open-swarm 0.1.1745275181__tar.gz → 0.1.1748636259__tar.gz

open_swarm-0.1.1748636259/PKG-INFO

@@ -0,0 +1,188 @@
+Metadata-Version: 2.4
+Name: open-swarm
+Version: 0.1.1748636259
+Summary: A tool for orchestrating Swarm using both CLI and API.
+Author-email: Matthew Hand <matthewhandau@gmail.com>
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: aiofiles>=24.1.0
+Requires-Dist: aiohttp>=3.11.11
+Requires-Dist: asyncio>=3.4.3
+Requires-Dist: asynctest>=0.13.0
+Requires-Dist: channels>=4.2.0
+Requires-Dist: colorama>=0.4.6
+Requires-Dist: django<5.0,>=4.2
+Requires-Dist: django-allauth>=65.3.1
+Requires-Dist: django-template-debug>=0.3.5
+Requires-Dist: djangorestframework>=3.15.2
+Requires-Dist: flask>=3.1.0
+Requires-Dist: jmespath>=1.0.1
+Requires-Dist: jsonschema-pydantic>=0.6
+Requires-Dist: mcp>=1.2.0
+Requires-Dist: openai>=1.58.1
+Requires-Dist: python-dotenv>=1.0.1
+Requires-Dist: redis>=5.2.1
+Requires-Dist: requests>=2.32.3
+Requires-Dist: whitenoise>=6.8.2
+Requires-Dist: drf-spectacular>=0.23.0
+Requires-Dist: pyinstaller>=6.12.0
+Requires-Dist: tiktoken>=0.9.0
+Requires-Dist: django-cors-headers>=4.7.0
+Requires-Dist: cachetools>=5.5.2
+Requires-Dist: openai-agents
+Provides-Extra: dev
+Requires-Dist: pytest>=8.3.4; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.25.1; extra == "dev"
+Requires-Dist: pytest-cov>=6.0.0; extra == "dev"
+Requires-Dist: pytest-django>=4.9.0; extra == "dev"
+Requires-Dist: pytest-mock>=3.14.0; extra == "dev"
+Requires-Dist: python-semantic-release>=9.20.0; extra == "dev"
+Provides-Extra: experimental
+Requires-Dist: nemoguardrails>=0.11.0; extra == "experimental"
+Dynamic: license-file
+
+Open-Swarm Update - 20250328
+
+This project is now repurposed due to OpenAI officially supporting the Swarm framework under the new name "openai-agents(-python)".
+
+Open-swarm now utilizes the openai-agents framework for enhanced capabilities, and the MCP logic has been offloaded to the openai-agents framework.
+
+Key focus areas of this open-swarm framework include:
+- **Blueprints**: A blueprint can be converted into an OpenAI-compatible REST endpoint (analogous to `/v1/chat/completions`, but with agents) and/or into CLI utilities on the shell.
+- **Config Loader**: Blueprints and configuration management form a core aspect of the project.
+
+Installation:
+-------------
+Open-swarm is available via PyPI. To install, run:
+```
+pip install open-swarm
+```
+
+Usage:
+------
+In development, after cloning the repository (`github.com/matthewhand/open-swarm`), you can run a blueprint directly with:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py
+```
+
+To run the blueprint with a specific instruction (for example, to list its tools), execute:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py --instruction "list your tools"
+```
+
+Alternatively, you can run the blueprint as an API endpoint using the swarm-api utility:
+```
+swarm-api --blueprint mcp_demo
+```
+
+In production, you can use the swarm-cli utility to manage and run blueprints. For example, to add an example blueprint:
+```
+swarm-cli add github:matthewhand/open-swarm/blueprints/mcp_demo
+```
+This command saves the blueprint to:
+```
+~/.swarm/blueprints/mcp_demo/
+```
+After adding the blueprint, you can convert it into a standalone CLI utility with:
+```
+swarm-cli install mcp_demo
+```
+
+Building a Basic Blueprint & Config File:
+------------------------------------------
+You can create your own blueprint to extend open-swarm's capabilities. Here is a walkthrough:
+
+1. **Create a Blueprint File:**
+   - In the `blueprints/` directory, create a new Python file, for example `blueprints/my_blueprint.py`.
+   - Define a new class that inherits from `BlueprintBase` and implement the required abstract methods, such as `metadata` and `create_agents()`. For instance:
+     ```
+     from swarm.extensions.blueprint.blueprint_base import BlueprintBase
+
+     class MyBlueprint(BlueprintBase):
+         @property
+         def metadata(self):
+             return {
+                 "title": "MyBlueprint",
+                 "env_vars": [],
+                 "required_mcp_servers": [],
+                 "max_context_tokens": 8000,
+                 "max_context_messages": 50
+             }
+
+         def create_agents(self):
+             # Create and return agents as a dictionary.
+             return {"MyAgent": ...}  # Implement your agent creation logic here.
+
+     if __name__ == "__main__":
+         MyBlueprint.main()
+     ```
+
+2. **Create a Configuration File:**
+   - Create a configuration file (e.g., `swarm_config.json`) at the root of the project. This file can include settings for LLM models and MCP servers. For example:
+     ```
+     {
+       "llm": {
+         "default": {
+           "provider": "openai",
+           "model": "gpt-4",
+           "api_key": "your-openai-api-key",
+           "base_url": null
+         }
+       },
+       "mcpServers": {
+         "mcp_llms_txt_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         },
+         "everything_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         }
+       }
+     }
+     ```
+
+3. **Running Your Blueprint:**
+   - To run your blueprint in development mode, use:
+     ```
+     uv run blueprints/my_blueprint.py
+     ```
+   - Ensure your configuration file is properly loaded by your blueprint (this might require modifications in your blueprint's initialization logic or passing a `--config` parameter).
+
+Installation & Deployment via swarm-cli:
+--------------------------------------------
+After creating your blueprint and config file, you can manage it with the swarm-cli utility. For example:
+- **Adding your blueprint:**
+  ```
+  swarm-cli add github:matthewhand/open-swarm/blueprints/my_blueprint
+  ```
+- **Installing as a standalone CLI utility:**
+  ```
+  swarm-cli install my_blueprint
+  ```
+
+Examples:
+---------
+**Blueprint "mcp_demo":**
+
+The blueprint located in `blueprints/mcp_demo` demonstrates a key design principle:
+- It creates a primary agent named **Sage** that leverages the MCP framework to incorporate external capabilities.
+- **Sage** uses another agent, **Explorer**, as a tool to extend its functionality.
+
+This hierarchical agent design illustrates how blueprints can compose agents that call on subagents as tools. This model serves as a prototype for creating powerful agent-driven workflows and can be deployed both as a REST endpoint and as a CLI tool.
+
+Production Environment:
+-----------------------
+After installing the package via pip, you can manage blueprints with `swarm-cli` and launch them as standalone utilities or REST services.
+
+For help with swarm-cli:
+```
+swarm-cli --help
+```
+
+For help with swarm-api:
+```
+swarm-api --help

