mockstack 0.7.3__tar.gz → 0.8.0__tar.gz

{mockstack-0.7.3 → mockstack-0.8.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mockstack
-Version: 0.7.3
+Version: 0.8.0
 Summary: An API mocking workhorse
 Author-email: Adam Ever-Hadani <mockstack.contact@gmail.com>
 License-Expression: MIT
@@ -27,6 +27,8 @@ Requires-Dist: pyyaml>=6.0.2
 Requires-Dist: rich>=14.0.0
 Requires-Dist: typer>=0.15.2
 Requires-Dist: uvicorn>=0.34.2
+Provides-Extra: llm
+Requires-Dist: ollama>=0.4.8; extra == "llm"
 Dynamic: license-file
 
 ![mockstack logo](https://github.com/adamhadani/mockstack/raw/main/docs/assets/mockstack.png)
@@ -50,7 +52,7 @@ Use mockstack for:
 
 * **Integration Testing** :ok_hand:. Creating a consistent environment for running integration tests on a single component, using fixture responses.
 
-* **LLM-powered Workflows** :alien:. Speeding up development of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs, you often want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace. **mockstack** can give you a solid foundation for this.
+* **LLM-powered Workflows** :alien:. Speeding up development and reducing per-token costs of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs. Optional [Ollama](https://ollama.com/) integration allows for realistic mocking of 3rd party LLMs without changing a line of code in your project. In addition, when you want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace you can create various fixture responses with varying levels of dynamic content that is template-driven. **mockstack** can give you a solid foundation for this.
 
 * **Chaos Engineering** :boom:. mockstack can simulate various real-world runtime error scenarios such as timeouts, http error codes, and invalid response payloads. This can be a great way to do some upfront [Chaos Monkey](https://github.com/Netflix/chaosmonkey) type of testing on software components.
 

{mockstack-0.7.3 → mockstack-0.8.0}/README.md

@@ -19,7 +19,7 @@ Use mockstack for:
 
 * **Integration Testing** :ok_hand:. Creating a consistent environment for running integration tests on a single component, using fixture responses.
 
-* **LLM-powered Workflows** :alien:. Speeding up development of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs, you often want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace. **mockstack** can give you a solid foundation for this.
+* **LLM-powered Workflows** :alien:. Speeding up development and reducing per-token costs of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs. Optional [Ollama](https://ollama.com/) integration allows for realistic mocking of 3rd party LLMs without changing a line of code in your project. In addition, when you want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace you can create various fixture responses with varying levels of dynamic content that is template-driven. **mockstack** can give you a solid foundation for this.
 
 * **Chaos Engineering** :boom:. mockstack can simulate various real-world runtime error scenarios such as timeouts, http error codes, and invalid response payloads. This can be a great way to do some upfront [Chaos Monkey](https://github.com/Netflix/chaosmonkey) type of testing on software components.
 

mockstack-0.8.0/docs/ollama.md

@@ -0,0 +1,86 @@
+# Ollama Integration
+
+Mockstack provides integration with [Ollama](https://ollama.com/), allowing you to use real LLM responses in your mock templates. This is particularly useful for development, debugging, and integration testing scenarios where you want to capture the non-deterministic nature of LLM responses.
+
+## Prerequisites
+
+To use the Ollama integration, you'll need:
+
+1. Mockstack installed with the optional `llm` dependencies:
+   ```bash
+   uv pip install mockstack[llm]
+   ```
+
+2. [Ollama](https://ollama.com/) installed locally with at least one model (e.g., "llama3.2")
+
+## Basic Usage
+
+The Ollama integration works by routing requests to a template file that uses the special `ollama` template function.
+
+1. Configure your LLM client to hit an endpoint that maps to a template filepath calling the `ollama` method, e.g.:
+   ```python
+   from langchain_openai import ChatOpenAI
+
+   llm = ChatOpenAI(
+       model="gpt-4o",
+       base_url="http://localhost:8000/ollama/openai/v1",
+       api_key="SOME_STRING_THAT_DOES_NOT_MATTER",
+   )
+   ```
+
+2. Make requests as you normally would:
+   ```python
+   messages = [
+       (
+           "system",
+           "You are a helpful assistant that translates English to French. Translate the user sentence.",
+       ),
+       ("human", "mockstack is pretty cool. But LLMs are way cooler"),
+   ]
+   ai_msg = llm.invoke(messages)
+   print(ai_msg.content)
+   ```
+
+## Integration with Templates
+
+You can also use Ollama responses within your Jinja templates. This allows you to:
+
+1. Mix static and dynamic content
+2. Apply transformations to the LLM responses
+3. Create conditional logic based on the responses
+
+Example template structure:
+```jinja
+{
+  "id": "chatcmpl-123",
+  "object": "chat.completion",
+  "created": 1677652288,
+  "model": "{{ model }}",
+  "choices": [{
+    "index": 0,
+    "message": {
+      "role": "assistant",
+      "content": "{{ ollama(request_json.messages, 'llama3.2') | json_escape }}"
+    },
+    "finish_reason": "stop"
+  }],
+  "usage": {
+    "prompt_tokens": 9,
+    "completion_tokens": 12,
+    "total_tokens": 21
+  }
+}
+```
+
+## Best Practices
+
+1. **Caching**: Consider implementing caching for frequently used responses to improve performance
+2. **Model Selection**: Choose the appropriate model based on your testing needs
+3. **Error Handling**: Implement proper error handling in your templates
+4. **Performance**: Be mindful of response times when using real LLM responses
+
+## Limitations
+
+- The integration requires a local Ollama instance
+- Response times will be slower than static templates
+- Model availability depends on your local setup

