type
stringclasses 5
values | name
stringlengths 1
55
| qualified_name
stringlengths 5
143
| docstring
stringlengths 0
3.59k
⌀ | filepath
stringclasses 180
values | is_public
bool 2
classes | is_private
bool 2
classes | line_start
float64 0
1.54k
⌀ | line_end
float64 0
1.56k
⌀ | annotation
stringclasses 8
values | returns
stringclasses 236
values | parameters
listlengths 0
74
⌀ | parent_class
stringclasses 298
values | value
stringclasses 112
values | bases
listlengths 0
3
⌀ | api_element_summary
stringlengths 199
23k
|
---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
class
|
ModelProvider
|
fenic.core._inference.model_catalog.ModelProvider
|
Enum representing different model providers supported by the system.
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 7 | 14 | null | null | null | null | null |
[
"Enum"
] |
Type: class
Member Name: ModelProvider
Qualified Name: fenic.core._inference.model_catalog.ModelProvider
Docstring: Enum representing different model providers supported by the system.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
TieredTokenCost
|
fenic.core._inference.model_catalog.TieredTokenCost
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 16 | 27 | null | null | null | null | null |
[] |
Type: class
Member Name: TieredTokenCost
Qualified Name: fenic.core._inference.model_catalog.TieredTokenCost
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core._inference.model_catalog.TieredTokenCost.__init__
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 17 | 27 | null | null |
[
"self",
"input_token_cost",
"cached_input_token_read_cost",
"output_token_cost",
"cached_input_token_write_cost"
] |
TieredTokenCost
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core._inference.model_catalog.TieredTokenCost.__init__
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "input_token_cost", "cached_input_token_read_cost", "output_token_cost", "cached_input_token_write_cost"]
Returns: none
Parent Class: TieredTokenCost
|
class
|
CompletionModelParameters
|
fenic.core._inference.model_catalog.CompletionModelParameters
|
Parameters for completion models including costs and context window size.
Attributes:
input_token_cost: Cost per input token in USD
cached_input_token_read_cost: Cost per cached input token read in USD
cached_input_token_write_cost: Cost per cached input token write in USD
output_token_cost: Cost per output token in USD
context_window_length: Maximum number of tokens in the context window
max_output_tokens: Maximum number of tokens the model can generate in a single request.
max_temperature: Maximum temperature for the model.
supports_profiles: Whether the model supports parameter profiles.
supports_reasoning: Whether the model supports reasoning parameter.
supports_minimal_reasoning: Whether the model supports minimal reasoning parameter. (Introduced with OpenAI gpt5 models)
supports_custom_temperature: Whether the model supports custom temperature.
supports_verbosity: Whether the model supports verbosity. (Introduced with OpenAI gpt5 models)
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 30 | 77 | null | null | null | null | null |
[] |
Type: class
Member Name: CompletionModelParameters
Qualified Name: fenic.core._inference.model_catalog.CompletionModelParameters
Docstring: Parameters for completion models including costs and context window size.
Attributes:
input_token_cost: Cost per input token in USD
cached_input_token_read_cost: Cost per cached input token read in USD
cached_input_token_write_cost: Cost per cached input token write in USD
output_token_cost: Cost per output token in USD
context_window_length: Maximum number of tokens in the context window
max_output_tokens: Maximum number of tokens the model can generate in a single request.
max_temperature: Maximum temperature for the model.
supports_profiles: Whether the model supports parameter profiles.
supports_reasoning: Whether the model supports reasoning parameter.
supports_minimal_reasoning: Whether the model supports minimal reasoning parameter. (Introduced with OpenAI gpt5 models)
supports_custom_temperature: Whether the model supports custom temperature.
supports_verbosity: Whether the model supports verbosity. (Introduced with OpenAI gpt5 models)
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core._inference.model_catalog.CompletionModelParameters.__init__
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 48 | 77 | null | null |
[
"self",
"input_token_cost",
"output_token_cost",
"context_window_length",
"max_output_tokens",
"max_temperature",
"cached_input_token_write_cost",
"cached_input_token_read_cost",
"tiered_token_costs",
"supports_profiles",
"supports_reasoning",
"supports_minimal_reasoning",
"supports_custom_temperature",
"supports_verbosity"
] |
CompletionModelParameters
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core._inference.model_catalog.CompletionModelParameters.__init__
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "input_token_cost", "output_token_cost", "context_window_length", "max_output_tokens", "max_temperature", "cached_input_token_write_cost", "cached_input_token_read_cost", "tiered_token_costs", "supports_profiles", "supports_reasoning", "supports_minimal_reasoning", "supports_custom_temperature", "supports_verbosity"]
Returns: none
Parent Class: CompletionModelParameters
|
class
|
EmbeddingModelParameters
|
fenic.core._inference.model_catalog.EmbeddingModelParameters
|
Parameters for embedding models including costs and output dimensions.
Attributes:
input_token_cost: Cost per input token in USD
output_dimensions: Number of dimensions in the embedding output
max_input_size: Maximum number of tokens in the input string
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 80 | 128 | null | null | null | null | null |
[] |
Type: class
Member Name: EmbeddingModelParameters
Qualified Name: fenic.core._inference.model_catalog.EmbeddingModelParameters
Docstring: Parameters for embedding models including costs and output dimensions.
Attributes:
input_token_cost: Cost per input token in USD
output_dimensions: Number of dimensions in the embedding output
max_input_size: Maximum number of tokens in the input string
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core._inference.model_catalog.EmbeddingModelParameters.__init__
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 89 | 113 | null | null |
[
"self",
"input_token_cost",
"allowed_output_dimensions",
"max_input_size",
"default_dimensionality"
] |
EmbeddingModelParameters
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core._inference.model_catalog.EmbeddingModelParameters.__init__
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "input_token_cost", "allowed_output_dimensions", "max_input_size", "default_dimensionality"]
Returns: none
Parent Class: EmbeddingModelParameters
|
method
|
get_possible_dimensions
|
fenic.core._inference.model_catalog.EmbeddingModelParameters.get_possible_dimensions
|
Get the possible dimensions for the model.
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 116 | 121 | null |
list[int]
|
[
"self"
] |
EmbeddingModelParameters
| null | null |
Type: method
Member Name: get_possible_dimensions
Qualified Name: fenic.core._inference.model_catalog.EmbeddingModelParameters.get_possible_dimensions
Docstring: Get the possible dimensions for the model.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self"]
Returns: list[int]
Parent Class: EmbeddingModelParameters
|
method
|
supports_dimensions
|
fenic.core._inference.model_catalog.EmbeddingModelParameters.supports_dimensions
|
Validate if requested dimensions are supported.