open_swarm-0.1.1748636259/README.md

@@ -0,0 +1,144 @@
+Open-Swarm Update - 20250328
+
+This project is now repurposed due to OpenAI officially supporting the Swarm framework under the new name "openai-agents(-python)".
+
+Open-swarm now utilizes the openai-agents framework for enhanced capabilities, and the MCP logic has been offloaded to the openai-agents framework.
+
+Key focus areas of this open-swarm framework include:
+- **Blueprints**: A blueprint can be converted into an OpenAI-compatible REST endpoint (analogous to `/v1/chat/completions`, but with agents) and/or into CLI utilities on the shell.
+- **Config Loader**: Blueprints and configuration management form a core aspect of the project.
+
+Installation:
+-------------
+Open-swarm is available via PyPI. To install, run:
+```
+pip install open-swarm
+```
+
+Usage:
+------
+In development, after cloning the repository (`github.com/matthewhand/open-swarm`), you can run a blueprint directly with:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py
+```
+
+To run the blueprint with a specific instruction (for example, to list its tools), execute:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py --instruction "list your tools"
+```
+
+Alternatively, you can run the blueprint as an API endpoint using the swarm-api utility:
+```
+swarm-api --blueprint mcp_demo
+```
+
+In production, you can use the swarm-cli utility to manage and run blueprints. For example, to add an example blueprint:
+```
+swarm-cli add github:matthewhand/open-swarm/blueprints/mcp_demo
+```
+This command saves the blueprint to:
+```
+~/.swarm/blueprints/mcp_demo/
+```
+After adding the blueprint, you can convert it into a standalone CLI utility with:
+```
+swarm-cli install mcp_demo
+```
+
+Building a Basic Blueprint & Config File:
+------------------------------------------
+You can create your own blueprint to extend open-swarm's capabilities. Here is a walkthrough:
+
+1. **Create a Blueprint File:**
+   - In the `blueprints/` directory, create a new Python file, for example `blueprints/my_blueprint.py`.
+   - Define a new class that inherits from `BlueprintBase` and implement the required abstract methods, such as `metadata` and `create_agents()`. For instance:
+     ```
+     from swarm.extensions.blueprint.blueprint_base import BlueprintBase
+
+     class MyBlueprint(BlueprintBase):
+         @property
+         def metadata(self):
+             return {
+                 "title": "MyBlueprint",
+                 "env_vars": [],
+                 "required_mcp_servers": [],
+                 "max_context_tokens": 8000,
+                 "max_context_messages": 50
+             }
+
+         def create_agents(self):
+             # Create and return agents as a dictionary.
+             return {"MyAgent": ...}  # Implement your agent creation logic here.
+
+     if __name__ == "__main__":
+         MyBlueprint.main()
+     ```
+
+2. **Create a Configuration File:**
+   - Create a configuration file (e.g., `swarm_config.json`) at the root of the project. This file can include settings for LLM models and MCP servers. For example:
+     ```
+     {
+       "llm": {
+         "default": {
+           "provider": "openai",
+           "model": "gpt-4",
+           "api_key": "your-openai-api-key",
+           "base_url": null
+         }
+       },
+       "mcpServers": {
+         "mcp_llms_txt_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         },
+         "everything_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         }
+       }
+     }
+     ```
+
+3. **Running Your Blueprint:**
+   - To run your blueprint in development mode, use:
+     ```
+     uv run blueprints/my_blueprint.py
+     ```
+   - Ensure your configuration file is properly loaded by your blueprint (this might require modifications in your blueprint's initialization logic or passing a `--config` parameter).
+
+Installation & Deployment via swarm-cli:
+--------------------------------------------
+After creating your blueprint and config file, you can manage it with the swarm-cli utility. For example:
+- **Adding your blueprint:**
+  ```
+  swarm-cli add github:matthewhand/open-swarm/blueprints/my_blueprint
+  ```
+- **Installing as a standalone CLI utility:**
+  ```
+  swarm-cli install my_blueprint
+  ```
+
+Examples:
+---------
+**Blueprint "mcp_demo":**
+
+The blueprint located in `blueprints/mcp_demo` demonstrates a key design principle:
+- It creates a primary agent named **Sage** that leverages the MCP framework to incorporate external capabilities.
+- **Sage** uses another agent, **Explorer**, as a tool to extend its functionality.
+
+This hierarchical agent design illustrates how blueprints can compose agents that call on subagents as tools. This model serves as a prototype for creating powerful agent-driven workflows and can be deployed both as a REST endpoint and as a CLI tool.
+
+Production Environment:
+-----------------------
+After installing the package via pip, you can manage blueprints with `swarm-cli` and launch them as standalone utilities or REST services.
+
+For help with swarm-cli:
+```
+swarm-cli --help
+```
+
+For help with swarm-api:
+```
+swarm-api --help

