haive.core.contracts.prompt_library

Prompt template library for reusable prompts.

This module provides a library of reusable prompt templates with versioning and composition support, extracted from scattered prompt management.

Classes

PromptCategory	Category of prompt templates.
PromptLibrary	Library of reusable prompt templates.
PromptTemplate	Versioned prompt template.

Module Contents

class haive.core.contracts.prompt_library.PromptCategory(/, **data)

Bases: pydantic.BaseModel

Category of prompt templates.

Parameters:

data (Any)

name

Category name.

description

Category description.

templates

Templates in this category.

subcategories

Nested categories.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

class haive.core.contracts.prompt_library.PromptLibrary(/, **data)

Bases: pydantic.BaseModel

Library of reusable prompt templates.

Provides: - Template storage with versioning - Category-based organization - Template composition - Usage tracking - Template evolution

Parameters:

data (Any)

templates

Templates by name and version.

categories

Template categories.

tag_index

Templates indexed by tag.

latest_versions

Latest version of each template.

composition_rules

Rules for template composition.

Examples

Add a template:

>>> library = PromptLibrary()
>>> library.add_template(
...     name="analysis",
...     template=analysis_prompt,
...     contract=analysis_contract
... )

Get latest version:

>>> prompt = library.get_latest("analysis")

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

add_template(name, template, contract, version='1.0.0', tags=None, category=None)

Add a template to the library.

Parameters:

• name (str) – Template name.

• template (langchain_core.prompts.BasePromptTemplate) – Prompt template.

• contract (haive.core.contracts.prompt_config.PromptContract) – Template contract.

• version (str) – Template version.

• tags (Optional[Set[str]]) – Template tags.

• category (Optional[str]) – Template category.

Returns:

Self for chaining.

Return type:

PromptLibrary

compose_templates(template_names, composed_name, mode='sequential')

Compose multiple templates into one.

Parameters:

• template_names (List[str]) – Templates to compose.

• composed_name (str) – Name for composed template.

• mode (str) – Composition mode.

Returns:

Composed template.

Return type:

langchain_core.prompts.BasePromptTemplate

find_by_category(category)

Find templates by category.

Parameters:

category (str) – Category name.

Returns:

List of template keys.

Return type:

List[str]

find_by_tag(tag)

Find templates by tag.

Parameters:

tag (str) – Tag to search for.

Returns:

List of template keys.

Return type:

List[str]

fork_template(name, new_name, new_version='1.0.0', modifications=None)

Fork a template to create a new version.

Parameters:

• name (str) – Original template name.

• new_name (str) – New template name.

• new_version (str) – New version.

• modifications (Optional[Dict[str, Any]]) – Modifications to apply.

Returns:

Self for chaining.

Return type:

PromptLibrary

get_evolution_history(name)

Get evolution history of a template.

Parameters:

name (str) – Template name.

Returns:

List of versions in order.

Return type:

List[str]

get_latest(name)

Get latest version of a template.

Parameters:

name (str) – Template name.

Returns:

Latest template or None.

Return type:

Optional[langchain_core.prompts.BasePromptTemplate]

get_template(name, version=None)

Get a template by name and version.

Parameters:

• name (str) – Template name.

• version (Optional[str]) – Template version (latest if not specified).

Returns:

Template or None.

Return type:

Optional[langchain_core.prompts.BasePromptTemplate]

get_usage_stats()

Get usage statistics.

Returns:

Usage count by template.

Return type:

Dict[str, int]

to_dict()

Convert to dictionary for serialization.

Returns:

Library as dictionary.

Return type:

Dict[str, Any]

class haive.core.contracts.prompt_library.PromptTemplate(/, **data)

Bases: pydantic.BaseModel

Versioned prompt template.

Parameters:

data (Any)

name

Template identifier.

version

Template version.

template

The actual prompt template.

contract

Template contract.

tags

Categorization tags.

created_at

Creation timestamp.

updated_at

Last update timestamp.

usage_count

Number of times used.

parent_version

Parent version if forked.

Create a new model by parsing and validating input data from keyword arguments.

Raises [ValidationError][pydantic_core.ValidationError] if the input data cannot be validated to form a valid model.

self is explicitly positional-only to allow self as a field name.

class Config

Pydantic configuration.