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 123 | 128 | null |
bool
|
[
"self",
"requested_dimensions"
] |
EmbeddingModelParameters
| null | null |
Type: method
Member Name: supports_dimensions
Qualified Name: fenic.core._inference.model_catalog.EmbeddingModelParameters.supports_dimensions
Docstring: Validate if requested dimensions are supported.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "requested_dimensions"]
Returns: bool
Parent Class: EmbeddingModelParameters
|
attribute
|
CompletionModelCollection
|
fenic.core._inference.model_catalog.CompletionModelCollection
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 131 | 131 |
TypeAlias
| null | null | null |
Dict[str, CompletionModelParameters]
| null |
Type: attribute
Member Name: CompletionModelCollection
Qualified Name: fenic.core._inference.model_catalog.CompletionModelCollection
Docstring: none
Value: Dict[str, CompletionModelParameters]
Annotation: TypeAlias
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
EmbeddingModelCollection
|
fenic.core._inference.model_catalog.EmbeddingModelCollection
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 132 | 132 |
TypeAlias
| null | null | null |
Dict[str, EmbeddingModelParameters]
| null |
Type: attribute
Member Name: EmbeddingModelCollection
Qualified Name: fenic.core._inference.model_catalog.EmbeddingModelCollection
Docstring: none
Value: Dict[str, EmbeddingModelParameters]
Annotation: TypeAlias
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
OpenAILanguageModelName
|
fenic.core._inference.model_catalog.OpenAILanguageModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 133 | 162 | null | null | null | null |
Literal['gpt-5', 'gpt-5-2025-08-07', 'gpt-5-mini', 'gpt-5-mini-2025-08-07', 'gpt-5-nano', 'gpt-5-nano-2025-08-07', 'gpt-4.1', 'gpt-4.1-mini', 'gpt-4.1-nano', 'gpt-4.1-2025-04-14', 'gpt-4.1-mini-2025-04-14', 'gpt-4.1-nano-2025-04-14', 'gpt-4o', 'gpt-4o-2024-11-20', 'gpt-4o-2024-08-06', 'gpt-4o-2024-05-13', 'gpt-4o-mini', 'gpt-4o-mini-2024-07-18', 'gpt-4-turbo', 'gpt-4-turbo-2024-04-09', 'gpt-4', 'gpt-4-0314', 'gpt-4-0613', 'o1', 'o1-mini', 'o3', 'o3-mini', 'o4-mini']
| null |
Type: attribute
Member Name: OpenAILanguageModelName
Qualified Name: fenic.core._inference.model_catalog.OpenAILanguageModelName
Docstring: none
Value: Literal['gpt-5', 'gpt-5-2025-08-07', 'gpt-5-mini', 'gpt-5-mini-2025-08-07', 'gpt-5-nano', 'gpt-5-nano-2025-08-07', 'gpt-4.1', 'gpt-4.1-mini', 'gpt-4.1-nano', 'gpt-4.1-2025-04-14', 'gpt-4.1-mini-2025-04-14', 'gpt-4.1-nano-2025-04-14', 'gpt-4o', 'gpt-4o-2024-11-20', 'gpt-4o-2024-08-06', 'gpt-4o-2024-05-13', 'gpt-4o-mini', 'gpt-4o-mini-2024-07-18', 'gpt-4-turbo', 'gpt-4-turbo-2024-04-09', 'gpt-4', 'gpt-4-0314', 'gpt-4-0613', 'o1', 'o1-mini', 'o3', 'o3-mini', 'o4-mini']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
OpenAIEmbeddingModelName
|
fenic.core._inference.model_catalog.OpenAIEmbeddingModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 164 | 166 | null | null | null | null |
Literal['text-embedding-3-small', 'text-embedding-3-large']
| null |
Type: attribute
Member Name: OpenAIEmbeddingModelName
Qualified Name: fenic.core._inference.model_catalog.OpenAIEmbeddingModelName
Docstring: none
Value: Literal['text-embedding-3-small', 'text-embedding-3-large']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
GoogleVertexEmbeddingModelName
|
fenic.core._inference.model_catalog.GoogleVertexEmbeddingModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 168 | 173 | null | null | null | null |
Literal['gemini-embedding-001', 'gemini-embedding-exp-03-07', 'text-multilingual-embedding-002', 'text-embedding-005']
| null |
Type: attribute
Member Name: GoogleVertexEmbeddingModelName
Qualified Name: fenic.core._inference.model_catalog.GoogleVertexEmbeddingModelName
Docstring: none
Value: Literal['gemini-embedding-001', 'gemini-embedding-exp-03-07', 'text-multilingual-embedding-002', 'text-embedding-005']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
GoogleDeveloperEmbeddingModelName
|
fenic.core._inference.model_catalog.GoogleDeveloperEmbeddingModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 175 | 179 | null | null | null | null |
Literal['gemini-embedding-001', 'gemini-embedding-exp-03-07', 'text-embedding-004']
| null |
Type: attribute
Member Name: GoogleDeveloperEmbeddingModelName
Qualified Name: fenic.core._inference.model_catalog.GoogleDeveloperEmbeddingModelName
Docstring: none
Value: Literal['gemini-embedding-001', 'gemini-embedding-exp-03-07', 'text-embedding-004']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
AnthropicLanguageModelName
|
fenic.core._inference.model_catalog.AnthropicLanguageModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 181 | 200 | null | null | null | null |
Literal['claude-3-7-sonnet-latest', 'claude-3-7-sonnet-20250219', 'claude-3-5-haiku-latest', 'claude-3-5-haiku-20241022', 'claude-sonnet-4-20250514', 'claude-sonnet-4-0', 'claude-4-sonnet-20250514', 'claude-3-5-sonnet-latest', 'claude-3-5-sonnet-20241022', 'claude-3-5-sonnet-20240620', 'claude-opus-4-1', 'claude-opus-4-1-20250805', 'claude-opus-4-0', 'claude-opus-4-20250514', 'claude-4-opus-20250514', 'claude-3-opus-latest', 'claude-3-opus-20240229', 'claude-3-haiku-20240307']
| null |
Type: attribute
Member Name: AnthropicLanguageModelName
Qualified Name: fenic.core._inference.model_catalog.AnthropicLanguageModelName
Docstring: none
Value: Literal['claude-3-7-sonnet-latest', 'claude-3-7-sonnet-20250219', 'claude-3-5-haiku-latest', 'claude-3-5-haiku-20241022', 'claude-sonnet-4-20250514', 'claude-sonnet-4-0', 'claude-4-sonnet-20250514', 'claude-3-5-sonnet-latest', 'claude-3-5-sonnet-20241022', 'claude-3-5-sonnet-20240620', 'claude-opus-4-1', 'claude-opus-4-1-20250805', 'claude-opus-4-0', 'claude-opus-4-20250514', 'claude-4-opus-20250514', 'claude-3-opus-latest', 'claude-3-opus-20240229', 'claude-3-haiku-20240307']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
CohereEmbeddingModelName
|
fenic.core._inference.model_catalog.CohereEmbeddingModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 202 | 208 | null | null | null | null |
Literal['embed-v4.0', 'embed-english-v3.0', 'embed-english-light-v3.0', 'embed-multilingual-v3.0', 'embed-multilingual-light-v3.0']
| null |
Type: attribute
Member Name: CohereEmbeddingModelName
Qualified Name: fenic.core._inference.model_catalog.CohereEmbeddingModelName
Docstring: none
Value: Literal['embed-v4.0', 'embed-english-v3.0', 'embed-english-light-v3.0', 'embed-multilingual-v3.0', 'embed-multilingual-light-v3.0']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
GoogleDeveloperLanguageModelName
|
fenic.core._inference.model_catalog.GoogleDeveloperLanguageModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 211 | 220 | null | null | null | null |
Literal['gemini-2.5-pro', 'gemini-2.5-flash', 'gemini-2.5-flash-lite', 'gemini-2.0-flash-lite', 'gemini-2.0-flash-lite-001', 'gemini-2.0-flash', 'gemini-2.0-flash-001', 'gemini-2.0-flash-exp']
| null |
Type: attribute
Member Name: GoogleDeveloperLanguageModelName
Qualified Name: fenic.core._inference.model_catalog.GoogleDeveloperLanguageModelName
Docstring: none
Value: Literal['gemini-2.5-pro', 'gemini-2.5-flash', 'gemini-2.5-flash-lite', 'gemini-2.0-flash-lite', 'gemini-2.0-flash-lite-001', 'gemini-2.0-flash', 'gemini-2.0-flash-001', 'gemini-2.0-flash-exp']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
GoogleVertexLanguageModelName
|
fenic.core._inference.model_catalog.GoogleVertexLanguageModelName
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 221 | 221 | null | null | null | null |
GoogleDeveloperLanguageModelName
| null |
Type: attribute
Member Name: GoogleVertexLanguageModelName
Qualified Name: fenic.core._inference.model_catalog.GoogleVertexLanguageModelName
Docstring: none
Value: GoogleDeveloperLanguageModelName
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
ProviderModelCollection
|
fenic.core._inference.model_catalog.ProviderModelCollection
|
A collection of models for a specific provider.
