工作树就必须被防线覆盖。
+plugins/**/loot/
+
+# ---- Python ----
+__pycache__/
+*.py[cod]
+.venv/
+venv/
+dist/
+build/
+*.egg-info/
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.coverage
+htmlcov/
+coverage.xml
+
+# ---- 工作流运行态(本机私有,勿提交;.zcode/workflows/ 保存的工作流除外)----
+.zcode/workflow-drafts/
+.zcode/workflow-runs/
+
+# ---- Node / 桌面端(desktop/ 是打包正式出处;构建产物可达 ~1GB,严禁入库)----
+node_modules/
+desktop/src-tauri/target/
+desktop/src-tauri/gen/
+desktop/.venv-build/
+desktop/build-pyi/
+desktop/.pyinstaller-cache/
+desktop/ui/
+desktop/src-tauri/binaries/
+
+# ---- GitNexus 本地索引(知识图谱 DB,本机私有;WAL 损坏勿提交)----
+.gitnexus/
+# gitnexus analyze 的自动副作用副本:MYIA 用全局技能(~/.agents/skills)+ AGENTS.md,不提交
+# (将来要手写 CLAUDE.md 时,先移除下一行规则)
+.claude/
+CLAUDE.md
+
+# ---- 编辑器 / OS ----
+.idea/
+.vscode/
+.DS_Store
+*.bak
+*~
+Thumbs.db
+
+# ---- 冒烟运行态(本机私有,勿提交)----
+.zcode/smoke/
+
+# ---- 技能工具侧副本(本机私有,勿提交)----
+# .zcode/skills/ 下 trellis-* 系列是仓库内容(已跟踪,trellis 工作流件);
+# 其余(frontend-design/tauri/webapp-testing 等)是技能工具落盘的本地副本
+# (与 ~/.agents/skills 全局件同源),不入库——只挡副本,不误伤 trellis-*。
+.zcode/skills/*
+!.zcode/skills/trellis-*/
+
+# ---- 仓库根误落的运行时产物(防误提交)----
+# 消息屏目录/别名是运行态数据,正常应落数据根(MYIA_HOME);desktop serve 在
+# 仓库 cwd 下直跑时缺省写到仓库根——内容含真实平台 chat_id 与群名(私有
+# 标识,同 myssia.db 零容忍口径),且会随每次 discover 变化,严禁入库。
+/channel_aliases.json
+/channel_directory.json
+/channel_dead.json
+# credhunter 全文密钥库(指纹→原文,含真实密钥——比 channel_* 更敏感,
+# 零容忍口径同 myssia.db):正常应落数据根(MYIA_HOME),引擎缺省路径在仓库
+# cwd 直跑时会落仓库根,严禁入库(10-03-aipocket-fusion Q9)。
+/credhunter-keystore.json
+# 技能工具链的本地锁文件(本机技能版本快照,非仓库配置)
+/skills-lock.json

myssia_classifier-0.0.1/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 xinzhuzi
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

