agentic-systems 1.0.2__tar.gz → 1.0.4__tar.gz

{agentic_systems-1.0.2/src/agentic_systems.egg-info → agentic_systems-1.0.4}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentic-systems
-Version: 1.0.2
+Version: 1.0.4
 Summary: Agentic Systems: auditable agentic systems with tools, skills, agents, systems, graphs, environments and evals across Python Direct, OpenAI Runtime, Bedrock Runtime and vLLM Runtime.
 Author-email: Jacobo Gerardo Gonzalez Leon <jacobogerardo.gonzalez@bbva.com>
 License-Expression: MIT
@@ -28,6 +28,12 @@ Requires-Dist: langgraph>=0.2.0; extra == "langgraph"
 Requires-Dist: typing-extensions>=4.10.0; extra == "langgraph"
 Provides-Extra: openai
 Requires-Dist: openai>=1.0.0; extra == "openai"
+Provides-Extra: vllm
+Requires-Dist: vllm>=0.9.0; extra == "vllm"
+Requires-Dist: openai>=1.0.0; extra == "vllm"
+Provides-Extra: vll
+Requires-Dist: vllm>=0.9.0; extra == "vll"
+Requires-Dist: openai>=1.0.0; extra == "vll"
 Provides-Extra: tutorials
 Requires-Dist: ipykernel>=6.0.0; extra == "tutorials"
 Requires-Dist: jupyterlab>=4.0.0; extra == "tutorials"
@@ -42,6 +48,7 @@ Requires-Dist: boto3>=1.34.0; extra == "all"
 Requires-Dist: botocore>=1.34.0; extra == "all"
 Requires-Dist: langgraph>=0.2.0; extra == "all"
 Requires-Dist: openai>=1.0.0; extra == "all"
+Requires-Dist: vllm>=0.9.0; extra == "all"
 Requires-Dist: ipykernel>=6.0.0; extra == "all"
 Requires-Dist: jupyterlab>=4.0.0; extra == "all"
 Requires-Dist: pytest>=8.0.0; extra == "all"
@@ -174,7 +181,18 @@ VLLM_MODEL
 VLLM_API_KEY
 ```
 
-`vllm-runtime` expects a running OpenAI-compatible vLLM server. The package does not start or install the GPU server by default.
+`vllm-runtime` expects a running OpenAI-compatible vLLM server. Base install does not install the GPU server. Use `agentic-systems[vllm]` or alias `agentic-systems[vll]` when you want the vLLM server dependency, and `agentic-systems[all]` when you want every optional dependency including vLLM.
+
+For Colab/L4 or Colab/T4, keep the vLLM server dependency explicit and aligned with the CUDA runtime. If a vLLM import fails with `libcudart.so.13`, install a CUDA 12 compatible vLLM build, restart the runtime, then install Agentic Systems with the OpenAI-compatible client:
+
+```python
+%pip uninstall -y vllm
+%pip install -U "vllm==0.10.2"
+%pip install -U "transformers>=4.55.2,<5" "tokenizers>=0.21.1,<0.22" "openai>=1.99.1" "huggingface_hub>=0.23.0"
+%pip install -U "agentic-systems[openai]"
+```
+
+The vLLM tutorial follows the Qwen recipe used in the working Colab template: download the model locally first, clean GPU/process state, launch `vllm.entrypoints.openai.api_server` against the local model path, and then point `VLLM_BASE_URL` at the OpenAI-compatible server.
 
 ### Bedrock Runtime
 
@@ -552,7 +570,7 @@ docs/ROADMAP_CHECKPOINTS.md
 Current verified status:
 
 ```text
-Version: 1.0.2
+Version: 1.0.3
 PyPI package: agentic-systems
 Tests: 304 passed, 0 skipped
 Coverage: 100.00%