This class maintains a collection of models for a specific provider,
including their costs, context windows, and other parameters. It provides methods
to query model information and calculate costs for different operations.
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 223 | 270 | null | null | null | null | null |
[] |
Type: class
Member Name: ProviderModelCollection
Qualified Name: fenic.core._inference.model_catalog.ProviderModelCollection
Docstring: A collection of models for a specific provider.
This class maintains a collection of models for a specific provider,
including their costs, context windows, and other parameters. It provides methods
to query model information and calculate costs for different operations.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core._inference.model_catalog.ProviderModelCollection.__init__
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 232 | 235 | null |
None
|
[
"self",
"provider"
] |
ProviderModelCollection
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core._inference.model_catalog.ProviderModelCollection.__init__
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "provider"]
Returns: None
Parent Class: ProviderModelCollection
|
method
|
add_model
|
fenic.core._inference.model_catalog.ProviderModelCollection.add_model
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 238 | 270 | null | null |
[
"self",
"name",
"parameters",
"snapshots"
] |
ProviderModelCollection
| null | null |
Type: method
Member Name: add_model
Qualified Name: fenic.core._inference.model_catalog.ProviderModelCollection.add_model
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "name", "parameters", "snapshots"]
Returns: none
Parent Class: ProviderModelCollection
|
class
|
ModelCatalog
|
fenic.core._inference.model_catalog.ModelCatalog
|
Catalog of supported models and their parameters for different providers.
This class maintains a registry of all supported models across different providers,
including their costs, context windows, and other parameters. It provides methods
to query model information and calculate costs for different operations.
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 273 | 1,180 | null | null | null | null | null |
[] |
Type: class
Member Name: ModelCatalog
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog
Docstring: Catalog of supported models and their parameters for different providers.
This class maintains a registry of all supported models across different providers,
including their costs, context windows, and other parameters. It provides methods
to query model information and calculate costs for different operations.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core._inference.model_catalog.ModelCatalog.__init__
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 281 | 285 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.__init__
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_models
|
Initialize all models in the catalog using the new API.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 287 | 293 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_models
Docstring: Initialize all models in the catalog using the new API.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_anthropic_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_anthropic_models
|
Initialize Anthropic models in the catalog.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 295 | 414 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_anthropic_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_anthropic_models
Docstring: Initialize Anthropic models in the catalog.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_openai_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_openai_models
|
Initialize OpenAI models in the catalog.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 416 | 666 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_openai_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_openai_models
Docstring: Initialize OpenAI models in the catalog.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_google_vertex_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_google_vertex_models
|
Initialize the Google Vertex Models.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 668 | 781 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_google_vertex_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_google_vertex_models
Docstring: Initialize the Google Vertex Models.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_google_gla_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_google_gla_models
|
Initialize Google models in the catalog.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 783 | 887 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_google_gla_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_google_gla_models
Docstring: Initialize Google models in the catalog.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_initialize_cohere_models
|
fenic.core._inference.model_catalog.ModelCatalog._initialize_cohere_models
|
Initialize Cohere models in the catalog.
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 889 | 945 | null | null |
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _initialize_cohere_models
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._initialize_cohere_models
Docstring: Initialize Cohere models in the catalog.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ModelCatalog
|
method
|
get_completion_model_parameters
|
fenic.core._inference.model_catalog.ModelCatalog.get_completion_model_parameters
|
Gets the parameters for a specific completion model.
Args:
model_provider: The provider of the model
model_name: The name of the model
Returns:
Model parameters if found, None otherwise
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 948 | 959 | null |
CompletionModelParameters | None
|
[
"self",
"model_provider",
"model_name"
] |
ModelCatalog
| null | null |
Type: method
Member Name: get_completion_model_parameters
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.get_completion_model_parameters
Docstring: Gets the parameters for a specific completion model.
Args:
model_provider: The provider of the model
model_name: The name of the model
Returns:
Model parameters if found, None otherwise
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name"]
Returns: CompletionModelParameters | None
Parent Class: ModelCatalog
|
method
|
get_embedding_model_parameters
|
fenic.core._inference.model_catalog.ModelCatalog.get_embedding_model_parameters
|
Gets the parameters for a specific embedding model.
Args:
model_provider: The provider of the model
model_name: The name of the model
Returns:
Model parameters if found, None otherwise
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 961 | 971 | null |
EmbeddingModelParameters | None
|
[
"self",
"model_provider",
"model_name"
] |
ModelCatalog
| null | null |
Type: method
Member Name: get_embedding_model_parameters
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.get_embedding_model_parameters
Docstring: Gets the parameters for a specific embedding model.
Args:
model_provider: The provider of the model
model_name: The name of the model
Returns:
Model parameters if found, None otherwise
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name"]
Returns: EmbeddingModelParameters | None
Parent Class: ModelCatalog
|
method
|
generate_unsupported_completion_model_error_message
|
fenic.core._inference.model_catalog.ModelCatalog.generate_unsupported_completion_model_error_message
|
Generates an error message for unsupported completion models.
Args:
model_provider: The provider of the unsupported model
model_name: The name of the unsupported model
Returns:
Error message string
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 973 | 984 | null |
str
|
[
"self",
"model_provider",
"model_name"
] |
ModelCatalog
| null | null |
Type: method
Member Name: generate_unsupported_completion_model_error_message
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.generate_unsupported_completion_model_error_message
Docstring: Generates an error message for unsupported completion models.
Args:
model_provider: The provider of the unsupported model
model_name: The name of the unsupported model
Returns:
Error message string
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name"]
Returns: str
Parent Class: ModelCatalog
|
method
|
generate_unsupported_embedding_model_error_message
|
fenic.core._inference.model_catalog.ModelCatalog.generate_unsupported_embedding_model_error_message
|
Generates an error message for unsupported embedding models.
Args:
model_provider: The provider of the unsupported model
model_name: The name of the unsupported model
Returns:
Error message string
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 986 | 997 | null |
str
|
[
"self",
"model_provider",
"model_name"
] |
ModelCatalog
| null | null |
Type: method
Member Name: generate_unsupported_embedding_model_error_message
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.generate_unsupported_embedding_model_error_message
Docstring: Generates an error message for unsupported embedding models.
