integrum 0.1.0__tar.gz

integrum-0.1.0/.gitignore

@@ -0,0 +1,42 @@
+# Python
+__pycache__/
+*.py[cod]
+*$py.class
+*.so
+.Python
+build/
+develop-eggs/
+dist/
+downloads/
+eggs/
+.eggs/
+lib/
+lib64/
+parts/
+sdist/
+var/
+wheels/
+*.egg-info/
+.installed.cfg
+*.egg
+
+# Virtual environments
+.venv/
+env/
+venv/
+ENV/
+
+# Pytest / Coverage
+.pytest_cache/
+.coverage
+htmlcov/
+
+# IDE / Editors
+.vscode/
+.idea/
+*.swp
+*.swo
+
+# Environment variables
+.env.local
+.env

integrum-0.1.0/PKG-INFO

@@ -0,0 +1,111 @@
+Metadata-Version: 2.5
+Name: integrum
+Version: 0.1.0
+Summary: Convert any Python module or library into an MCP server with FastMCP and stateful word-based object management.
+Author: Integrum Team
+License: MIT
+Keywords: fastmcp,llm,mcp,reflection,scikit-learn,tools
+Classifier: Development Status :: 4 - Beta
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: MIT License
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3.10
+Classifier: Programming Language :: Python :: 3.11
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Topic :: Software Development :: Libraries :: Python Modules
+Requires-Python: >=3.10
+Requires-Dist: click>=8.1.0
+Requires-Dist: docstring-parser>=0.16
+Requires-Dist: fastmcp>=4.0.0
+Requires-Dist: pydantic>=2.0.0
+Provides-Extra: test
+Requires-Dist: openai>=1.30.0; extra == 'test'
+Requires-Dist: pytest-asyncio>=0.23.0; extra == 'test'
+Requires-Dist: pytest>=8.0.0; extra == 'test'
+Requires-Dist: python-dotenv>=1.0.0; extra == 'test'
+Requires-Dist: scikit-learn>=1.4.0; extra == 'test'
+Requires-Dist: tenacity>=8.2.0; extra == 'test'
+Description-Content-Type: text/markdown
+
+<p align="center">
+  <img src="src/integrum/logo.png" alt="Integrum Logo" width="180"/>
+</p>
+
+<h1 align="center">Integrum</h1>
+
+<p align="center">
+  <em>Convert any Python module or library into a FastMCP Model Context Protocol server.</em>
+</p>
+
+**Integrum** turns any Python module, library, function, or class into a [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server powered by [FastMCP](https://github.com/jlowin/fastmcp).
+
+Integrum uses Python reflection (`inspect`), regex filtering, and an in-memory **Word-Based Object Store** (`playful_llama`, `swift_falcon`) so that object-oriented libraries (such as `scikit-learn`, `pandas`, `PyTorch`) can be used naturally by LLMs across multi-turn conversations without token bloat.
+
+## Features
+
+- **Module Reflection**: Automatically introspects functions and classes using `inspect`.
+- **Regex Filtering**: Expose only what you need with `--include` and `--exclude` regex patterns.
+- **Word-Based State Management**: Object IDs are short, memorable words (e.g. `playful_llama`) to save tokens and avoid UUID hallucination.
+- **Multi-User Isolation**: Supports FastMCP `Context` sessions to partition objects per connecting client.
+- **Docstring Instruction Injection**: Transforms Python docstrings into LLM tool documentation and automatically injects state-handling instructions.
+- **Dot-Notation Attribute Resolution**: Pass `swift_falcon.data` or `swift_falcon.target` directly to subsequent tool calls like `fit()`.
+- **CLI & Python API**: Run from the command line (`integrum expose`) or integrate directly into Python scripts.
+
+## Installation
+
+```bash
+uv pip install integrum
+# or
+pip install integrum
+```
+
+## Quick Start (CLI)
+
+Expose `scikit-learn` dataset loaders and RandomForestClassifier as an MCP server:
+
+```bash
+integrum expose sklearn.datasets.load_iris sklearn.ensemble.RandomForestClassifier
+```
+
+Preview what tools will be exposed:
+
+```bash
+integrum list sklearn.datasets.load_iris sklearn.ensemble.RandomForestClassifier
+```
+
+Filter functions and methods with regex:
+
+```bash
+integrum expose sklearn.datasets sklearn.ensemble \
+  --include "^(load_iris|RandomForestClassifier)$"
+```
+
+Run with HTTP or SSE transport:
+
+```bash
+integrum expose sklearn.datasets.load_iris --transport http --port 8000
+```
+
+## Python API
+
+```python
+from integrum import IntegrumServer
+
+server = IntegrumServer(name="ml-server")
+server.expose(["sklearn.datasets.load_iris", "sklearn.ensemble.RandomForestClassifier"])
+
+# Run over stdio (default for MCP desktop clients)
+server.run(transport="stdio")
+```
+
+## How State Management Works
+
+When a class is exposed, Integrum registers:
+1. `<ClassName>_create`: Instantiates the class and stores it under a word ID (e.g., `playful_llama`).
+2. `<ClassName>_<method>`: Accepts a virtual parameter `obj_id: str` (e.g., `obj_id="playful_llama"`), resolves the underlying instance, and executes the method.
+
+Returned objects (such as datasets or numpy arrays) are also registered in the store with word IDs, allowing subsequent tools to receive them as arguments.
+
+## Examples
+
+Check out [`examples/gemma_sklearn_session.md`](examples/gemma_sklearn_session.md) and [`examples/gemma_sklearn_session.json`](examples/gemma_sklearn_session.json) for a complete, verified multi-turn transcript of Gemma 4 26B loading the Iris dataset, performing `train_test_split`, training a `RandomForestClassifier`, and evaluating accuracy.