{mockstack-0.7.3 → mockstack-0.8.0}/examples/llm/mockstack-langchain-example.ipynb

@@ -11,7 +11,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 26,
+   "execution_count": 7,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -22,11 +22,31 @@
     "#!uv pip install langchain langchain-openai"
    ]
   },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "from ollama import chat\n",
+    "from ollama import ChatResponse\n",
+    "\n",
+    "response: ChatResponse = chat(model='llama3.2', messages=[\n",
+    "  {\n",
+    "    'role': 'user',\n",
+    "    'content': 'Why is the sky blue?',\n",
+    "  },\n",
+    "])\n",
+    "print(response['message']['content'])\n",
+    "# or access fields directly from the response object\n",
+    "print(response.message.content)"
+   ]
+  },
   {
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Example #1: Template-based mocking (`filefixtures` strategy)\n",
+    "## Example #1: Static template-based mocking (`filefixtures` strategy)\n",
     "\n",
     "Here we simply use the **filefixtures** strategy to route requesets coming in for a certain URL to a template file with the appropriate name.\n",
     "\n",
@@ -42,7 +62,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 38,
+   "execution_count": 2,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -86,7 +106,7 @@
   },
   {
    "cell_type": "code",
-   "execution_count": 40,
+   "execution_count": 3,
    "metadata": {},
    "outputs": [],
    "source": [
@@ -120,17 +140,62 @@
    "cell_type": "markdown",
    "metadata": {},
    "source": [
-    "## Example #2: `llm` strategy\n",
+    "## Example #2: Dynamic template-based mocking with ollama integration\n",
     "\n",
-    "** COMING SOON **"
+    "In this example we use the [ollama-python](https://github.com/ollama/ollama-python) integration to return mock responses that actually come from a real LLM running on your host! This is a great way to go a step further in development, debugging, and integration testing scenarios where LLMs and their non-determinism are important to capture.\n",
+    "\n",
+    "For this example you will need to make sure you have the following:\n",
+    "\n",
+    "* mockstack installed with the optional `llm` dependencies:\n",
+    "\n",
+    "    ```bash\n",
+    "    uv pip install mockstack[llm]\n",
+    "    ```\n",
+    "\n",
+    "* [ollama](https://ollama.com/) installed locally with the \"llama3.2\" model (which is typically the default model installed)"
    ]
   },
   {
    "cell_type": "code",
    "execution_count": null,
    "metadata": {},
-   "outputs": [],
-   "source": []
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "Le mockstack est très amusant, mais les LLMs sont beaucoup plus froids.\n",
+      "\n",
+      "(Note: I used \"froids\" instead of \"cool\" as it's a more common translation in French. If you want to keep the exact word order and meaning, I can suggest an alternative: Le mockstack est très amusant, mais les LLMs sont vraiment cool.)\n"
+     ]
+    }
+   ],
+   "source": [
+    "from langchain_openai import ChatOpenAI\n",
+    "\n",
+    "\n",
+    "llm = ChatOpenAI(\n",
+    "    model=\"gpt-4o\",\n",
+    "    temperature=0,\n",
+    "    max_tokens=None,\n",
+    "    timeout=None,\n",
+    "    max_retries=2,\n",
+    "    base_url=\"http://localhost:8000/ollama/openai/v1\",\n",
+    "    api_key=\"SOME_STRING_THAT_DOES_NOT_MATTER\",\n",
+    ")\n",
+    "\n",
+    "messages = [\n",
+    "    (\n",
+    "        \"system\",\n",
+    "        \"You are a helpful assistant that translates English to French. Translate the user sentence.\",\n",
+    "    ),\n",
+    "    (\"human\", \"mockstack is pretty cool. But LLMs are way cooler\"),\n",
+    "]\n",
+    "ai_msg = llm.invoke(messages)\n",
+    "\n",
+    "# Output would not be deterministic, but should be a valid French translation.\n",
+    "print(ai_msg.content)"
+   ]
   },
   {
    "cell_type": "markdown",