Args:
model_provider: The provider of the unsupported model
model_name: The name of the unsupported model
Returns:
Error message string
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name"]
Returns: str
Parent Class: ModelCatalog
|
method
|
get_supported_completions_models_as_string
|
fenic.core._inference.model_catalog.ModelCatalog.get_supported_completions_models_as_string
|
Returns a comma-separated string of all supported completion models.
Returns:
Comma-separated string of model names in format 'provider:model'
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 999 | 1,009 | null |
str
|
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: get_supported_completions_models_as_string
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.get_supported_completions_models_as_string
Docstring: Returns a comma-separated string of all supported completion models.
Returns:
Comma-separated string of model names in format 'provider:model'
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self"]
Returns: str
Parent Class: ModelCatalog
|
method
|
get_supported_embeddings_models_as_string
|
fenic.core._inference.model_catalog.ModelCatalog.get_supported_embeddings_models_as_string
|
Returns a comma-separated string of all supported embedding models.
Returns:
Comma-separated string of model names
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 1,011 | 1,021 | null |
str
|
[
"self"
] |
ModelCatalog
| null | null |
Type: method
Member Name: get_supported_embeddings_models_as_string
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.get_supported_embeddings_models_as_string
Docstring: Returns a comma-separated string of all supported embedding models.
Returns:
Comma-separated string of model names
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self"]
Returns: str
Parent Class: ModelCatalog
|
method
|
calculate_completion_model_cost
|
fenic.core._inference.model_catalog.ModelCatalog.calculate_completion_model_cost
|
Calculates the total cost for a completion model operation.
Args:
model_provider: The provider of the model
model_name: The name of the model
uncached_input_tokens: Number of uncached input tokens
cached_input_tokens_read: Number of cached input tokens read
output_tokens: Number of output tokens
cached_input_tokens_written: Number of cached input tokens written
Returns:
Total cost in USD
Raises:
ValueError: If the model is not supported
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 1,023 | 1,067 | null |
float
|
[
"self",
"model_provider",
"model_name",
"uncached_input_tokens",
"cached_input_tokens_read",
"output_tokens",
"cached_input_tokens_written"
] |
ModelCatalog
| null | null |
Type: method
Member Name: calculate_completion_model_cost
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.calculate_completion_model_cost
Docstring: Calculates the total cost for a completion model operation.
Args:
model_provider: The provider of the model
model_name: The name of the model
uncached_input_tokens: Number of uncached input tokens
cached_input_tokens_read: Number of cached input tokens read
output_tokens: Number of output tokens
cached_input_tokens_written: Number of cached input tokens written
Returns:
Total cost in USD
Raises:
ValueError: If the model is not supported
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name", "uncached_input_tokens", "cached_input_tokens_read", "output_tokens", "cached_input_tokens_written"]
Returns: float
Parent Class: ModelCatalog
|
method
|
calculate_embedding_model_cost
|
fenic.core._inference.model_catalog.ModelCatalog.calculate_embedding_model_cost
|
Calculates the total cost for an embedding model operation.
Args:
model_provider: The provider of the model
model_name: The name of the model
billable_inputs: Number of tokens or characters to embed (some Google models charge per character)
Returns:
Total cost in USD
Raises:
ValueError: If the model is not supported
|
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 1,069 | 1,092 | null |
float
|
[
"self",
"model_provider",
"model_name",
"billable_inputs"
] |
ModelCatalog
| null | null |
Type: method
Member Name: calculate_embedding_model_cost
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog.calculate_embedding_model_cost
Docstring: Calculates the total cost for an embedding model operation.
Args:
model_provider: The provider of the model
model_name: The name of the model
billable_inputs: Number of tokens or characters to embed (some Google models charge per character)
Returns:
Total cost in USD
Raises:
ValueError: If the model is not supported
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "model_provider", "model_name", "billable_inputs"]
Returns: float
Parent Class: ModelCatalog
|
method
|
_add_model_to_catalog
|
fenic.core._inference.model_catalog.ModelCatalog._add_model_to_catalog
|
A helper method to add a model to the catalog, adding it to the appropriate provider collection.
Args:
model_provider: The provider of the model
name: The name of the model
parameters: The parameters of the model
snapshots: The optional list of snapshot names for the model
Raises:
InternalError: If the model already exists in the catalog
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 1,094 | 1,116 | null | null |
[
"self",
"model_provider",
"name",
"parameters",
"snapshots"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _add_model_to_catalog
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._add_model_to_catalog
Docstring: A helper method to add a model to the catalog, adding it to the appropriate provider collection.
Args:
model_provider: The provider of the model
name: The name of the model
parameters: The parameters of the model
snapshots: The optional list of snapshot names for the model
Raises:
InternalError: If the model already exists in the catalog
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "model_provider", "name", "parameters", "snapshots"]
Returns: none
Parent Class: ModelCatalog
|
method
|
_get_supported_completions_models_by_provider
|
fenic.core._inference.model_catalog.ModelCatalog._get_supported_completions_models_by_provider
|
Returns the collection of completion models for a specific provider.
Args:
model_provider: The provider to get models for
Returns:
Collection of completion models for the specified provider, including snapshots
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 1,118 | 1,129 | null |
CompletionModelCollection
|
[
"self",
"model_provider"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _get_supported_completions_models_by_provider
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._get_supported_completions_models_by_provider
Docstring: Returns the collection of completion models for a specific provider.
Args:
model_provider: The provider to get models for
Returns:
Collection of completion models for the specified provider, including snapshots
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "model_provider"]
Returns: CompletionModelCollection
Parent Class: ModelCatalog
|
method
|
_get_supported_embeddings_models_by_provider
|
fenic.core._inference.model_catalog.ModelCatalog._get_supported_embeddings_models_by_provider
|
Returns the collection of embedding models for a specific provider.
Args:
model_provider: The provider to get models for
Returns:
Collection of embedding models for the specified provider, including snapshots
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 1,131 | 1,142 | null |
EmbeddingModelCollection
|
[
"self",
"model_provider"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _get_supported_embeddings_models_by_provider
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._get_supported_embeddings_models_by_provider
Docstring: Returns the collection of embedding models for a specific provider.
Args:
model_provider: The provider to get models for
Returns:
Collection of embedding models for the specified provider, including snapshots
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "model_provider"]
Returns: EmbeddingModelCollection
Parent Class: ModelCatalog
|
method
|
_get_supported_completions_models_by_provider_as_string
|
fenic.core._inference.model_catalog.ModelCatalog._get_supported_completions_models_by_provider_as_string
|
Returns a comma-separated string of supported completion model names for a provider.
Args:
model_provider: The provider to get model names for
Returns:
Comma-separated string of model names
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 1,144 | 1,161 | null |
str
|
[
"self",
"model_provider"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _get_supported_completions_models_by_provider_as_string
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._get_supported_completions_models_by_provider_as_string
Docstring: Returns a comma-separated string of supported completion model names for a provider.
