aegis-gateway-pack-classification 2.0.0a0__tar.gz

aegis_gateway_pack_classification-2.0.0a0/.gitignore

@@ -0,0 +1,100 @@
+# Node/npm local installs
+node_modules/
+npm-debug.log*
+yarn-debug.log*
+yarn-error.log*
+.pnpm-store/
+
+# Local environment files
+.env
+.env.*
+!.env.example
+
+## jj
+.jj
+
+# Codex local configuration
+.codex/
+AGENTS.md
+
+# Claude Flow runtime data
+.claude-flow/data/
+.claude-flow/logs/
+CLAUDE.md
+.claude
+.claude-flow
+
+# MCP
+.mcp.json
+ruvector.db
+
+# ruflo agents
+.agents
+.swarm
+.agentdb
+agentdb.db
+
+# Environment variables
+.env
+.env.local
+.env.*.local
+
+# Created by https://www.toptal.com/developers/gitignore/api/visualstudiocode
+# Edit at https://www.toptal.com/developers/gitignore?templates=visualstudiocode
+
+### VisualStudioCode ###
+.vscode/*
+!.vscode/settings.json
+!.vscode/tasks.json
+!.vscode/launch.json
+!.vscode/extensions.json
+!.vscode/*.code-snippets
+.vscode
+
+# Local History for Visual Studio Code
+.history/
+
+# Built Visual Studio Code Extensions
+*.vsix
+
+### VisualStudioCode Patch ###
+# Ignore all local history of files
+.history
+.ionide
+
+### Misc files
+gemma4
+.postman/
+.openclaude/
+postman/
+
+### OpenClaude
+.openclaude-profile.json
+
+# Python
+__pycache__/
+*.py[cod]
+*.egg-info/
+.pytest_cache/
+dist/
+build/
+*.egg
+.venv/
+.uv-cache/
+.tmp/
+.ruff_cache/
+
+# Docker
+.docker/
+
+### Build plans
+BUILD_PLAN.md
+CLAUDE.md
+PROJECT_SPEC.md
+AMEND_PLAN.md
+SOURCE_OF_TRUTH.md
+repo-seed/
+.import_linter_cache/
+site/
+
+# End of https://www.toptal.com/developers/gitignore/api/visualstudiocode

aegis_gateway_pack_classification-2.0.0a0/PKG-INFO

@@ -0,0 +1,10 @@
+Metadata-Version: 2.4
+Name: aegis-gateway-pack-classification
+Version: 2.0.0a0
+Summary: Aegis v2 — content classification policy pack (regex labeler).
+Project-URL: Homepage, https://github.com/e-choness/aegis
+Project-URL: Repository, https://github.com/e-choness/aegis
+Project-URL: Issues, https://github.com/e-choness/aegis/issues
+License: MIT
+Requires-Python: >=3.12
+Requires-Dist: aegis-gateway-core

aegis_gateway_pack_classification-2.0.0a0/pyproject.toml

@@ -0,0 +1,26 @@
+[project]
+name = "aegis-gateway-pack-classification"
+version = "2.0.0a0"
+description = "Aegis v2 — content classification policy pack (regex labeler)."
+license = { text = "MIT" }
+requires-python = ">=3.12"
+
+dependencies = ["aegis-gateway-core"]
+
+[project.urls]
+Homepage = "https://github.com/e-choness/aegis"
+Repository = "https://github.com/e-choness/aegis"
+Issues = "https://github.com/e-choness/aegis/issues"
+
+[tool.uv.sources]
+aegis-gateway-core = { workspace = true }
+
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/aegis_pack_classification"]
+
+[project.entry-points."aegis.nodes"]
+classification = "aegis_pack_classification:ClassificationNode"

aegis_gateway_pack_classification-2.0.0a0/src/aegis_pack_classification/__init__.py

@@ -0,0 +1,5 @@
+"""aegis-pack-classification — regex content labeler."""
+
+from aegis_pack_classification.node import ClassificationNode
+
+__all__ = ["ClassificationNode"]