integrum-0.1.0/README.md

@@ -0,0 +1,82 @@
+<p align="center">
+  <img src="src/integrum/logo.png" alt="Integrum Logo" width="180"/>
+</p>
+
+<h1 align="center">Integrum</h1>
+
+<p align="center">
+  <em>Convert any Python module or library into a FastMCP Model Context Protocol server.</em>
+</p>
+
+**Integrum** turns any Python module, library, function, or class into a [Model Context Protocol (MCP)](https://modelcontextprotocol.io) server powered by [FastMCP](https://github.com/jlowin/fastmcp).
+
+Integrum uses Python reflection (`inspect`), regex filtering, and an in-memory **Word-Based Object Store** (`playful_llama`, `swift_falcon`) so that object-oriented libraries (such as `scikit-learn`, `pandas`, `PyTorch`) can be used naturally by LLMs across multi-turn conversations without token bloat.
+
+## Features
+
+- **Module Reflection**: Automatically introspects functions and classes using `inspect`.
+- **Regex Filtering**: Expose only what you need with `--include` and `--exclude` regex patterns.
+- **Word-Based State Management**: Object IDs are short, memorable words (e.g. `playful_llama`) to save tokens and avoid UUID hallucination.
+- **Multi-User Isolation**: Supports FastMCP `Context` sessions to partition objects per connecting client.
+- **Docstring Instruction Injection**: Transforms Python docstrings into LLM tool documentation and automatically injects state-handling instructions.
+- **Dot-Notation Attribute Resolution**: Pass `swift_falcon.data` or `swift_falcon.target` directly to subsequent tool calls like `fit()`.
+- **CLI & Python API**: Run from the command line (`integrum expose`) or integrate directly into Python scripts.
+
+## Installation
+
+```bash
+uv pip install integrum
+# or
+pip install integrum
+```
+
+## Quick Start (CLI)
+
+Expose `scikit-learn` dataset loaders and RandomForestClassifier as an MCP server:
+
+```bash
+integrum expose sklearn.datasets.load_iris sklearn.ensemble.RandomForestClassifier
+```
+
+Preview what tools will be exposed:
+
+```bash
+integrum list sklearn.datasets.load_iris sklearn.ensemble.RandomForestClassifier
+```
+
+Filter functions and methods with regex:
+
+```bash
+integrum expose sklearn.datasets sklearn.ensemble \
+  --include "^(load_iris|RandomForestClassifier)$"
+```
+
+Run with HTTP or SSE transport:
+
+```bash
+integrum expose sklearn.datasets.load_iris --transport http --port 8000
+```
+
+## Python API
+
+```python
+from integrum import IntegrumServer
+
+server = IntegrumServer(name="ml-server")
+server.expose(["sklearn.datasets.load_iris", "sklearn.ensemble.RandomForestClassifier"])
+
+# Run over stdio (default for MCP desktop clients)
+server.run(transport="stdio")
+```
+
+## How State Management Works
+
+When a class is exposed, Integrum registers:
+1. `<ClassName>_create`: Instantiates the class and stores it under a word ID (e.g., `playful_llama`).
+2. `<ClassName>_<method>`: Accepts a virtual parameter `obj_id: str` (e.g., `obj_id="playful_llama"`), resolves the underlying instance, and executes the method.
+
+Returned objects (such as datasets or numpy arrays) are also registered in the store with word IDs, allowing subsequent tools to receive them as arguments.
+
+## Examples
+
+Check out [`examples/gemma_sklearn_session.md`](examples/gemma_sklearn_session.md) and [`examples/gemma_sklearn_session.json`](examples/gemma_sklearn_session.json) for a complete, verified multi-turn transcript of Gemma 4 26B loading the Iris dataset, performing `train_test_split`, training a `RandomForestClassifier`, and evaluating accuracy.