mockstack-0.8.0/examples/llm/templates/ollama-openai-v1-chat-completions.j2

@@ -0,0 +1,35 @@
+ {
+   "id": "chatcmpl-B9MBs8CjcvOU2jLn4n570S5qMJKcT",
+   "object": "chat.completion",
+   "created": 1741569952,
+   "model": "gpt-4.1-2025-04-14",
+   "choices": [
+     {
+       "index": 0,
+       "message": {
+         "role": "assistant",
+         "content": "{{ ollama(request_json.messages, 'llama3.2') | json_escape }}",
+         "refusal": null,
+         "annotations": []
+       },
+       "logprobs": null,
+       "finish_reason": "stop"
+     }
+   ],
+   "usage": {
+     "prompt_tokens": 19,
+     "completion_tokens": 10,
+     "total_tokens": 29,
+     "prompt_tokens_details": {
+       "cached_tokens": 0,
+       "audio_tokens": 0
+     },
+     "completion_tokens_details": {
+       "reasoning_tokens": 0,
+       "audio_tokens": 0,
+       "accepted_prediction_tokens": 0,
+       "rejected_prediction_tokens": 0
+     }
+   },
+   "service_tier": "default"
+ }

{mockstack-0.7.3 → mockstack-0.8.0}/mkdocs.yml

@@ -49,6 +49,14 @@ markdown_extensions:
   - pymdownx.tabbed:
       alternate_style: true
 
+nav:
+  - Home: README.md
+  - Configuration: configuration.md
+  - Strategies:
+    - Overview: strategies/index.md
+    - FileFixtures: strategies/filefixtures.md
+  - LLM Integrations:
+    - Ollama: ollama.md
 
 extra:
   social:

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack/config.py

@@ -20,13 +20,13 @@ from mockstack.constants import (
 class OpenTelemetrySettings(BaseSettings):
     """Settings for OpenTelemetry."""
 
-    enabled: bool = False
+    enabled: CliImplicitFlag[bool] = False
 
     endpoint: str = "http://localhost:4317/"
 
     # whether to capture the response body.
     # this can be heavy, sensitive (PII) and/or not needed depending on the use case.
-    capture_response_body: bool = False
+    capture_response_body: CliImplicitFlag[bool] = False
 
 
 class Settings(BaseSettings):

mockstack-0.8.0/mockstack/exceptions.py

@@ -0,0 +1,6 @@
+"""Custom exceptions for mockstack."""
+
+
+def raise_for_missing(message: str, *args, **kwargs):
+    """Raise an exception for a missing dependency."""
+    raise RuntimeError(message)

mockstack-0.8.0/mockstack/llm/ollama.py

@@ -0,0 +1,50 @@
+"""Ollama integration"""
+
+from typing import List, Dict
+
+try:
+    from ollama import chat
+    from ollama import ChatResponse
+
+    IS_OLLAMA_AVAILABLE = True
+except ImportError:
+    IS_OLLAMA_AVAILABLE = False
+
+
+if IS_OLLAMA_AVAILABLE:
+
+    class OllamaLLM:
+        def __init__(self, model: str = "llama3.2"):
+            self.model = model
+
+        def __call__(
+            self,
+            messages: List[Dict[str, str]],
+            max_tokens: int = 4096,
+            temperature: float = 0.7,
+        ) -> str:
+            response: ChatResponse = chat(
+                model=self.model,
+                messages=messages,
+                options={"num_ctx": max_tokens, "temperature": temperature},
+            )
+
+            return content(response)
+
+    def content(response: ChatResponse) -> str:
+        """Extract the message content from the LLM response."""
+        return response["message"]["content"]
+
+    def ollama(
+        messages: List[Dict[str, str]],
+        model: str = "llama3.2",
+        *args,
+        **kwargs,
+    ) -> str:
+        """Fluent interface for Ollama to be used in templates."""
+
+        return OllamaLLM(model)(
+            messages,
+            *args,
+            **kwargs,
+        )