Args:
model_provider: The provider to get model names for
Returns:
Comma-separated string of model names
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "model_provider"]
Returns: str
Parent Class: ModelCatalog
|
method
|
_get_supported_embeddings_models_by_provider_as_string
|
fenic.core._inference.model_catalog.ModelCatalog._get_supported_embeddings_models_by_provider_as_string
|
Returns a comma-separated string of supported completion model names for a provider.
Args:
model_provider: The provider to get model names for
Returns:
Comma-separated string of model names
|
site-packages/fenic/core/_inference/model_catalog.py
| false | true | 1,163 | 1,180 | null |
str
|
[
"self",
"model_provider"
] |
ModelCatalog
| null | null |
Type: method
Member Name: _get_supported_embeddings_models_by_provider_as_string
Qualified Name: fenic.core._inference.model_catalog.ModelCatalog._get_supported_embeddings_models_by_provider_as_string
Docstring: Returns a comma-separated string of supported completion model names for a provider.
Args:
model_provider: The provider to get model names for
Returns:
Comma-separated string of model names
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "model_provider"]
Returns: str
Parent Class: ModelCatalog
|
attribute
|
model_catalog
|
fenic.core._inference.model_catalog.model_catalog
| null |
site-packages/fenic/core/_inference/model_catalog.py
| true | false | 1,184 | 1,184 | null | null | null | null |
ModelCatalog()
| null |
Type: attribute
Member Name: model_catalog
Qualified Name: fenic.core._inference.model_catalog.model_catalog
Docstring: none
Value: ModelCatalog()
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
module
|
mcp
|
fenic.core.mcp
|
MCP Tool Generation/FastMCP Server Management.
|
site-packages/fenic/core/mcp/__init__.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: mcp
Qualified Name: fenic.core.mcp
Docstring: MCP Tool Generation/FastMCP Server Management.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
__all__
|
fenic.core.mcp.__all__
| null |
site-packages/fenic/core/mcp/__init__.py
| false | false | 8 | 13 | null | null | null | null |
['BoundToolParam', 'ParameterizedToolDefinition', 'ToolParam']
| null |
Type: attribute
Member Name: __all__
Qualified Name: fenic.core.mcp.__all__
Docstring: none
Value: ['BoundToolParam', 'ParameterizedToolDefinition', 'ToolParam']
Annotation: none
is Public? : false
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
module
|
_binder
|
fenic.core.mcp._binder
| null |
site-packages/fenic/core/mcp/_binder.py
| false | true | null | null | null | null | null | null | null | null |
Type: module
Member Name: _binder
Qualified Name: fenic.core.mcp._binder
Docstring: none
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: none
Returns: none
Parent Class: none
|
function
|
collect_unresolved_parameters
|
fenic.core.mcp._binder.collect_unresolved_parameters
|
Return the set of parameter names referenced by UnresolvedLiteralExpr in the plan.
|
site-packages/fenic/core/mcp/_binder.py
| true | false | 18 | 21 | null |
dict[str, UnresolvedLiteralExpr]
|
[
"plan"
] | null | null | null |
Type: function
Member Name: collect_unresolved_parameters
Qualified Name: fenic.core.mcp._binder.collect_unresolved_parameters
Docstring: Return the set of parameter names referenced by UnresolvedLiteralExpr in the plan.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["plan"]
Returns: dict[str, UnresolvedLiteralExpr]
Parent Class: none
|
function
|
bind_parameters
|
fenic.core.mcp._binder.bind_parameters
|
Bind provided params into a LogicalPlan by replacing placeholder literals.
- Replaces every UnresolvedLiteralExpr with a LiteralExpr(value, data_type)
- Raises PlanError if any placeholder parameter is missing in params
- Raises TypeMismatchError if provided value does not match placeholder data_type
- Operates in-place on the plan/expressions and returns the same plan reference
Note: This only substitutes literal values. It does not change node schemas
and assumes placeholder data types match the provided values semantically.
|
site-packages/fenic/core/mcp/_binder.py
| true | false | 24 | 128 | null |
LogicalPlan
|
[
"plan",
"params",
"tool_params"
] | null | null | null |
Type: function
Member Name: bind_parameters
Qualified Name: fenic.core.mcp._binder.bind_parameters
Docstring: Bind provided params into a LogicalPlan by replacing placeholder literals.
- Replaces every UnresolvedLiteralExpr with a LiteralExpr(value, data_type)
- Raises PlanError if any placeholder parameter is missing in params
- Raises TypeMismatchError if provided value does not match placeholder data_type
- Operates in-place on the plan/expressions and returns the same plan reference
Note: This only substitutes literal values. It does not change node schemas
and assumes placeholder data types match the provided values semantically.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["plan", "params", "tool_params"]
Returns: LogicalPlan
Parent Class: none
|
module
|
_tools
|
fenic.core.mcp._tools
| null |
site-packages/fenic/core/mcp/_tools.py
| false | true | null | null | null | null | null | null | null | null |
Type: module
Member Name: _tools
Qualified Name: fenic.core.mcp._tools
Docstring: none
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
logger
|
fenic.core.mcp._tools.logger
| null |
site-packages/fenic/core/mcp/_tools.py
| true | false | 22 | 22 | null | null | null | null |
logging.getLogger(__name__)
| null |
Type: attribute
Member Name: logger
Qualified Name: fenic.core.mcp._tools.logger
Docstring: none
Value: logging.getLogger(__name__)
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
function
|
bind_tool
|
fenic.core.mcp._tools.bind_tool
|
Create a tool from a query and a set of parameters.
Raises PlanError if the logical plan contains unresolved parameters that are not in the tool parameters.
|
site-packages/fenic/core/mcp/_tools.py
| true | false | 24 | 94 | null |
ParameterizedToolDefinition
|
[
"name",
"description",
"params",
"result_limit",
"query"
] | null | null | null |
Type: function
Member Name: bind_tool
Qualified Name: fenic.core.mcp._tools.bind_tool
Docstring: Create a tool from a query and a set of parameters.
Raises PlanError if the logical plan contains unresolved parameters that are not in the tool parameters.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["name", "description", "params", "result_limit", "query"]
Returns: ParameterizedToolDefinition
Parent Class: none
|
function
|
create_pydantic_model_for_tool
|
fenic.core.mcp._tools.create_pydantic_model_for_tool
|
Create a Pydantic model for a tool.
|
site-packages/fenic/core/mcp/_tools.py
| true | false | 97 | 133 | null |
type[BaseModel]
|
[
"tool"
] | null | null | null |
Type: function
Member Name: create_pydantic_model_for_tool
Qualified Name: fenic.core.mcp._tools.create_pydantic_model_for_tool
Docstring: Create a Pydantic model for a tool.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["tool"]
Returns: type[BaseModel]
Parent Class: none
|
module
|
types
|
fenic.core.mcp.types
|
Exported Types related to Parameterized View/MCP Tool Generation.
|
site-packages/fenic/core/mcp/types.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: types
Qualified Name: fenic.core.mcp.types
Docstring: Exported Types related to Parameterized View/MCP Tool Generation.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
ToolParameterType
|
fenic.core.mcp.types.ToolParameterType
| null |
site-packages/fenic/core/mcp/types.py
| true | false | 13 | 13 | null | null | null | null |
Union[str, int, float, bool, list, dict]
| null |
Type: attribute
Member Name: ToolParameterType
Qualified Name: fenic.core.mcp.types.ToolParameterType
Docstring: none
Value: Union[str, int, float, bool, list, dict]
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
TableFormat
|
fenic.core.mcp.types.TableFormat
| null |
site-packages/fenic/core/mcp/types.py
| true | false | 14 | 14 | null | null | null | null |
Literal['structured', 'markdown']
| null |
Type: attribute
Member Name: TableFormat
Qualified Name: fenic.core.mcp.types.TableFormat
Docstring: none
Value: Literal['structured', 'markdown']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
ToolParam
|
fenic.core.mcp.types.ToolParam
|
A parameter for a parameterized view tool.