integrum-0.1.0/examples/gemma_sklearn_session.json

@@ -0,0 +1,109 @@
+[
+  {
+    "role": "system",
+    "content": "You are an expert AI data scientist using tools from an Integrum MCP server.\nKey Rules:\n- When a tool returns an object, it assigns an object ID (e.g. 'lucid_yak').\n- Use 'obj_id' to call methods on that object (e.g. RandomForestClassifier_fit, RandomForestClassifier_score).\n- To split data with train_test_split, pass arrays as a list, e.g. arrays=['<id>.data', '<id>.target'] (or arrays=['<id>_0', '<id>_1']), along with test_size (e.g. 0.2) and random_state.\n- train_test_split returns 4 arrays: X_train, X_test, y_train, y_test (accessible via '<split_id>.X_train', '<split_id>.X_test', etc.).\n- Fit the classifier on the training set (X_train, y_train).\n- Evaluate test accuracy using RandomForestClassifier_score (passing obj_id, X='<split_id>.X_test', y='<split_id>.y_test') or accuracy_score.\n- Always report the final accuracy score in your final reply."
+  },
+  {
+    "role": "user",
+    "content": "Please perform the following end-to-end task using the tools:\n1. Load the iris dataset using load_iris.\n2. Split the dataset into train and test sets using train_test_split (test_size=0.2, random_state=42).\n3. Create a RandomForestClassifier instance.\n4. Train / fit the classifier on the training set (X_train and y_train).\n5. Evaluate accuracy on the test set (X_test and y_test) and report the final test accuracy."
+  },
+  {
+    "role": "assistant",
+    "content": null,
+    "tool_calls": [
+      {
+        "id": "call_load_iris_001",
+        "type": "function",
+        "function": {
+          "name": "load_iris",
+          "arguments": "{\"return_X_y\": true}"
+        }
+      }
+    ]
+  },
+  {
+    "role": "tool",
+    "tool_call_id": "call_load_iris_001",
+    "content": "{\"status\":\"success\",\"object_id\":\"warm_hawk\",\"type\":\"builtins.tuple\",\"shape\":\"len=2\",\"items\":[\"warm_hawk_0\",\"warm_hawk_1\"],\"message\":\"Stored tuple with object_id 'warm_hawk'. Access items using: ['warm_hawk_0', 'warm_hawk_1'].\"}"
+  },
+  {
+    "role": "assistant",
+    "content": null,
+    "tool_calls": [
+      {
+        "id": "call_train_test_split_002",
+        "type": "function",
+        "function": {
+          "name": "train_test_split",
+          "arguments": "{\"arrays\": [\"warm_hawk_0\", \"warm_hawk_1\"], \"random_state\": 42, \"test_size\": 0.2}"
+        }
+      }
+    ]
+  },
+  {
+    "role": "tool",
+    "tool_call_id": "call_train_test_split_002",
+    "content": "{\"status\":\"success\",\"object_id\":\"serene_panda\",\"type\":\"builtins.list\",\"shape\":\"len=4\",\"items\":[\"serene_panda_0\",\"serene_panda_1\",\"serene_panda_2\",\"serene_panda_3\"],\"sub_keys\":[\"serene_panda.X_train\",\"serene_panda.X_test\",\"serene_panda.y_train\",\"serene_panda.y_test\"],\"message\":\"train_test_split returned 4 arrays stored under 'serene_panda':\\n  - X_train: 'serene_panda.X_train' (or 'serene_panda_0')\\n  - X_test:  'serene_panda.X_test' (or 'serene_panda_1')\\n  - y_train: 'serene_panda.y_train' (or 'serene_panda_2')\\n  - y_test:  'serene_panda.y_test' (or 'serene_panda_3')\"}"
+  },
+  {
+    "role": "assistant",
+    "content": null,
+    "tool_calls": [
+      {
+        "id": "call_rf_create_003",
+        "type": "function",
+        "function": {
+          "name": "RandomForestClassifier_create",
+          "arguments": "{\"random_state\": 42}"
+        }
+      }
+    ]
+  },
+  {
+    "role": "tool",
+    "tool_call_id": "call_rf_create_003",
+    "content": "{\"status\":\"success\",\"object_id\":\"radiant_newt\",\"class\":\"RandomForestClassifier\",\"message\":\"Created RandomForestClassifier instance with object_id 'radiant_newt'. Use obj_id='radiant_newt' in subsequent method calls like RandomForestClassifier_fit.\"}"
+  },
+  {
+    "role": "assistant",
+    "content": null,
+    "tool_calls": [
+      {
+        "id": "call_rf_fit_004",
+        "type": "function",
+        "function": {
+          "name": "RandomForestClassifier_fit",
+          "arguments": "{\"obj_id\": \"radiant_newt\", \"X\": \"serene_panda.X_train\", \"y\": \"serene_panda.y_train\"}"
+        }
+      }
+    ]
+  },
+  {
+    "role": "tool",
+    "tool_call_id": "call_rf_fit_004",
+    "content": "{\"status\":\"success\",\"object_id\":\"radiant_newt\",\"message\":\"Successfully executed on object 'radiant_newt' (returned self).\"}"
+  },
+  {
+    "role": "assistant",
+    "content": null,
+    "tool_calls": [
+      {
+        "id": "call_rf_score_005",
+        "type": "function",
+        "function": {
+          "name": "RandomForestClassifier_score",
+          "arguments": "{\"obj_id\": \"radiant_newt\", \"X\": \"serene_panda.X_test\", \"y\": \"serene_panda.y_test\"}"
+        }
+      }
+    ]
+  },
+  {
+    "role": "tool",
+    "tool_call_id": "call_rf_score_005",
+    "content": "1.0"
+  },
+  {
+    "role": "assistant",
+    "content": "The end-to-end task has been completed successfully. Here is the summary of the steps performed:\n\n1.  **Dataset Loaded**: The Iris dataset was loaded.\n2.  **Data Split**: The dataset was split into training and testing sets (80% train, 20% test) using `random_state=42`.\n3.  **Model Created**: A `RandomForestClassifier` instance was initialized.\n4.  **Model Trained**: The classifier was fitted on the training data (`X_train`, `y_train`).\n5.  **Model Evaluated**: The model's accuracy was evaluated on the test set (`X_test`, `y_test`).\n\n**Final Test Accuracy: 1.0** (100%)"
+  }
+]