{agentic_systems-1.0.2 → agentic_systems-1.0.4}/README.md

@@ -124,7 +124,18 @@ VLLM_MODEL
 VLLM_API_KEY
 ```
 
-`vllm-runtime` expects a running OpenAI-compatible vLLM server. The package does not start or install the GPU server by default.
+`vllm-runtime` expects a running OpenAI-compatible vLLM server. Base install does not install the GPU server. Use `agentic-systems[vllm]` or alias `agentic-systems[vll]` when you want the vLLM server dependency, and `agentic-systems[all]` when you want every optional dependency including vLLM.
+
+For Colab/L4 or Colab/T4, keep the vLLM server dependency explicit and aligned with the CUDA runtime. If a vLLM import fails with `libcudart.so.13`, install a CUDA 12 compatible vLLM build, restart the runtime, then install Agentic Systems with the OpenAI-compatible client:
+
+```python
+%pip uninstall -y vllm
+%pip install -U "vllm==0.10.2"
+%pip install -U "transformers>=4.55.2,<5" "tokenizers>=0.21.1,<0.22" "openai>=1.99.1" "huggingface_hub>=0.23.0"
+%pip install -U "agentic-systems[openai]"
+```
+
+The vLLM tutorial follows the Qwen recipe used in the working Colab template: download the model locally first, clean GPU/process state, launch `vllm.entrypoints.openai.api_server` against the local model path, and then point `VLLM_BASE_URL` at the OpenAI-compatible server.
 
 ### Bedrock Runtime
 
@@ -502,7 +513,7 @@ docs/ROADMAP_CHECKPOINTS.md
 Current verified status:
 
 ```text
-Version: 1.0.2
+Version: 1.0.3
 PyPI package: agentic-systems
 Tests: 304 passed, 0 skipped
 Coverage: 100.00%

{agentic_systems-1.0.2 → agentic_systems-1.0.4}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "agentic-systems"
-version = "1.0.2"
+version = "1.0.4"
 description = "Agentic Systems: auditable agentic systems with tools, skills, agents, systems, graphs, environments and evals across Python Direct, OpenAI Runtime, Bedrock Runtime and vLLM Runtime."
 readme = "README.md"
 requires-python = ">=3.10"
@@ -36,6 +36,8 @@ LinkedIn = "https://www.linkedin.com/in/jacoboggleon/"
 bedrock = ["boto3>=1.34.0", "botocore>=1.34.0"]
 langgraph = ["langgraph>=0.2.0", "typing-extensions>=4.10.0"]
 openai = ["openai>=1.0.0"]
+vllm = ["vllm>=0.9.0", "openai>=1.0.0"]
+vll = ["vllm>=0.9.0", "openai>=1.0.0"]
 tutorials = ["ipykernel>=6.0.0", "jupyterlab>=4.0.0", "pandas>=2.0.0", "rich>=13.0.0"]
 dev = ["pytest>=8.0.0", "pytest-cov>=5.0.0", "ruff>=0.6.1"]
 all = [
@@ -43,6 +45,7 @@ all = [
     "botocore>=1.34.0",
     "langgraph>=0.2.0",
     "openai>=1.0.0",
+    "vllm>=0.9.0",
     "ipykernel>=6.0.0",
     "jupyterlab>=4.0.0",
     "pytest>=8.0.0",

{agentic_systems-1.0.2 → agentic_systems-1.0.4}/src/agentic_systems/__init__.py

@@ -87,6 +87,6 @@ from .utils import (
     tool_result_summary,
 )
 
-__version__ = "1.0.2"
+__version__ = "1.0.3"
 
 __all__ = list(PUBLIC_API)

{agentic_systems-1.0.2 → agentic_systems-1.0.4}/src/agentic_systems/providers/bedrock_runtime.py

@@ -41,7 +41,7 @@ from botocore.exceptions import ClientError
 from pydantic import BaseModel, ConfigDict, Field, create_model
 
 
-__version__ = "1.0.2"
+__version__ = "1.0.3"
 
 
 class ToolEnvelope(BaseModel):