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack/strategies/filefixtures.py

@@ -2,21 +2,24 @@
 
 import logging
 import os
+from functools import cached_property
 from pathlib import Path
 
 from fastapi import HTTPException, Request, Response, status
-from jinja2 import Environment, FileSystemLoader
+from jinja2 import Environment
 
 from mockstack.config import Settings
 from mockstack.intent import (
     looks_like_a_command,
     looks_like_a_search,
+    wants_json,
 )
 from mockstack.strategies.base import BaseStrategy
 from mockstack.strategies.create_mixin import CreateMixin
 from mockstack.templating import (
     iter_possible_template_arguments,
     missing_template_detail,
+    templates_env_provider,
 )
 
 
@@ -37,8 +40,6 @@ class FileFixturesStrategy(BaseStrategy, CreateMixin):
         self.created_resource_metadata = settings.created_resource_metadata
         self.missing_resource_fields = settings.missing_resource_fields
 
-        self.env = Environment(loader=FileSystemLoader(self.templates_dir))
-
     def __str__(self) -> str:
         return (
             f"[medium_purple]filefixtures[/medium_purple]\n "
@@ -46,6 +47,11 @@ class FileFixturesStrategy(BaseStrategy, CreateMixin):
             f"enable_templates_for_post: [medium_purple]{self.enable_templates_for_post}[/medium_purple]. "
         )
 
+    @cached_property
+    def env(self) -> Environment:
+        """Jinja2 environment for the filefixtures strategy."""
+        return templates_env_provider(self.templates_dir)
+
     async def apply(self, request: Request) -> Response:
         match request.method:
             case "GET":
@@ -74,9 +80,10 @@ class FileFixturesStrategy(BaseStrategy, CreateMixin):
         We also allow a configuration to specify a default intent.
 
         """
+        request_json = (await request.json()) if wants_json(request) else None
         if self.enable_templates_for_post:
             try:
-                return self._response_from_template(request)
+                return self._response_from_template(request, request_json=request_json)
             except HTTPException as e:
                 if e.status_code == status.HTTP_404_NOT_FOUND:
                     # If the template is not found, we try to create the resource with logic below.
@@ -86,12 +93,12 @@ class FileFixturesStrategy(BaseStrategy, CreateMixin):
 
         if looks_like_a_search(request):
             # Searching for resources with a complex query that cannot be expressed in a URI.
-            return self._response_from_template(request)
+            return self._response_from_template(request, request_json=request_json)
         elif looks_like_a_command(request):
             # Executing a 'command' of some sort, like a workflow or a batch job.
             # We return a 201 CREATED status code with response from template.
             return self._response_from_template(
-                request, status_code=status.HTTP_201_CREATED
+                request, request_json=request_json, status_code=status.HTTP_201_CREATED
             )
         else:
             # simulate resource creation:
@@ -133,9 +140,15 @@ class FileFixturesStrategy(BaseStrategy, CreateMixin):
         return Response(status_code=status.HTTP_204_NO_CONTENT)
 
     def _response_from_template(
-        self, request: Request, status_code: int = status.HTTP_200_OK
+        self,
+        request: Request,
+        *,
+        request_json: dict | None = None,
+        status_code: int = status.HTTP_200_OK,
     ) -> Response:
-        for template_args in iter_possible_template_arguments(request):
+        for template_args in iter_possible_template_arguments(
+            request, request_json=request_json
+        ):
             filename = self.templates_dir / template_args["name"]
             self.logger.debug("Looking for template filename: %s", filename)
             if not os.path.exists(filename):

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack/strategies/proxyrules.py

@@ -18,6 +18,7 @@ from mockstack.constants import ProxyRulesRedirectVia
 from mockstack.intent import looks_like_a_create
 from mockstack.strategies.base import BaseStrategy
 from mockstack.strategies.create_mixin import CreateMixin
+from mockstack.templating import templates_env_provider
 
 
 class Rule:
@@ -73,8 +74,6 @@ class ProxyRulesStrategy(BaseStrategy, CreateMixin):
         self.simulate_create_on_missing = settings.proxyrules_simulate_create_on_missing
         self.created_resource_metadata = settings.created_resource_metadata
 