integrum-0.1.0/examples/gemma_sklearn_session.md

@@ -0,0 +1,207 @@
+# Example: End-to-End ML Pipeline with Gemma 4 26B & Integrum
+
+This example documents the exact multi-turn chat and tool calling session between **Gemma 4 26B** and the **Integrum FastMCP Server** exposing `scikit-learn`.
+
+The model loads the Iris dataset, performs a train-test split, instantiates a `RandomForestClassifier`, fits the classifier on `X_train` and `y_train`, evaluates accuracy on `X_test` and `y_test`, and reports the final result.
+
+---
+
+## 1. System Prompt & Setup
+
+### Exposed Tools
+- `load_iris`: Loads Iris dataset.
+- `train_test_split`: Splits arrays into `X_train`, `X_test`, `y_train`, `y_test`.
+- `RandomForestClassifier_create`: Instantiates the classifier and returns an object ID.
+- `RandomForestClassifier_fit`: Trains the classifier instance using virtual parameter `obj_id`.
+- `RandomForestClassifier_score`: Evaluates test set accuracy using virtual parameter `obj_id`.
+
+### System Message
+```text
+You are an expert AI data scientist using tools from an Integrum MCP server.
+Key Rules:
+- When a tool returns an object, it assigns an object ID (e.g. 'lucid_yak').
+- Use 'obj_id' to call methods on that object (e.g. RandomForestClassifier_fit, RandomForestClassifier_score).
+- To split data with train_test_split, pass arrays as a list, e.g. arrays=['<id>.data', '<id>.target'] (or arrays=['<id>_0', '<id>_1']), along with test_size (e.g. 0.2) and random_state.
+- train_test_split returns 4 arrays: X_train, X_test, y_train, y_test (accessible via '<split_id>.X_train', '<split_id>.X_test', etc.).
+- Fit the classifier on the training set (X_train, y_train).
+- Evaluate test accuracy using RandomForestClassifier_score (passing obj_id, X='<split_id>.X_test', y='<split_id>.y_test') or accuracy_score.
+- Always report the final accuracy score in your final reply.
+```
+
+### User Request
+```text
+Please perform the following end-to-end task using the tools:
+1. Load the iris dataset using load_iris.
+2. Split the dataset into train and test sets using train_test_split (test_size=0.2, random_state=42).
+3. Create a RandomForestClassifier instance.
+4. Train / fit the classifier on the training set (X_train and y_train).
+5. Evaluate accuracy on the test set (X_test and y_test) and report the final test accuracy.
+```
+
+---
+
+## 2. Multi-Turn Execution Transcript
+
+### Turn 1: Load Dataset
+
+**Gemma Tool Call:**
+```json
+{
+  "name": "load_iris",
+  "arguments": {
+    "return_X_y": true
+  }
+}
+```
+
+**Integrum FastMCP Output:**
+```json
+{
+  "status": "success",
+  "object_id": "warm_hawk",
+  "type": "builtins.tuple",
+  "shape": "len=2",
+  "items": [
+    "warm_hawk_0",
+    "warm_hawk_1"
+  ],
+  "message": "Stored tuple with object_id 'warm_hawk'. Access items using: ['warm_hawk_0', 'warm_hawk_1']."
+}
+```
+
+---
+
+### Turn 2: Train/Test Split
+
+**Gemma Tool Call:**
+```json
+{
+  "name": "train_test_split",
+  "arguments": {
+    "arrays": [
+      "warm_hawk_0",
+      "warm_hawk_1"