{agentic_systems-1.0.2 → agentic_systems-1.0.4/src/agentic_systems.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agentic-systems
-Version: 1.0.2
+Version: 1.0.4
 Summary: Agentic Systems: auditable agentic systems with tools, skills, agents, systems, graphs, environments and evals across Python Direct, OpenAI Runtime, Bedrock Runtime and vLLM Runtime.
 Author-email: Jacobo Gerardo Gonzalez Leon <jacobogerardo.gonzalez@bbva.com>
 License-Expression: MIT
@@ -28,6 +28,12 @@ Requires-Dist: langgraph>=0.2.0; extra == "langgraph"
 Requires-Dist: typing-extensions>=4.10.0; extra == "langgraph"
 Provides-Extra: openai
 Requires-Dist: openai>=1.0.0; extra == "openai"
+Provides-Extra: vllm
+Requires-Dist: vllm>=0.9.0; extra == "vllm"
+Requires-Dist: openai>=1.0.0; extra == "vllm"
+Provides-Extra: vll
+Requires-Dist: vllm>=0.9.0; extra == "vll"
+Requires-Dist: openai>=1.0.0; extra == "vll"
 Provides-Extra: tutorials
 Requires-Dist: ipykernel>=6.0.0; extra == "tutorials"
 Requires-Dist: jupyterlab>=4.0.0; extra == "tutorials"
@@ -42,6 +48,7 @@ Requires-Dist: boto3>=1.34.0; extra == "all"
 Requires-Dist: botocore>=1.34.0; extra == "all"
 Requires-Dist: langgraph>=0.2.0; extra == "all"
 Requires-Dist: openai>=1.0.0; extra == "all"
+Requires-Dist: vllm>=0.9.0; extra == "all"
 Requires-Dist: ipykernel>=6.0.0; extra == "all"
 Requires-Dist: jupyterlab>=4.0.0; extra == "all"
 Requires-Dist: pytest>=8.0.0; extra == "all"
@@ -174,7 +181,18 @@ VLLM_MODEL
 VLLM_API_KEY
 ```
 
-`vllm-runtime` expects a running OpenAI-compatible vLLM server. The package does not start or install the GPU server by default.
+`vllm-runtime` expects a running OpenAI-compatible vLLM server. Base install does not install the GPU server. Use `agentic-systems[vllm]` or alias `agentic-systems[vll]` when you want the vLLM server dependency, and `agentic-systems[all]` when you want every optional dependency including vLLM.
+
+For Colab/L4 or Colab/T4, keep the vLLM server dependency explicit and aligned with the CUDA runtime. If a vLLM import fails with `libcudart.so.13`, install a CUDA 12 compatible vLLM build, restart the runtime, then install Agentic Systems with the OpenAI-compatible client:
+
+```python
+%pip uninstall -y vllm
+%pip install -U "vllm==0.10.2"
+%pip install -U "transformers>=4.55.2,<5" "tokenizers>=0.21.1,<0.22" "openai>=1.99.1" "huggingface_hub>=0.23.0"
+%pip install -U "agentic-systems[openai]"
+```
+
+The vLLM tutorial follows the Qwen recipe used in the working Colab template: download the model locally first, clean GPU/process state, launch `vllm.entrypoints.openai.api_server` against the local model path, and then point `VLLM_BASE_URL` at the OpenAI-compatible server.
 
 ### Bedrock Runtime
 
@@ -552,7 +570,7 @@ docs/ROADMAP_CHECKPOINTS.md
 Current verified status:
 
 ```text
-Version: 1.0.2
+Version: 1.0.3
 PyPI package: agentic-systems
 Tests: 304 passed, 0 skipped
 Coverage: 100.00%

{agentic_systems-1.0.2 → agentic_systems-1.0.4}/src/agentic_systems.egg-info/requires.txt

@@ -8,6 +8,7 @@ boto3>=1.34.0
 botocore>=1.34.0
 langgraph>=0.2.0
 openai>=1.0.0
+vllm>=0.9.0
 ipykernel>=6.0.0
 jupyterlab>=4.0.0
 pytest>=8.0.0
@@ -35,3 +36,11 @@ ipykernel>=6.0.0
 jupyterlab>=4.0.0
 pandas>=2.0.0
 rich>=13.0.0
+
+[vll]
+vllm>=0.9.0
+openai>=1.0.0
+
+[vllm]
+vllm>=0.9.0
+openai>=1.0.0