A parameter is a named value that can be passed to a tool. These are matched to the
parameter names of the `tool_param` UnresolvedLiteralExpr expressions captured in the Logical Plan.
Attributes:
name: The name of the parameter.
description: The description of the parameter.
allowed_values: The allowed values for the parameter.
has_default: Whether the parameter has a default value.
default_value: The default value for the parameter.
|
site-packages/fenic/core/mcp/types.py
| true | false | 17 | 50 | null | null | null | null | null |
[
"BaseModel"
] |
Type: class
Member Name: ToolParam
Qualified Name: fenic.core.mcp.types.ToolParam
Docstring: A parameter for a parameterized view tool.
A parameter is a named value that can be passed to a tool. These are matched to the
parameter names of the `tool_param` UnresolvedLiteralExpr expressions captured in the Logical Plan.
Attributes:
name: The name of the parameter.
description: The description of the parameter.
allowed_values: The allowed values for the parameter.
has_default: Whether the parameter has a default value.
default_value: The default value for the parameter.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
_check_default_value
|
fenic.core.mcp.types.ToolParam._check_default_value
| null |
site-packages/fenic/core/mcp/types.py
| false | true | 36 | 41 | null | null |
[
"self"
] |
ToolParam
| null | null |
Type: method
Member Name: _check_default_value
Qualified Name: fenic.core.mcp.types.ToolParam._check_default_value
Docstring: none
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self"]
Returns: none
Parent Class: ToolParam
|
class
|
BoundToolParam
|
fenic.core.mcp.types.BoundToolParam
|
A bound tool parameter.
A bound tool parameter is a parameter that has been bound to a specific, typed,
`tool_param` usage within a Dataframe.
|
site-packages/fenic/core/mcp/types.py
| true | false | 53 | 66 | null | null | null | null | null |
[] |
Type: class
Member Name: BoundToolParam
Qualified Name: fenic.core.mcp.types.BoundToolParam
Docstring: A bound tool parameter.
A bound tool parameter is a parameter that has been bound to a specific, typed,
`tool_param` usage within a Dataframe.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
ParameterizedToolDefinition
|
fenic.core.mcp.types.ParameterizedToolDefinition
|
A tool that has been bound to a specific Parameterized View.
|
site-packages/fenic/core/mcp/types.py
| true | false | 69 | 76 | null | null | null | null | null |
[] |
Type: class
Member Name: ParameterizedToolDefinition
Qualified Name: fenic.core.mcp.types.ParameterizedToolDefinition
Docstring: A tool that has been bound to a specific Parameterized View.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
module
|
_server
|
fenic.core.mcp._server
|
MCP server and tool integration for Fenic.
This module exposes a small wrapper that registers Tool Definitions
with FastMCP, collects results, and formats them as table-like output
for model consumption. FastMCP is an optional dependency.
Install with:
pip install "fenic[mcp]"
# or
pip install fastmcp
|
site-packages/fenic/core/mcp/_server.py
| false | true | null | null | null | null | null | null | null | null |
Type: module
Member Name: _server
Qualified Name: fenic.core.mcp._server
Docstring: MCP server and tool integration for Fenic.
This module exposes a small wrapper that registers Tool Definitions
with FastMCP, collects results, and formats them as table-like output
for model consumption. FastMCP is an optional dependency.
Install with:
pip install "fenic[mcp]"
# or
pip install fastmcp
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
logger
|
fenic.core.mcp._server.logger
| null |
site-packages/fenic/core/mcp/_server.py
| true | false | 35 | 35 | null | null | null | null |
logging.getLogger(__name__)
| null |
Type: attribute
Member Name: logger
Qualified Name: fenic.core.mcp._server.logger
Docstring: none
Value: logging.getLogger(__name__)
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
MCPResultSet
|
fenic.core.mcp._server.MCPResultSet
|
Structured result returned to the MCP client.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 38 | 42 | null | null | null | null | null |
[
"BaseModel"
] |
Type: class
Member Name: MCPResultSet
Qualified Name: fenic.core.mcp._server.MCPResultSet
Docstring: Structured result returned to the MCP client.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
MCPTransport
|
fenic.core.mcp._server.MCPTransport
| null |
site-packages/fenic/core/mcp/_server.py
| true | false | 44 | 44 | null | null | null | null |
Literal['http', 'stdio']
| null |
Type: attribute
Member Name: MCPTransport
Qualified Name: fenic.core.mcp._server.MCPTransport
Docstring: none
Value: Literal['http', 'stdio']
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
FenicMCPServer
|
fenic.core.mcp._server.FenicMCPServer
|
Register Fenic tools and serve them via FastMCP.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 46 | 146 | null | null | null | null | null |
[] |
Type: class
Member Name: FenicMCPServer
Qualified Name: fenic.core.mcp._server.FenicMCPServer
Docstring: Register Fenic tools and serve them via FastMCP.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
__init__
|
fenic.core.mcp._server.FenicMCPServer.__init__
|
Initialize the server with a Fenic session state and tool list.
Args:
session_state: Fenic session state to use for tool execution.
paramaterized_tools: List of user-created tools
server_name: Name of the MCP server.
concurrency_limit: Maximum number of concurrent tool executions.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 49 | 80 | null | null |
[
"self",
"session_state",
"paramaterized_tools",
"server_name",
"concurrency_limit"
] |
FenicMCPServer
| null | null |
Type: method
Member Name: __init__
Qualified Name: fenic.core.mcp._server.FenicMCPServer.__init__
Docstring: Initialize the server with a Fenic session state and tool list.
Args:
session_state: Fenic session state to use for tool execution.
paramaterized_tools: List of user-created tools
server_name: Name of the MCP server.
concurrency_limit: Maximum number of concurrent tool executions.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "session_state", "paramaterized_tools", "server_name", "concurrency_limit"]
Returns: none
Parent Class: FenicMCPServer
|
method
|
run_async
|
fenic.core.mcp._server.FenicMCPServer.run_async
|
Run the MCP server asynchronously.
Args:
transport: Transport protocol to use (http, stdio).
kwargs: Additional transport-specific arguments to pass to FastMCP.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 82 | 91 | null | null |
[
"self",
"transport",
"kwargs"
] |
FenicMCPServer
| null | null |
Type: method
Member Name: run_async
Qualified Name: fenic.core.mcp._server.FenicMCPServer.run_async
Docstring: Run the MCP server asynchronously.
Args:
transport: Transport protocol to use (http, stdio).
kwargs: Additional transport-specific arguments to pass to FastMCP.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "transport", "kwargs"]
Returns: none
Parent Class: FenicMCPServer
|
method
|
run
|
fenic.core.mcp._server.FenicMCPServer.run
|
Run the MCP server. This is a synchronous function.