+    ],
+    "random_state": 42,
+    "test_size": 0.2
+  }
+}
+```
+
+**Integrum FastMCP Output:**
+```json
+{
+  "status": "success",
+  "object_id": "serene_panda",
+  "type": "builtins.list",
+  "shape": "len=4",
+  "items": [
+    "serene_panda_0",
+    "serene_panda_1",
+    "serene_panda_2",
+    "serene_panda_3"
+  ],
+  "sub_keys": [
+    "serene_panda.X_train",
+    "serene_panda.X_test",
+    "serene_panda.y_train",
+    "serene_panda.y_test"
+  ],
+  "message": "train_test_split returned 4 arrays stored under 'serene_panda':\n  - X_train: 'serene_panda.X_train' (or 'serene_panda_0')\n  - X_test:  'serene_panda.X_test' (or 'serene_panda_1')\n  - y_train: 'serene_panda.y_train' (or 'serene_panda_2')\n  - y_test:  'serene_panda.y_test' (or 'serene_panda_3')"
+}
+```
+
+---
+
+### Turn 3: Instantiate RandomForestClassifier
+
+**Gemma Tool Call:**
+```json
+{
+  "name": "RandomForestClassifier_create",
+  "arguments": {
+    "random_state": 42
+  }
+}
+```
+
+**Integrum FastMCP Output:**
+```json
+{
+  "status": "success",
+  "object_id": "radiant_newt",
+  "class": "RandomForestClassifier",
+  "message": "Created RandomForestClassifier instance with object_id 'radiant_newt'. Use obj_id='radiant_newt' in subsequent method calls like RandomForestClassifier_fit."
+}
+```
+
+---
+
+### Turn 4: Fit Model on Training Data
+
+**Gemma Tool Call:**
+```json
+{
+  "name": "RandomForestClassifier_fit",
+  "arguments": {
+    "obj_id": "radiant_newt",
+    "X": "serene_panda.X_train",
+    "y": "serene_panda.y_train"
+  }
+}
+```
+
+**Integrum FastMCP Output:**
+```json
+{
+  "status": "success",
+  "object_id": "radiant_newt",
+  "message": "Successfully executed on object 'radiant_newt' (returned self)."
+}
+```
+
+---
+
+### Turn 5: Evaluate Model on Test Data
+
+**Gemma Tool Call:**
+```json
+{
+  "name": "RandomForestClassifier_score",
+  "arguments": {
+    "obj_id": "radiant_newt",
+    "X": "serene_panda.X_test",
+    "y": "serene_panda.y_test"
+  }
+}
+```
+
+**Integrum FastMCP Output:**
+```text
+1.0
+```
+
+---
+
+### Turn 6: Assistant Final Response
+
+**Gemma Response:**
+```markdown
+The end-to-end task has been completed successfully. Here is the summary of the steps performed:
+
+1. **Dataset Loaded**: The Iris dataset was loaded.
+2. **Data Split**: The dataset was split into training and testing sets (80% train, 20% test) using `random_state=42`.
+3. **Model Created**: A `RandomForestClassifier` instance was initialized.
+4. **Model Trained**: The classifier was fitted on the training data (`X_train`, `y_train`).
+5. **Model Evaluated**: The model's accuracy was evaluated on the test set (`X_test`, `y_test`).
+
+**Final Test Accuracy: 1.0** (100%)
+```
+
+---
+
+## 3. Verification Assertion
+
+The test asserted:
+- **Object Store Verification**: Model `radiant_newt` in `ObjectStore` has 100 decision tree estimators and 3 target classes.
+- **Independent Test Set Score**: Re-running `score(X_test, y_test)` on the store instance yielded `1.0000` (100% test accuracy).