open_swarm-0.1.1748636259/pyproject.toml

@@ -0,0 +1,62 @@
+[project]
+name = "open-swarm"
+version = "0.1.1748636259"
+description = "A tool for orchestrating Swarm using both CLI and API."
+readme = "README.md"
+authors = [
+   { name = "Matthew Hand", email = "matthewhandau@gmail.com" }
+]
+requires-python = ">=3.10"
+dependencies = [
+   "aiofiles>=24.1.0",
+   "aiohttp>=3.11.11",
+   "asyncio>=3.4.3",
+   "asynctest>=0.13.0",
+   "channels>=4.2.0",
+   "colorama>=0.4.6",
+   "django>=4.2,<5.0",
+   "django-allauth>=65.3.1",
+   "django-template-debug>=0.3.5",
+   "djangorestframework>=3.15.2",
+   "flask>=3.1.0",
+   "jmespath>=1.0.1",
+   "jsonschema-pydantic>=0.6",
+   "mcp>=1.2.0",
+   "openai>=1.58.1",
+   "python-dotenv>=1.0.1",
+   "redis>=5.2.1",
+   "requests>=2.32.3",
+   "whitenoise>=6.8.2",
+   "drf-spectacular>=0.23.0",
+   "pyinstaller>=6.12.0",
+   "tiktoken>=0.9.0",
+   "django-cors-headers>=4.7.0",
+   "cachetools>=5.5.2",
+   "openai-agents",
+]
+
+[project.scripts]
+swarm-cli = "swarm.extensions.launchers.swarm_cli:main"
+swarm-api = "swarm.extensions.launchers.swarm_api:main"
+
+[project.optional-dependencies]
+dev = [
+   "pytest>=8.3.4",
+   "pytest-asyncio>=0.25.1",
+   "pytest-cov>=6.0.0",
+   "pytest-django>=4.9.0",
+   "pytest-mock>=3.14.0",
+   "python-semantic-release>=9.20.0",
+]
+
+experimental = [
+   "nemoguardrails>=0.11.0",
+]
+
+[build-system]
+requires = ["setuptools", "wheel"]
+build-backend = "setuptools.build_meta"
+
+[tool.pytest.ini_options]
+DJANGO_SETTINGS_MODULE = "src.swarm.settings"
+pythonpath = ["src"]