myssia_classifier-0.0.1/PKG-INFO

@@ -0,0 +1,98 @@
+Metadata-Version: 2.5
+Name: myssia-classifier
+Version: 0.0.1
+Summary: Zero-dependency seven-category keyword classifier with free/paid dual-signal adjudication, data-driven keyword tables, and a safe custom-rule API. 七大类双信号关键词分类引擎。
+Project-URL: Homepage, https://github.com/xinzhuzi/myia
+Project-URL: Repository, https://github.com/xinzhuzi/myia
+Project-URL: Issues, https://github.com/xinzhuzi/myia/issues
+Author: xinzhuzi
+License: MIT
+License-File: LICENSE
+Keywords: chinese,classifier,information-filtering,keyword-matching,myia,rule-engine,text-classification
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Natural Language :: Chinese (Simplified)
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Topic :: Text Processing :: Filters
+Classifier: Typing :: Typed
+Requires-Python: >=3.11
+Description-Content-Type: text/markdown
+
+# myssia-classifier
+
+myssia 七大类关键词分类引擎,独立 pip 包:零运行时依赖、免费/付费双信号裁决、数据与代码分离、命中可追溯。
+
+从 myssia 核心的分类层(`myssia.classify`)拆出,忠实移植生产爬虫关键词分类器(2026-09-28 纠错版,金测集钉住语义)。可脱离 myssia 单独引用,myssia 核心也以固定版本范围依赖本包。
+
+## 特性
+
+- **零依赖**:纯 Python 标准库(3.11+),`pip install` 即跑,无任何传递依赖
+- **双信号裁决**:免费信号绝对优先于付费信号;注册引流过滤;死词/噪声词过滤
+- **数据与代码分离**:词表是包内 `data/keywords.json`,调关键词/正则不用改代码
+- **命中追溯**:每条结果都带命中明细(哪张表、哪个关键词/正则、命中区间)
+- **自定义规则 API**:`when` 表达式走白名单 AST,永不 `eval`,未知构造加载期即拒
+- **类型完备**:全量类型标注(`py.typed`),mypy strict 友好
+
+## 安装
+
+尚未上架 PyPI;在那之前从 myssia 仓库以 uv workspace 安装(`git clone https://github.com/xinzhuzi/myia && cd myssia && uv sync`),PyPI 上架后即可 `pip install myssia-classifier`。
+
+## 最小示例
+
+```python
+from myssia_classifier import classify_item
+
+result = classify_item({"title": "Claude 免费领100$"})
+print(result.category)  # "token"
+print(result.tags)      # 自定义规则命中的 tag(无规则时为 [])
+for evidence in result.matched:
+    print(evidence.describe())  # 命中追溯: 表:类型=值
+```
+
+七大类:`credit-card` / `proxy-node` / `buying-agent` / `server` / `token` / `ai-news` / `freebie`;另有第八个生产标签 `channel`(付费信号且无免费信号的导购/渠道帖)。被过滤标题的 `category` 为 `None`。
+
+## 自定义规则
+
+规则读取条目字段(字典键或对象属性),在粗分类之上打 tag:
+
+```python
+from myssia_classifier import Rule, classify_item
+
+rule = Rule(name="big-move", when="abs(change_pct) >= 3", tag="大波动")
+result = classify_item({"title": "NVDA 盘前大跌", "change_pct": -4.2}, rules=[rule])
+print(result.tags)  # ["大波动"]
+```
+
+`when` 表达式支持的语法(白名单 AST):字面量、字段名、list/tuple、四则与比较运算、`and`/`or`/`not`、白名单函数(`abs/min/max/round/len/int/float/str`)。属性访问、下标、lambda、f-string、推导式、任意调用等一律在构造期抛 `RuleSyntaxError`。运行期单条规则失败只按"未命中"处理并记 WARNING,绝不打断整批评测。
+
+## 自定义词表
+
+默认词表随包分发(`myssia_classifier/data/keywords.json`);也可以加载自己的 JSON:
+
+```python
+from myssia_classifier import load_table, classify_title
+
+table = load_table()  # 缺省即打包词表;自定义时传自己的 JSON 路径(结构非法时结构化报错)
+print(classify_title("便宜出极速服务器三台", table).category)  # "server"(打包词表实测命中)
+```
+
+词表 JSON 支持在 `keywords`(普通关键词)与 `patterns`(正则)两列增改,`case_insensitive` 按节可调;顶层未知字段与断裂引用(如 `fallback_target` 指向不存在的类目)加载即报错,不做静默忽略。
+
+## 从 YAML 加载规则
+
+`load_rules` 可直接读取 myssia 插件 YAML 的 `classify.rules` 节(需要环境里装有 PyYAML;本包不把它列为依赖):
+
+```python
+from myssia_classifier import load_rules
+
+rules = load_rules("plugin.yaml")
+```
+
+## 许可
+
+MIT