-        self.env = Environment()
-
     def __str__(self) -> str:
         return (
             f"[medium_purple]proxyrules[/medium_purple]\n "
@@ -84,6 +83,11 @@ class ProxyRulesStrategy(BaseStrategy, CreateMixin):
             f"reverse_proxy_timeout: {self.reverse_proxy_timeout}"
         )
 
+    @cached_property
+    def env(self) -> Environment:
+        """Jinja2 environment for the proxy rules strategy."""
+        return templates_env_provider()
+
     @cached_property
     def rules(self) -> list[Rule]:
         return self.load_rules()

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack/templating.py

@@ -1,14 +1,39 @@
 """Templates related functionality."""
 
 from collections import OrderedDict
+from functools import partial
 from pathlib import Path
 from typing import Generator
 
 from fastapi import Request
+from jinja2 import Environment, FileSystemLoader
 
+from mockstack.exceptions import raise_for_missing
 from mockstack.identifiers import looks_like_id, prefixes
 
 
+def templates_env_provider(templates_dir: Path | str | None = None) -> Environment:
+    """Provide a Jinja2 environment for the templates."""
+    # TODO refactor a bit to be more generic for optional dependencies.
+    from mockstack.llm import ollama
+
+    loader = FileSystemLoader(templates_dir) if templates_dir else None
+
+    env = Environment(loader=loader)
+
+    env.filters["json_escape"] = json_escape
+
+    if ollama.IS_OLLAMA_AVAILABLE:
+        env.globals["ollama"] = ollama.ollama
+    else:
+        env.globals["ollama"] = partial(
+            raise_for_missing,
+            "Ollama is not available. Install with optional dependency mockstack[llm] to use it.",
+        )
+
+    return env
+
+
 def missing_template_detail(request: Request, *, templates_dir: Path) -> str:
     """Return a detailed message for a missing template."""
     return (
@@ -21,6 +46,8 @@ def missing_template_detail(request: Request, *, templates_dir: Path) -> str:
 
 def iter_possible_template_arguments(
     request: Request,
+    *,
+    request_json: dict | None = None,
     default_identifier_key: str = "id",
     default_media_type: str = "application/json",
     default_template_name: str = "index.j2",
@@ -47,9 +74,9 @@ def iter_possible_template_arguments(
     )
     media_type = request.headers.get("Content-Type", default_media_type)
 
-    # hydrate template context with additional request details such as query params.
     context = dict(
         query=dict(request.query_params),
+        request_json=request_json,
         headers=dict(request.headers),
         **identifiers,
     )
@@ -119,3 +146,19 @@ def iter_possible_template_filenames(
         yield template_file_separator.join(name_segments) + template_file_extension
 
     yield default_template_name
+
+
+def json_escape(text: str) -> str:
+    """Escape quotes, newlines, tabs, etc. in a string.
+
+    Used as a filter in templates to make the output JSON-compliant.
+
+    """
+    return (
+        text.replace('"', '\\"')
+        .replace("\n", "\\n")
+        .replace("\r", "\\r")
+        .replace("\t", "\\t")
+        .replace("\f", "\\f")
+        .replace("\v", "\\v")
+    )