Args:
transport: Transport protocol to use (http, stdio).
kwargs: Additional transport-specific arguments to pass to FastMCP.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 93 | 102 | null | null |
[
"self",
"transport",
"kwargs"
] |
FenicMCPServer
| null | null |
Type: method
Member Name: run
Qualified Name: fenic.core.mcp._server.FenicMCPServer.run
Docstring: Run the MCP server. This is a synchronous function.
Args:
transport: Transport protocol to use (http, stdio).
kwargs: Additional transport-specific arguments to pass to FastMCP.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "transport", "kwargs"]
Returns: none
Parent Class: FenicMCPServer
|
method
|
http_app
|
fenic.core.mcp._server.FenicMCPServer.http_app
|
Create a Starlette ASGI app for the MCP server.
|
site-packages/fenic/core/mcp/_server.py
| true | false | 104 | 106 | null | null |
[
"self",
"kwargs"
] |
FenicMCPServer
| null | null |
Type: method
Member Name: http_app
Qualified Name: fenic.core.mcp._server.FenicMCPServer.http_app
Docstring: Create a Starlette ASGI app for the MCP server.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["self", "kwargs"]
Returns: none
Parent Class: FenicMCPServer
|
method
|
_build_parameterized_tool
|
fenic.core.mcp._server.FenicMCPServer._build_parameterized_tool
|
Build a Pydantic single-parameter tool function for ResolvedTool.
|
site-packages/fenic/core/mcp/_server.py
| false | true | 108 | 146 | null | null |
[
"self",
"tool"
] |
FenicMCPServer
| null | null |
Type: method
Member Name: _build_parameterized_tool
Qualified Name: fenic.core.mcp._server.FenicMCPServer._build_parameterized_tool
Docstring: Build a Pydantic single-parameter tool function for ResolvedTool.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["self", "tool"]
Returns: none
Parent Class: FenicMCPServer
|
function
|
_to_snake_case
|
fenic.core.mcp._server._to_snake_case
| null |
site-packages/fenic/core/mcp/_server.py
| false | true | 149 | 157 | null |
str
|
[
"name"
] | null | null | null |
Type: function
Member Name: _to_snake_case
Qualified Name: fenic.core.mcp._server._to_snake_case
Docstring: none
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["name"]
Returns: str
Parent Class: none
|
function
|
_render_markdown_preview
|
fenic.core.mcp._server._render_markdown_preview
| null |
site-packages/fenic/core/mcp/_server.py
| false | true | 159 | 168 | null |
str
|
[
"rows"
] | null | null | null |
Type: function
Member Name: _render_markdown_preview
Qualified Name: fenic.core.mcp._server._render_markdown_preview
Docstring: none
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["rows"]
Returns: str
Parent Class: none
|
module
|
_serde
|
fenic.core._serde
|
LogicalPlan serialization implementations.
|
site-packages/fenic/core/_serde/__init__.py
| false | true | null | null | null | null | null | null | null | null |
Type: module
Member Name: _serde
Qualified Name: fenic.core._serde
Docstring: LogicalPlan serialization implementations.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
__all__
|
fenic.core._serde.__all__
| null |
site-packages/fenic/core/_serde/__init__.py
| false | false | 7 | 7 | null | null | null | null |
['CloudPickleSerde', 'LogicalPlanSerde', 'ProtoSerde']
| null |
Type: attribute
Member Name: __all__
Qualified Name: fenic.core._serde.__all__
Docstring: none
Value: ['CloudPickleSerde', 'LogicalPlanSerde', 'ProtoSerde']
Annotation: none
is Public? : false
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
module
|
serde_protocol
|
fenic.core._serde.serde_protocol
|
Protocol for serialization/deserialization of LogicalPlan objects.
|
site-packages/fenic/core/_serde/serde_protocol.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: serde_protocol
Qualified Name: fenic.core._serde.serde_protocol
Docstring: Protocol for serialization/deserialization of LogicalPlan objects.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
SupportsLogicalPlanSerde
|
fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde
|
Protocol for LogicalPlan serialization.
|
site-packages/fenic/core/_serde/serde_protocol.py
| true | false | 10 | 38 | null | null | null | null | null |
[
"Protocol"
] |
Type: class
Member Name: SupportsLogicalPlanSerde
Qualified Name: fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde
Docstring: Protocol for LogicalPlan serialization.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
serialize
|
fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde.serialize
|
Serialize a LogicalPlan to bytes.
Removes any local session state refs from the plan.
Args:
plan: The LogicalPlan to serialize
Returns:
bytes: The serialized plan
|
site-packages/fenic/core/_serde/serde_protocol.py
| true | false | 14 | 26 | null |
bytes
|
[
"plan"
] |
SupportsLogicalPlanSerde
| null | null |
Type: method
Member Name: serialize
Qualified Name: fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde.serialize
Docstring: Serialize a LogicalPlan to bytes.
Removes any local session state refs from the plan.
Args:
plan: The LogicalPlan to serialize
Returns:
bytes: The serialized plan
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["plan"]
Returns: bytes
Parent Class: SupportsLogicalPlanSerde
|
method
|
deserialize
|
fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde.deserialize
|
Deserialize bytes back into a LogicalPlan.
Args:
data: The serialized plan data
Returns:
The deserialized plan
|
site-packages/fenic/core/_serde/serde_protocol.py
| true | false | 28 | 38 | null |
LogicalPlan
|
[
"data"
] |
SupportsLogicalPlanSerde
| null | null |
Type: method
Member Name: deserialize
Qualified Name: fenic.core._serde.serde_protocol.SupportsLogicalPlanSerde.deserialize
Docstring: Deserialize bytes back into a LogicalPlan.
Args:
data: The serialized plan data
Returns:
The deserialized plan
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["data"]
Returns: LogicalPlan
Parent Class: SupportsLogicalPlanSerde
|
module
|
cloudpickle_serde
|
fenic.core._serde.cloudpickle_serde
|
CloudPickle-based implementation of LogicalPlan serialization.
|
site-packages/fenic/core/_serde/cloudpickle_serde.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: cloudpickle_serde
Qualified Name: fenic.core._serde.cloudpickle_serde
Docstring: CloudPickle-based implementation of LogicalPlan serialization.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
CloudPickleSerde
|
fenic.core._serde.cloudpickle_serde.CloudPickleSerde
|
CloudPickle-based LogicalPlan serialization implementation.
|
site-packages/fenic/core/_serde/cloudpickle_serde.py
| true | false | 13 | 39 | null | null | null | null | null |
[
"SupportsLogicalPlanSerde"
] |
Type: class
Member Name: CloudPickleSerde
Qualified Name: fenic.core._serde.cloudpickle_serde.CloudPickleSerde
Docstring: CloudPickle-based LogicalPlan serialization implementation.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
serialize
|
fenic.core._serde.cloudpickle_serde.CloudPickleSerde.serialize
|
Serialize a LogicalPlan to bytes using cloudpickle.
Args:
plan: The LogicalPlan to serialize
Returns:
bytes: The serialized plan
|
site-packages/fenic/core/_serde/cloudpickle_serde.py
| true | false | 16 | 26 | null |
bytes
|
[
"plan"
] |
CloudPickleSerde
| null | null |
Type: method
Member Name: serialize
Qualified Name: fenic.core._serde.cloudpickle_serde.CloudPickleSerde.serialize
Docstring: Serialize a LogicalPlan to bytes using cloudpickle.
