Skip to content

Preset Base

Preset

Bases: ABC

Abstract factory for creating pre-wired MCPStack pipelines.

A Preset bundles a curated set of tools and configuration so users can spin up a working pipeline with zero boilerplate. Think of it as an opinionated recipe that returns a fully configured MCPStackCore you can build() and run() immediately.

Example Preset: Clinical Research Stack

This preset might combine:

  • MIMIC-IV Tool 🏥 Conversational agent over the MIMIC-IV database. Lets the LLM query structured EHR data naturally (SQL hidden under the hood).

  • Jupyter Notebook Tool 📓 Programmatic notebook creation and execution. Great for generating reproducible analyses or rich reports directly from pipeline outputs.

  • Scikit-Longitudinal Tool 📈 Utilities for longitudinal EHR data classification. Provides ready-to-use classifiers and pipelines (e.g., Cox regression, survival models, patient-level trajectory classification).

Why presets?

  • Ship opinionated tool bundles (MIMIC-IV + Jupyter + SKLong).
  • Enable reproducible clinical research workflows.
  • Save time for non-technical users (clinicians, researchers).
A quick sketch
+---------------------+      +---------------------+      +------------------------+
|   MIMIC-IV Tool     | ---> |  Scikit-Longitudinal| ---> |   Jupyter Notebook     |
| (EHR conversational)|      | (classification)    |      | (generate + run nb)    |
+---------------------+      +---------------------+      +------------------------+

Preset wires tools together, sets defaults (paths, API keys, DB creds), and returns an MCPStackCore ready to .build() and .run().

Extensibility

Presets are starting points. Users can extend the returned stack with .with_tool(...) or override configs before building.

Source code in src/MCPStack/core/preset/base.py 
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
@beartype
class Preset(ABC):
    """Abstract factory for creating pre-wired MCPStack pipelines.

    A **Preset** bundles a curated set of tools and configuration so users can
    spin up a working pipeline with **zero boilerplate**. Think of it as an
    opinionated recipe that returns a fully configured `MCPStackCore` you can
    `build()` and `run()` immediately.

    ### Example Preset: Clinical Research Stack

    This preset might combine:

    - **MIMIC-IV Tool** 🏥
      Conversational agent over the MIMIC-IV database. Lets the LLM query
      structured EHR data naturally (SQL hidden under the hood).

    - **Jupyter Notebook Tool** 📓
      Programmatic notebook creation and execution. Great for generating
      reproducible analyses or rich reports directly from pipeline outputs.

    - **Scikit-Longitudinal Tool** 📈
      Utilities for longitudinal EHR data classification. Provides ready-to-use
      classifiers and pipelines (e.g., Cox regression, survival models,
      patient-level trajectory classification).

    !!! tip "Why presets?"
        * Ship **opinionated tool bundles** (MIMIC-IV + Jupyter + SKLong).
        * Enable reproducible **clinical research workflows**.
        * Save time for non-technical users (clinicians, researchers).

    ### A quick sketch

    ```text
    +---------------------+      +---------------------+      +------------------------+
    |   MIMIC-IV Tool     | ---> |  Scikit-Longitudinal| ---> |   Jupyter Notebook     |
    | (EHR conversational)|      | (classification)    |      | (generate + run nb)    |
    +---------------------+      +---------------------+      +------------------------+
    ```

    Preset wires tools together, sets defaults (paths, API keys, DB creds),
    and returns an `MCPStackCore` ready to `.build()` and `.run()`.

    !!! note "Extensibility"
        Presets are **starting points**. Users can extend the returned stack with
        `.with_tool(...)` or override configs before building.
    """

    @classmethod
    @abstractmethod
    def create(cls, config: StackConfig | None = None, **kwargs: dict):
        """Return an MCPStackCore instance configured with tools/pipeline.

        Implementations should:
          1. Accept an optional :class:`StackConfig` (or create a default).
          2. Instantiate and compose the required tools **in dependency order**.
          3. Apply sensible defaults and merge any `**kwargs` into tool params.
          4. Return the configured (but **not yet built**) `MCPStackCore`.

        Args:
            config: Optional shared configuration (env/logging/paths). If `None`,
                implementations may construct a default `StackConfig`.
            **kwargs: Implementation-specific parameters (e.g., `mimic_db_url`,
                `notebook_dir`, `classifier_type`).

        Returns:
            MCPStackCore: A stack with tools added and ready for `build()`.

        !!! warning "Do not build here"
            `create(...)` must **not** call `.build()` or `.run()`. Leave that to
            the caller so they can modify or inspect the stack.
        """

create(config=None, **kwargs) abstractmethod classmethod

Return an MCPStackCore instance configured with tools/pipeline.

Implementations should
  1. Accept an optional :class:StackConfig (or create a default).
  2. Instantiate and compose the required tools in dependency order.
  3. Apply sensible defaults and merge any **kwargs into tool params.
  4. Return the configured (but not yet built) MCPStackCore.

Parameters:

Name Type Description Default
config StackConfig | None

Optional shared configuration (env/logging/paths). If None, implementations may construct a default StackConfig.

 None
**kwargs dict

Implementation-specific parameters (e.g., mimic_db_url, notebook_dir, classifier_type).

 {}

Returns:

Name Type Description
MCPStackCore

A stack with tools added and ready for build().

Do not build here

create(...) must not call .build() or .run(). Leave that to the caller so they can modify or inspect the stack.

Source code in src/MCPStack/core/preset/base.py 
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
@classmethod
@abstractmethod
def create(cls, config: StackConfig | None = None, **kwargs: dict):
    """Return an MCPStackCore instance configured with tools/pipeline.

    Implementations should:
      1. Accept an optional :class:`StackConfig` (or create a default).
      2. Instantiate and compose the required tools **in dependency order**.
      3. Apply sensible defaults and merge any `**kwargs` into tool params.
      4. Return the configured (but **not yet built**) `MCPStackCore`.

    Args:
        config: Optional shared configuration (env/logging/paths). If `None`,
            implementations may construct a default `StackConfig`.
        **kwargs: Implementation-specific parameters (e.g., `mimic_db_url`,
            `notebook_dir`, `classifier_type`).

    Returns:
        MCPStackCore: A stack with tools added and ready for `build()`.

    !!! warning "Do not build here"
        `create(...)` must **not** call `.build()` or `.run()`. Leave that to
        the caller so they can modify or inspect the stack.
    """