myssia_classifier-0.0.1/README.md

@@ -0,0 +1,73 @@
+# myssia-classifier
+
+myssia 七大类关键词分类引擎,独立 pip 包:零运行时依赖、免费/付费双信号裁决、数据与代码分离、命中可追溯。
+
+从 myssia 核心的分类层(`myssia.classify`)拆出,忠实移植生产爬虫关键词分类器(2026-09-28 纠错版,金测集钉住语义)。可脱离 myssia 单独引用,myssia 核心也以固定版本范围依赖本包。
+
+## 特性
+
+- **零依赖**:纯 Python 标准库(3.11+),`pip install` 即跑,无任何传递依赖
+- **双信号裁决**:免费信号绝对优先于付费信号;注册引流过滤;死词/噪声词过滤
+- **数据与代码分离**:词表是包内 `data/keywords.json`,调关键词/正则不用改代码
+- **命中追溯**:每条结果都带命中明细(哪张表、哪个关键词/正则、命中区间)
+- **自定义规则 API**:`when` 表达式走白名单 AST,永不 `eval`,未知构造加载期即拒
+- **类型完备**:全量类型标注(`py.typed`),mypy strict 友好
+
+## 安装
+
+尚未上架 PyPI;在那之前从 myssia 仓库以 uv workspace 安装(`git clone https://github.com/xinzhuzi/myia && cd myssia && uv sync`),PyPI 上架后即可 `pip install myssia-classifier`。
+
+## 最小示例
+
+```python
+from myssia_classifier import classify_item
+
+result = classify_item({"title": "Claude 免费领100$"})
+print(result.category)  # "token"
+print(result.tags)      # 自定义规则命中的 tag(无规则时为 [])
+for evidence in result.matched:
+    print(evidence.describe())  # 命中追溯: 表:类型=值
+```
+
+七大类:`credit-card` / `proxy-node` / `buying-agent` / `server` / `token` / `ai-news` / `freebie`;另有第八个生产标签 `channel`(付费信号且无免费信号的导购/渠道帖)。被过滤标题的 `category` 为 `None`。
+
+## 自定义规则
+
+规则读取条目字段(字典键或对象属性),在粗分类之上打 tag:
+
+```python
+from myssia_classifier import Rule, classify_item
+
+rule = Rule(name="big-move", when="abs(change_pct) >= 3", tag="大波动")
+result = classify_item({"title": "NVDA 盘前大跌", "change_pct": -4.2}, rules=[rule])
+print(result.tags)  # ["大波动"]
+```
+
+`when` 表达式支持的语法(白名单 AST):字面量、字段名、list/tuple、四则与比较运算、`and`/`or`/`not`、白名单函数(`abs/min/max/round/len/int/float/str`)。属性访问、下标、lambda、f-string、推导式、任意调用等一律在构造期抛 `RuleSyntaxError`。运行期单条规则失败只按"未命中"处理并记 WARNING,绝不打断整批评测。
+
+## 自定义词表
+
+默认词表随包分发(`myssia_classifier/data/keywords.json`);也可以加载自己的 JSON:
+
+```python
+from myssia_classifier import load_table, classify_title
+
+table = load_table()  # 缺省即打包词表;自定义时传自己的 JSON 路径(结构非法时结构化报错)
+print(classify_title("便宜出极速服务器三台", table).category)  # "server"(打包词表实测命中)
+```
+
+词表 JSON 支持在 `keywords`(普通关键词)与 `patterns`(正则)两列增改,`case_insensitive` 按节可调;顶层未知字段与断裂引用(如 `fallback_target` 指向不存在的类目)加载即报错,不做静默忽略。
+
+## 从 YAML 加载规则
+
+`load_rules` 可直接读取 myssia 插件 YAML 的 `classify.rules` 节(需要环境里装有 PyYAML;本包不把它列为依赖):
+
+```python
+from myssia_classifier import load_rules
+
+rules = load_rules("plugin.yaml")
+```
+
+## 许可
+
+MIT