mockstack-0.8.0/mockstack/tests/llm/test_ollama.py

@@ -0,0 +1,95 @@
+"""Tests for the Ollama module."""
+
+import pytest
+from unittest.mock import patch, MagicMock
+from typing import List, Dict
+
+
+@pytest.fixture
+def mock_messages() -> List[Dict[str, str]]:
+    """Sample messages for testing."""
+    return [
+        {"role": "user", "content": "Hello"},
+        {"role": "assistant", "content": "Hi there!"},
+    ]
+
+
+@pytest.fixture
+def mock_chat_response() -> Dict:
+    """Mock chat response from Ollama."""
+    return {"message": {"content": "This is a test response"}}
+
+
+@pytest.fixture
+def mock_ollama_module():
+    """Mock the ollama module."""
+    # TODO: Should probably remove this in favor making sure ollama optional package
+    # is installed for running unit-tests, and instead mocking it as missing for the
+    # "not available" test cases.
+    with patch.dict("sys.modules", {"ollama": MagicMock()}) as mocked_dict:
+        yield mocked_dict
+
+
+class TestOllamaAvailable:
+    """Test cases when Ollama is available."""
+
+    @patch("mockstack.llm.ollama.IS_OLLAMA_AVAILABLE", True)
+    def test_ollama_llm_initialization(self, mock_ollama_module):
+        """Test OllamaLLM initialization with default model."""
+        from mockstack.llm.ollama import OllamaLLM
+
+        llm = OllamaLLM()
+        assert llm.model == "llama3.2"
+
+    @patch("mockstack.llm.ollama.IS_OLLAMA_AVAILABLE", True)
+    def test_ollama_llm_initialization_custom_model(self, mock_ollama_module):
+        """Test OllamaLLM initialization with custom model."""
+        from mockstack.llm.ollama import OllamaLLM
+
+        llm = OllamaLLM(model="custom-model")
+        assert llm.model == "custom-model"
+
+    @patch("mockstack.llm.ollama.IS_OLLAMA_AVAILABLE", True)
+    def test_ollama_llm_call(
+        self, mock_ollama_module, mock_messages, mock_chat_response
+    ):
+        """Test OllamaLLM.__call__ method."""
+        from mockstack.llm.ollama import OllamaLLM
+
+        mock_chat = MagicMock(return_value=mock_chat_response)
+        with patch("mockstack.llm.ollama.chat", mock_chat):
+            llm = OllamaLLM()
+            response = llm(mock_messages)
+
+            mock_chat.assert_called_once_with(
+                model="llama3.2",
+                messages=mock_messages,
+                options={"num_ctx": 4096, "temperature": 0.7},
+            )
+            assert response == "This is a test response"
+
+    @patch("mockstack.llm.ollama.IS_OLLAMA_AVAILABLE", True)
+    def test_content_function(self, mock_ollama_module, mock_chat_response):
+        """Test content function extraction."""
+        from mockstack.llm.ollama import content
+
+        result = content(mock_chat_response)
+        assert result == "This is a test response"
+
+    @patch("mockstack.llm.ollama.IS_OLLAMA_AVAILABLE", True)
+    def test_ollama_function(
+        self, mock_ollama_module, mock_messages, mock_chat_response
+    ):
+        """Test ollama function."""
+        from mockstack.llm.ollama import ollama
+
+        mock_chat = MagicMock(return_value=mock_chat_response)
+        with patch("mockstack.llm.ollama.chat", mock_chat):
+            response = ollama(mock_messages)
+
+            mock_chat.assert_called_once_with(
+                model="llama3.2",
+                messages=mock_messages,
+                options={"num_ctx": 4096, "temperature": 0.7},
+            )
+            assert response == "This is a test response"

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack/tests/test_templating.py

@@ -18,17 +18,32 @@ from mockstack.templating import (
             [
                 {
                     "name": "api-v1-projects.1234.j2",
-                    "context": {"projects": "1234", "query": {}, "headers": {}},
+                    "context": {
+                        "projects": "1234",
+                        "query": {},
+                        "headers": {},
+                        "request_json": None,
+                    },
                     "media_type": "application/json",
                 },
                 {
                     "name": "api-v1-projects.j2",
-                    "context": {"projects": "1234", "query": {}, "headers": {}},
+                    "context": {
+                        "projects": "1234",
+                        "query": {},
+                        "headers": {},
+                        "request_json": None,
+                    },
                     "media_type": "application/json",
                 },
                 {
                     "name": "index.j2",
-                    "context": {"projects": "1234", "query": {}, "headers": {}},
+                    "context": {
+                        "projects": "1234",
+                        "query": {},
+                        "headers": {},
+                        "request_json": None,
+                    },
                     "media_type": "application/json",
                 },
             ],
@@ -42,6 +57,7 @@ from mockstack.templating import (
                         "users": "3a4e5ad9-17ee-41af-972f-864dfccd4856",
                         "query": {},
                         "headers": {},
+                        "request_json": None,
                     },
                     "media_type": "application/json",
                 },
@@ -51,6 +67,7 @@ from mockstack.templating import (
                         "users": "3a4e5ad9-17ee-41af-972f-864dfccd4856",
                         "query": {},
                         "headers": {},
+                        "request_json": None,
                     },
                     "media_type": "application/json",
                 },
@@ -60,6 +77,7 @@ from mockstack.templating import (
                         "users": "3a4e5ad9-17ee-41af-972f-864dfccd4856",
                         "query": {},
                         "headers": {},
+                        "request_json": None,
                     },
                     "media_type": "application/json",
                 },
@@ -70,12 +88,12 @@ from mockstack.templating import (
             [
                 {
                     "name": "api-v1-projects.j2",
-                    "context": {"query": {}, "headers": {}},
+                    "context": {"query": {}, "headers": {}, "request_json": None},
                     "media_type": "application/json",
                 },
                 {
                     "name": "index.j2",
-                    "context": {"query": {}, "headers": {}},
+                    "context": {"query": {}, "headers": {}, "request_json": None},
                     "media_type": "application/json",
                 },
             ],
@@ -85,7 +103,12 @@ from mockstack.templating import (
             [
                 {
                     "name": "index.j2",
-                    "context": {"id": "1234", "query": {}, "headers": {}},
+                    "context": {
+                        "id": "1234",
+                        "query": {},
+                        "headers": {},
+                        "request_json": None,
+                    },
                     "media_type": "application/json",
                 },
             ],