aegis_gateway_pack_classification-2.0.0a0/src/aegis_pack_classification/node.py

@@ -0,0 +1,67 @@
+"""ClassificationNode — regex-based content labeler writing labels.classification."""
+
+from __future__ import annotations
+
+import re
+from collections.abc import Sequence
+
+from aegis_core.pipeline.state import RunState, RunStateDelta
+
+_DEFAULT_PATTERNS: list[tuple[str, str]] = [
+    # (label, pattern)
+    ("pii", r"\b[\w.+-]+@[\w-]+\.[a-z]{2,}\b"),          # email
+    ("pii", r"\b\d{3}[-.\s]\d{3}[-.\s]\d{4}\b"),          # US phone
+    ("financial", r"\b\d{4}[\s-]\d{4}[\s-]\d{4}[\s-]\d{4}\b"),  # credit card
+    ("secret", r"(?i)\b(api[_-]?key|password|secret|token)\s*[=:]\s*\S+"),
+    ("medical", r"(?i)\b(diagnosis|prescription|patient|hipaa)\b"),
+    ("legal", r"(?i)\b(attorney.client|privileged|confidential)\b"),
+    ("public", r".*"),  # catch-all fallback
+]
+
+
+class ClassificationNode:
+    """Pipeline node that classifies the last user message via regex rules.
+
+    Writes the matched label into ``state.labels["classification"]``.
+    If no pattern matches (impossible with the default catch-all, but possible
+    with a custom ``patterns`` list), the label is left unchanged.
+
+    Args:
+        patterns: Ordered list of ``(label, regex_pattern)`` pairs.  First
+            match wins.  Defaults to a built-in set covering PII, financial,
+            secrets, medical, legal, and public content.
+        name: Node name.
+    """
+
+    def __init__(
+        self,
+        patterns: Sequence[tuple[str, str]] = _DEFAULT_PATTERNS,
+        name: str = "classification",
+    ) -> None:
+        self.name = name
+        self._rules: list[tuple[str, re.Pattern[str]]] = [
+            (label, re.compile(pattern)) for label, pattern in patterns
+        ]
+
+    async def run(self, state: RunState) -> RunStateDelta:
+        """Classify the last user message and return a labels delta."""
+        text: str | None = None
+        for msg in reversed(state.messages):
+            if msg.role == "user":
+                text = msg.content
+                break
+
+        if text is None:
+            return RunStateDelta()
+
+        label = self._classify(text)
+        if label is None:
+            return RunStateDelta()
+
+        return RunStateDelta(labels={"classification": label})
+
+    def _classify(self, text: str) -> str | None:
+        for label, pattern in self._rules:
+            if pattern.search(text):
+                return label
+        return None

aegis_gateway_pack_classification-2.0.0a0/tests/test_classification.py

@@ -0,0 +1,125 @@
+"""Tests for aegis-pack-classification.
+
+Gate: DC uv run pytest packages/aegis-pack-classification -q
+"""
+
+from __future__ import annotations
+
+from aegis_pack_classification import ClassificationNode
+
+from aegis_core.pipeline.state import RunState
+from aegis_core.providers.models import Message
+
+# ---------------------------------------------------------------------------
+# Helpers
+# ---------------------------------------------------------------------------
+
+
+def _state(*contents: str, role: str = "user") -> RunState:
+    return RunState(
+        run_id="test",
+        route="default",
+        messages=[Message(role=role, content=c) for c in contents],
+    )
+
+
+# ---------------------------------------------------------------------------
+# ClassificationNode — basic labeling
+# ---------------------------------------------------------------------------
+
+
+class TestClassificationNode:
+    async def test_classifies_email_as_pii(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("My email is alice@example.com"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "pii"
+
+    async def test_classifies_phone_as_pii(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("Call me at 555-867-5309"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "pii"
+
+    async def test_classifies_credit_card_as_financial(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("Card: 4111 1111 1111 1111"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "financial"
+
+    async def test_classifies_api_key_as_secret(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("api_key=sk-abc123"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "secret"
+
+    async def test_classifies_password_as_secret(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("password: hunter2"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "secret"
+
+    async def test_classifies_medical(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("Patient diagnosis: hypertension"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "medical"
+
+    async def test_classifies_legal(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("This is attorney-client privileged"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "legal"
+
+    async def test_classifies_plain_text_as_public(self) -> None:
+        node = ClassificationNode()
+        delta = await node.run(_state("What is the weather today?"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "public"
+
+    async def test_uses_last_user_message(self) -> None:
+        node = ClassificationNode()
+        state = RunState(
+            run_id="r",
+            route="d",
+            messages=[
+                Message(role="user", content="What is the weather?"),
+                Message(role="assistant", content="It is sunny."),
+                Message(role="user", content="My email is bob@test.com"),
+            ],
+        )
+        delta = await node.run(state)
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "pii"
+
+    async def test_no_user_message_returns_empty_delta(self) -> None:
+        node = ClassificationNode()
+        state = RunState(
+            run_id="r",
+            route="d",
+            messages=[Message(role="assistant", content="hello")],
+        )
+        delta = await node.run(state)
+        assert delta.labels is None
+
+    async def test_custom_patterns_first_match_wins(self) -> None:
+        node = ClassificationNode(patterns=[
+            ("top_secret", r"classified"),
+            ("public", r".*"),
+        ])
+        delta = await node.run(_state("This document is classified"))
+        assert delta.labels is not None
+        assert delta.labels["classification"] == "top_secret"
+
+    async def test_no_match_with_empty_patterns(self) -> None:
+        node = ClassificationNode(patterns=[])
+        delta = await node.run(_state("anything"))
+        assert delta.labels is None
+
+    def test_node_name_default(self) -> None:
+        node = ClassificationNode()
+        assert node.name == "classification"
+
+    def test_node_name_custom(self) -> None:
+        node = ClassificationNode(name="my_classifier")
+        assert node.name == "my_classifier"