integrum-0.1.0/pyproject.toml

@@ -0,0 +1,51 @@
+[build-system]
+requires = ["hatchling"]
+build-backend = "hatchling.build"
+
+[project]
+name = "integrum"
+version = "0.1.0"
+description = "Convert any Python module or library into an MCP server with FastMCP and stateful word-based object management."
+readme = "README.md"
+requires-python = ">=3.10"
+license = { text = "MIT" }
+authors = [
+    { name = "Integrum Team" },
+]
+keywords = ["mcp", "fastmcp", "llm", "tools", "reflection", "scikit-learn"]
+classifiers = [
+    "Development Status :: 4 - Beta",
+    "Intended Audience :: Developers",
+    "License :: OSI Approved :: MIT License",
+    "Programming Language :: Python :: 3",
+    "Programming Language :: Python :: 3.10",
+    "Programming Language :: Python :: 3.11",
+    "Programming Language :: Python :: 3.12",
+    "Topic :: Software Development :: Libraries :: Python Modules",
+]
+dependencies = [
+    "fastmcp>=4.0.0",
+    "docstring-parser>=0.16",
+    "click>=8.1.0",
+    "pydantic>=2.0.0",
+]
+
+[project.optional-dependencies]
+test = [
+    "pytest>=8.0.0",
+    "pytest-asyncio>=0.23.0",
+    "scikit-learn>=1.4.0",
+    "openai>=1.30.0",
+    "python-dotenv>=1.0.0",
+    "tenacity>=8.2.0",
+]
+
+[project.scripts]
+integrum = "integrum.cli:main"
+
+[tool.hatch.build.targets.wheel]
+packages = ["src/integrum"]
+
+[tool.pytest.ini_options]
+asyncio_mode = "auto"
+testpaths = ["tests"]

integrum-0.1.0/src/integrum/__init__.py

@@ -0,0 +1,19 @@
+"""Integrum: Expose any Python module or library as an MCP server."""
+
+from integrum.filter import MemberFilter
+from integrum.inspector import import_target, inspect_class, inspect_function, inspect_module
+from integrum.server import IntegrumServer
+from integrum.store import ObjectStore, WordIdGenerator
+
+__version__ = "0.1.0"
+
+__all__ = [
+    "IntegrumServer",
+    "ObjectStore",
+    "WordIdGenerator",
+    "MemberFilter",
+    "import_target",
+    "inspect_class",
+    "inspect_function",
+    "inspect_module",
+]