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: mockstack
-Version: 0.7.3
+Version: 0.8.0
 Summary: An API mocking workhorse
 Author-email: Adam Ever-Hadani <mockstack.contact@gmail.com>
 License-Expression: MIT
@@ -27,6 +27,8 @@ Requires-Dist: pyyaml>=6.0.2
 Requires-Dist: rich>=14.0.0
 Requires-Dist: typer>=0.15.2
 Requires-Dist: uvicorn>=0.34.2
+Provides-Extra: llm
+Requires-Dist: ollama>=0.4.8; extra == "llm"
 Dynamic: license-file
 
 ![mockstack logo](https://github.com/adamhadani/mockstack/raw/main/docs/assets/mockstack.png)
@@ -50,7 +52,7 @@ Use mockstack for:
 
 * **Integration Testing** :ok_hand:. Creating a consistent environment for running integration tests on a single component, using fixture responses.
 
-* **LLM-powered Workflows** :alien:. Speeding up development of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs, you often want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace. **mockstack** can give you a solid foundation for this.
+* **LLM-powered Workflows** :alien:. Speeding up development and reducing per-token costs of LLM-based workflows and tools for use with frameworks such as [LangChain](https://python.langchain.com/docs/introduction/), [LangGraph](https://www.langchain.com/langgraph) and others. When developing LLM-driven execution graphs. Optional [Ollama](https://ollama.com/) integration allows for realistic mocking of 3rd party LLMs without changing a line of code in your project. In addition, when you want to have a consistent response from a tool while you're tuning prompts or debugging other aspects of a particular trace you can create various fixture responses with varying levels of dynamic content that is template-driven. **mockstack** can give you a solid foundation for this.
 
 * **Chaos Engineering** :boom:. mockstack can simulate various real-world runtime error scenarios such as timeouts, http error codes, and invalid response payloads. This can be a great way to do some upfront [Chaos Monkey](https://github.com/Netflix/chaosmonkey) type of testing on software components.
 

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack.egg-info/SOURCES.txt

@@ -16,6 +16,7 @@ uv.lock
 .github/workflows/publish-to-dockerhub.yml
 .github/workflows/publish-to-pypi.yml
 docs/configuration.md
+docs/ollama.md
 docs/assets/favicon.ico
 docs/assets/logo.png
 docs/assets/mockstack.png
@@ -28,6 +29,7 @@ examples/filefixtures-with-templates/templates/servicename-api-v1-items.j2
 examples/filefixtures-with-templates/templates/servicename2-api-v2-user.533ec889-7c68-45c8-b21e-4a7e455d1234.j2
 examples/llm/README.md
 examples/llm/mockstack-langchain-example.ipynb
+examples/llm/templates/ollama-openai-v1-chat-completions.j2
 examples/llm/templates/openai-v1-chat-completions.j2
 examples/llm/templates/openai-v1-deployments-gpt-4o-chat-completions.j2
 examples/proxyrules-with-rules-file/.env.example
@@ -37,6 +39,7 @@ mockstack/__init__.py
 mockstack/config.py
 mockstack/constants.py
 mockstack/display.py
+mockstack/exceptions.py
 mockstack/identifiers.py
 mockstack/intent.py
 mockstack/lifespan.py
@@ -50,6 +53,8 @@ mockstack.egg-info/dependency_links.txt
 mockstack.egg-info/entry_points.txt
 mockstack.egg-info/requires.txt
 mockstack.egg-info/top_level.txt
+mockstack/llm/__init__.py
+mockstack/llm/ollama.py
 mockstack/routers/__init__.py
 mockstack/routers/catchall.py
 mockstack/routers/homepage.py
@@ -69,6 +74,8 @@ mockstack/tests/test_templating.py
 mockstack/tests/fixtures/proxyrules.yml
 mockstack/tests/fixtures/templates/__init__.py
 mockstack/tests/fixtures/templates/example-template.j2
+mockstack/tests/llm/__init__.py
+mockstack/tests/llm/test_ollama.py
 mockstack/tests/routers/__init__.py
 mockstack/tests/routers/test_catchall.py
 mockstack/tests/routers/test_homepage.py