open_swarm-0.1.1748636259/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

open_swarm-0.1.1748636259/src/open_swarm.egg-info/PKG-INFO

@@ -0,0 +1,188 @@
+Metadata-Version: 2.4
+Name: open-swarm
+Version: 0.1.1748636259
+Summary: A tool for orchestrating Swarm using both CLI and API.
+Author-email: Matthew Hand <matthewhandau@gmail.com>
+Requires-Python: >=3.10
+Description-Content-Type: text/markdown
+License-File: LICENSE
+Requires-Dist: aiofiles>=24.1.0
+Requires-Dist: aiohttp>=3.11.11
+Requires-Dist: asyncio>=3.4.3
+Requires-Dist: asynctest>=0.13.0
+Requires-Dist: channels>=4.2.0
+Requires-Dist: colorama>=0.4.6
+Requires-Dist: django<5.0,>=4.2
+Requires-Dist: django-allauth>=65.3.1
+Requires-Dist: django-template-debug>=0.3.5
+Requires-Dist: djangorestframework>=3.15.2
+Requires-Dist: flask>=3.1.0
+Requires-Dist: jmespath>=1.0.1
+Requires-Dist: jsonschema-pydantic>=0.6
+Requires-Dist: mcp>=1.2.0
+Requires-Dist: openai>=1.58.1
+Requires-Dist: python-dotenv>=1.0.1
+Requires-Dist: redis>=5.2.1
+Requires-Dist: requests>=2.32.3
+Requires-Dist: whitenoise>=6.8.2
+Requires-Dist: drf-spectacular>=0.23.0
+Requires-Dist: pyinstaller>=6.12.0
+Requires-Dist: tiktoken>=0.9.0
+Requires-Dist: django-cors-headers>=4.7.0
+Requires-Dist: cachetools>=5.5.2
+Requires-Dist: openai-agents
+Provides-Extra: dev
+Requires-Dist: pytest>=8.3.4; extra == "dev"
+Requires-Dist: pytest-asyncio>=0.25.1; extra == "dev"
+Requires-Dist: pytest-cov>=6.0.0; extra == "dev"
+Requires-Dist: pytest-django>=4.9.0; extra == "dev"
+Requires-Dist: pytest-mock>=3.14.0; extra == "dev"
+Requires-Dist: python-semantic-release>=9.20.0; extra == "dev"
+Provides-Extra: experimental
+Requires-Dist: nemoguardrails>=0.11.0; extra == "experimental"
+Dynamic: license-file
+
+Open-Swarm Update - 20250328
+
+This project is now repurposed due to OpenAI officially supporting the Swarm framework under the new name "openai-agents(-python)".
+
+Open-swarm now utilizes the openai-agents framework for enhanced capabilities, and the MCP logic has been offloaded to the openai-agents framework.
+
+Key focus areas of this open-swarm framework include:
+- **Blueprints**: A blueprint can be converted into an OpenAI-compatible REST endpoint (analogous to `/v1/chat/completions`, but with agents) and/or into CLI utilities on the shell.
+- **Config Loader**: Blueprints and configuration management form a core aspect of the project.
+
+Installation:
+-------------
+Open-swarm is available via PyPI. To install, run:
+```
+pip install open-swarm
+```
+
+Usage:
+------
+In development, after cloning the repository (`github.com/matthewhand/open-swarm`), you can run a blueprint directly with:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py
+```
+
+To run the blueprint with a specific instruction (for example, to list its tools), execute:
+```
+uv run blueprints/mcp_demo/blueprint_mcp_demo.py --instruction "list your tools"
+```
+
+Alternatively, you can run the blueprint as an API endpoint using the swarm-api utility:
+```
+swarm-api --blueprint mcp_demo
+```
+
+In production, you can use the swarm-cli utility to manage and run blueprints. For example, to add an example blueprint:
+```
+swarm-cli add github:matthewhand/open-swarm/blueprints/mcp_demo
+```
+This command saves the blueprint to:
+```
+~/.swarm/blueprints/mcp_demo/
+```
+After adding the blueprint, you can convert it into a standalone CLI utility with:
+```
+swarm-cli install mcp_demo
+```
+