myssia_classifier-0.0.1/myssia_classifier/__init__.py

@@ -0,0 +1,139 @@
+"""myssia-classifier: standalone seven-category keyword classifier.
+
+Zero-dependency classification engine extracted from MYIA's classification
+layer. Public entry point is :func:`classify_item`: it runs the built-in
+dual-signal adjudication (seven categories + channel) over the item title
+and applies optional custom rules (safe ``when`` expressions) on top,
+returning ``category`` + ``tags`` + the hit trace for debugging.
+
+The keyword tables live in the packaged ``data/keywords.json`` so the owner
+or an AI can tune them without code changes.
+"""
+
+from __future__ import annotations
+
+import logging
+from dataclasses import dataclass
+from typing import Any, Iterable, Mapping
+
+from myssia_classifier.builtin import (
+    ALL_CATEGORIES,
+    CATEGORY_CHANNEL,
+    DEFAULT_TABLE_PATH,
+    SEVEN_CATEGORIES,
+    BuiltinResult,
+    ClassifyDataError,
+    ClassifyTable,
+    MatchEvidence,
+    category_label,
+    classify_title,
+    load_table,
+)
+from myssia_classifier.custom import (
+    ALLOWED_FUNCTIONS,
+    Rule,
+    RuleConfigError,
+    RuleEvalError,
+    RuleSyntaxError,
+    evaluate_expression,
+    load_rules,
+    rules_from_config,
+)
+
+__all__ = [
+    "ALL_CATEGORIES",
+    "ALLOWED_FUNCTIONS",
+    "CATEGORY_CHANNEL",
+    "DEFAULT_TABLE_PATH",
+    "SEVEN_CATEGORIES",
+    "BuiltinResult",
+    "ClassifyDataError",
+    "ClassifyResult",
+    "ClassifyTable",
+    "MatchEvidence",
+    "Rule",
+    "RuleConfigError",
+    "RuleEvalError",
+    "RuleSyntaxError",
+    "category_label",
+    "classify_item",
+    "classify_title",
+    "evaluate_expression",
+    "load_rules",
+    "load_table",
+    "rules_from_config",
+]
+
+__version__ = "0.0.1"
+
+logger = logging.getLogger(__name__)
+
+
+@dataclass(frozen=True)
+class ClassifyResult:
+    """Full classification outcome: built-in category + custom rule tags + trace."""
+
+    category: str | None  # category id, or None when the title is discarded
+    tags: list[str]  # tags from fired custom rules (falls back to rule name)
+    matched: tuple[MatchEvidence, ...]  # which keyword/pattern hit where
+    normalized_title: str
+
+
+def _item_title(item: Mapping[str, Any] | object) -> Any:
+    """Read the title; raise TypeError when the field is absent or non-str (fail fast)."""
+    if isinstance(item, Mapping):
+        if "title" not in item:
+            raise TypeError("分类条目缺少 title 字段")
+        return item["title"]
+    if not hasattr(item, "title"):
+        raise TypeError("分类条目缺少 title 属性")
+    return item.title
+
+
+def classify_item(
+    item: Mapping[str, Any] | object,
+    *,
+    table: ClassifyTable | None = None,
+    rules: Iterable[Rule] | None = None,
+) -> ClassifyResult:
+    """Classify one item: built-in keyword funnel, then custom rules.
+
+    The built-in stage reads only ``item.title`` (production semantics:
+    classification never looks at the URL). Custom rules may read any item
+    field, e.g. ``abs(change_pct) >= 3``.
+
+    Args:
+        item: dict-like item or object with a ``title`` attribute.
+        table: preloaded keyword table; defaults to the packaged
+            ``data/keywords.json``.
+        rules: optional custom rules (see :mod:`myssia_classifier.custom`).
+
+    Returns:
+        ClassifyResult with ``category`` (or ``None`` for dead/noise/
+        registration-lure/no-signal titles), fired ``tags`` and the hit
+        trace. A failing rule never breaks the batch (WARNING + skipped).
+
+    Raises:
+        TypeError: ``item.title`` is missing or not a string.
+        ClassifyDataError: keyword table malformed (fail fast at startup).
+    """
+    title = _item_title(item)
+    if not isinstance(title, str):
+        raise TypeError(f"分类条目缺少字符串 title 字段, got {type(title).__name__}")
+    builtin_result = classify_title(title, table)
+    tags: list[str] = []
+    for rule in rules or ():
+        if rule.evaluate(item):
+            tags.append(rule.tag or rule.name)
+    logger.debug(
+        "分类条目完成: category=%s tags=%s title=%r",
+        builtin_result.category,
+        tags,
+        builtin_result.normalized_title,
+    )
+    return ClassifyResult(
+        category=builtin_result.category,
+        tags=tags,
+        matched=builtin_result.matched,
+        normalized_title=builtin_result.normalized_title,
+    )