{mockstack-0.7.3 → mockstack-0.8.0}/mockstack.egg-info/requires.txt

@@ -10,3 +10,6 @@ pyyaml>=6.0.2
 rich>=14.0.0
 typer>=0.15.2
 uvicorn>=0.34.2
+
+[llm]
+ollama>=0.4.8

{mockstack-0.7.3 → mockstack-0.8.0}/pyproject.toml

@@ -41,6 +41,11 @@ Issues = "https://github.com/adamhadani/mockstack/issues"
 [project.scripts]
 mockstack = "mockstack.main:run"
 
+[project.optional-dependencies]
+llm = [
+    "ollama>=0.4.8",
+]
+
 [dependency-groups]
 dev = [
     "mypy>=1.15.0",

{mockstack-0.7.3 → mockstack-0.8.0}/uv.lock

@@ -493,6 +493,11 @@ dependencies = [
     { name = "uvicorn" },
 ]
 
+[package.optional-dependencies]
+llm = [
+    { name = "ollama" },
+]
+
 [package.dev-dependencies]
 dev = [
     { name = "mypy" },
@@ -509,6 +514,7 @@ requires-dist = [
     { name = "fastapi", extras = ["standard"], specifier = ">=0.115.12" },
     { name = "httpx", specifier = ">=0.28.1" },
     { name = "jinja2", specifier = ">=3.1.6" },
+    { name = "ollama", marker = "extra == 'llm'", specifier = ">=0.4.8" },
     { name = "opentelemetry-distro", extras = ["otlp"], specifier = ">=0.53b1" },
     { name = "opentelemetry-instrumentation-fastapi", specifier = ">=0.53b1" },
     { name = "pydantic", specifier = ">=2.11.3" },
@@ -518,6 +524,7 @@ requires-dist = [
     { name = "typer", specifier = ">=0.15.2" },
     { name = "uvicorn", specifier = ">=0.34.2" },
 ]
+provides-extras = ["llm"]
 
 [package.metadata.requires-dev]
 dev = [
@@ -569,6 +576,19 @@ wheels = [
     { url = "https://files.pythonhosted.org/packages/79/7b/2c79738432f5c924bef5071f933bcc9efd0473bac3b4aa584a6f7c1c8df8/mypy_extensions-1.1.0-py3-none-any.whl", hash = "sha256:1be4cccdb0f2482337c4743e60421de3a356cd97508abadd57d47403e94f5505", size = 4963, upload-time = "2025-04-22T14:54:22.983Z" },
 ]
 
+[[package]]
+name = "ollama"
+version = "0.4.8"
+source = { registry = "https://pypi.org/simple" }
+dependencies = [
+    { name = "httpx" },
+    { name = "pydantic" },
+]
+sdist = { url = "https://files.pythonhosted.org/packages/e2/64/709dc99030f8f46ec552f0a7da73bbdcc2da58666abfec4742ccdb2e800e/ollama-0.4.8.tar.gz", hash = "sha256:1121439d49b96fa8339842965d0616eba5deb9f8c790786cdf4c0b3df4833802", size = 12972, upload-time = "2025-04-16T21:55:14.101Z" }
+wheels = [
+    { url = "https://files.pythonhosted.org/packages/33/3f/164de150e983b3a16e8bf3d4355625e51a357e7b3b1deebe9cc1f7cb9af8/ollama-0.4.8-py3-none-any.whl", hash = "sha256:04312af2c5e72449aaebac4a2776f52ef010877c554103419d3f36066fe8af4c", size = 13325, upload-time = "2025-04-16T21:55:12.779Z" },
+]
+
 [[package]]
 name = "opentelemetry-api"
 version = "1.32.1"