+Building a Basic Blueprint & Config File:
+------------------------------------------
+You can create your own blueprint to extend open-swarm's capabilities. Here is a walkthrough:
+
+1. **Create a Blueprint File:**
+   - In the `blueprints/` directory, create a new Python file, for example `blueprints/my_blueprint.py`.
+   - Define a new class that inherits from `BlueprintBase` and implement the required abstract methods, such as `metadata` and `create_agents()`. For instance:
+     ```
+     from swarm.extensions.blueprint.blueprint_base import BlueprintBase
+
+     class MyBlueprint(BlueprintBase):
+         @property
+         def metadata(self):
+             return {
+                 "title": "MyBlueprint",
+                 "env_vars": [],
+                 "required_mcp_servers": [],
+                 "max_context_tokens": 8000,
+                 "max_context_messages": 50
+             }
+
+         def create_agents(self):
+             # Create and return agents as a dictionary.
+             return {"MyAgent": ...}  # Implement your agent creation logic here.
+
+     if __name__ == "__main__":
+         MyBlueprint.main()
+     ```
+
+2. **Create a Configuration File:**
+   - Create a configuration file (e.g., `swarm_config.json`) at the root of the project. This file can include settings for LLM models and MCP servers. For example:
+     ```
+     {
+       "llm": {
+         "default": {
+           "provider": "openai",
+           "model": "gpt-4",
+           "api_key": "your-openai-api-key",
+           "base_url": null
+         }
+       },
+       "mcpServers": {
+         "mcp_llms_txt_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         },
+         "everything_server": {
+           "command": "echo",
+           "args": [],
+           "env": {}
+         }
+       }
+     }
+     ```
+
+3. **Running Your Blueprint:**
+   - To run your blueprint in development mode, use:
+     ```
+     uv run blueprints/my_blueprint.py
+     ```
+   - Ensure your configuration file is properly loaded by your blueprint (this might require modifications in your blueprint's initialization logic or passing a `--config` parameter).
+
+Installation & Deployment via swarm-cli:
+--------------------------------------------
+After creating your blueprint and config file, you can manage it with the swarm-cli utility. For example:
+- **Adding your blueprint:**
+  ```
+  swarm-cli add github:matthewhand/open-swarm/blueprints/my_blueprint
+  ```
+- **Installing as a standalone CLI utility:**
+  ```
+  swarm-cli install my_blueprint
+  ```
+
+Examples:
+---------
+**Blueprint "mcp_demo":**
+
+The blueprint located in `blueprints/mcp_demo` demonstrates a key design principle:
+- It creates a primary agent named **Sage** that leverages the MCP framework to incorporate external capabilities.
+- **Sage** uses another agent, **Explorer**, as a tool to extend its functionality.
+
+This hierarchical agent design illustrates how blueprints can compose agents that call on subagents as tools. This model serves as a prototype for creating powerful agent-driven workflows and can be deployed both as a REST endpoint and as a CLI tool.
+
+Production Environment:
+-----------------------
+After installing the package via pip, you can manage blueprints with `swarm-cli` and launch them as standalone utilities or REST services.
+
+For help with swarm-cli:
+```
+swarm-cli --help
+```
+
+For help with swarm-api:
+```
+swarm-api --help