Args:
plan: The LogicalPlan to serialize
Returns:
bytes: The serialized plan
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["plan"]
Returns: bytes
Parent Class: CloudPickleSerde
|
method
|
deserialize
|
fenic.core._serde.cloudpickle_serde.CloudPickleSerde.deserialize
|
Deserialize bytes back into a LogicalPlan using cloudpickle.
Args:
data: The serialized plan data
Returns:
The deserialized plan
|
site-packages/fenic/core/_serde/cloudpickle_serde.py
| true | false | 28 | 39 | null |
LogicalPlan
|
[
"data"
] |
CloudPickleSerde
| null | null |
Type: method
Member Name: deserialize
Qualified Name: fenic.core._serde.cloudpickle_serde.CloudPickleSerde.deserialize
Docstring: Deserialize bytes back into a LogicalPlan using cloudpickle.
Args:
data: The serialized plan data
Returns:
The deserialized plan
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["data"]
Returns: LogicalPlan
Parent Class: CloudPickleSerde
|
module
|
serde
|
fenic.core._serde.serde
|
LogicalPlan serialization with pluggable backends.
|
site-packages/fenic/core/_serde/serde.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: serde
Qualified Name: fenic.core._serde.serde
Docstring: LogicalPlan serialization with pluggable backends.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
_default_serde_type
|
fenic.core._serde.serde._default_serde_type
| null |
site-packages/fenic/core/_serde/serde.py
| false | true | 13 | 13 | null | null | null | null |
ProtoSerde
| null |
Type: attribute
Member Name: _default_serde_type
Qualified Name: fenic.core._serde.serde._default_serde_type
Docstring: none
Value: ProtoSerde
Annotation: none
is Public? : false
is Private? : true
Parameters: none
Returns: none
Parent Class: none
|
class
|
LogicalPlanSerde
|
fenic.core._serde.serde.LogicalPlanSerde
|
Facade for LogicalPlan serialization with pluggable backends.
|
site-packages/fenic/core/_serde/serde.py
| true | false | 16 | 32 | null | null | null | null | null |
[
"SupportsLogicalPlanSerde"
] |
Type: class
Member Name: LogicalPlanSerde
Qualified Name: fenic.core._serde.serde.LogicalPlanSerde
Docstring: Facade for LogicalPlan serialization with pluggable backends.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
serialize
|
fenic.core._serde.serde.LogicalPlanSerde.serialize
|
Serialize a LogicalPlan to bytes.
|
site-packages/fenic/core/_serde/serde.py
| true | false | 21 | 24 | null |
bytes
|
[
"cls",
"plan"
] |
LogicalPlanSerde
| null | null |
Type: method
Member Name: serialize
Qualified Name: fenic.core._serde.serde.LogicalPlanSerde.serialize
Docstring: Serialize a LogicalPlan to bytes.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["cls", "plan"]
Returns: bytes
Parent Class: LogicalPlanSerde
|
method
|
deserialize
|
fenic.core._serde.serde.LogicalPlanSerde.deserialize
|
Deserialize a LogicalPlan from bytes.
|
site-packages/fenic/core/_serde/serde.py
| true | false | 26 | 32 | null |
LogicalPlan
|
[
"cls",
"serialized_plan"
] |
LogicalPlanSerde
| null | null |
Type: method
Member Name: deserialize
Qualified Name: fenic.core._serde.serde.LogicalPlanSerde.deserialize
Docstring: Deserialize a LogicalPlan from bytes.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["cls", "serialized_plan"]
Returns: LogicalPlan
Parent Class: LogicalPlanSerde
|
module
|
proto
|
fenic.core._serde.proto
| null |
site-packages/fenic/core/_serde/proto/__init__.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: proto
Qualified Name: fenic.core._serde.proto
Docstring: none
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
module
|
proto_serde
|
fenic.core._serde.proto.proto_serde
|
Main API for logical plan and expression serialization/deserialization.
|
site-packages/fenic/core/_serde/proto/proto_serde.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: proto_serde
Qualified Name: fenic.core._serde.proto.proto_serde
Docstring: Main API for logical plan and expression serialization/deserialization.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
ProtoSerde
|
fenic.core._serde.proto.proto_serde.ProtoSerde
|
Proto Serde implementation.
This implementation uses the Protobuf specs defined in the `protos` package to serialize
and deserialize logical plans. Provides the main API for converting between LogicalPlan
objects and their binary protobuf representation.
|
site-packages/fenic/core/_serde/proto/proto_serde.py
| true | false | 11 | 52 | null | null | null | null | null |
[
"SupportsLogicalPlanSerde"
] |
Type: class
Member Name: ProtoSerde
Qualified Name: fenic.core._serde.proto.proto_serde.ProtoSerde
Docstring: Proto Serde implementation.
This implementation uses the Protobuf specs defined in the `protos` package to serialize
and deserialize logical plans. Provides the main API for converting between LogicalPlan
objects and their binary protobuf representation.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
method
|
serialize
|
fenic.core._serde.proto.proto_serde.ProtoSerde.serialize
|
Serialize a logical plan to binary protobuf format.
Args:
logical_plan: The logical plan to serialize.
Returns:
Binary protobuf representation of the logical plan.
|
site-packages/fenic/core/_serde/proto/proto_serde.py
| true | false | 22 | 34 | null |
bytes
|
[
"cls",
"logical_plan"
] |
ProtoSerde
| null | null |
Type: method
Member Name: serialize
Qualified Name: fenic.core._serde.proto.proto_serde.ProtoSerde.serialize
Docstring: Serialize a logical plan to binary protobuf format.
Args:
logical_plan: The logical plan to serialize.
Returns:
Binary protobuf representation of the logical plan.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["cls", "logical_plan"]
Returns: bytes
Parent Class: ProtoSerde
|
method
|
deserialize
|
fenic.core._serde.proto.proto_serde.ProtoSerde.deserialize
|
Deserialize a logical plan from binary protobuf format.
Args:
data: Binary protobuf data to deserialize.
Returns:
The deserialized logical plan.
|
site-packages/fenic/core/_serde/proto/proto_serde.py
| true | false | 36 | 52 | null |
LogicalPlan
|
[
"cls",
"data"
] |
ProtoSerde
| null | null |
Type: method
Member Name: deserialize
Qualified Name: fenic.core._serde.proto.proto_serde.ProtoSerde.deserialize
Docstring: Deserialize a logical plan from binary protobuf format.
Args:
data: Binary protobuf data to deserialize.
Returns:
The deserialized logical plan.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["cls", "data"]
Returns: LogicalPlan
Parent Class: ProtoSerde
|
module
|
plan_serde
|
fenic.core._serde.proto.plan_serde
|
Logical plan serialization/deserialization using singledispatch.
This module provides the main dispatch functions for plan serialization.
The actual serialization implementations are organized in the plans/ subdirectory.
|
site-packages/fenic/core/_serde/proto/plan_serde.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: plan_serde
Qualified Name: fenic.core._serde.proto.plan_serde
Docstring: Logical plan serialization/deserialization using singledispatch.
This module provides the main dispatch functions for plan serialization.
The actual serialization implementations are organized in the plans/ subdirectory.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
function
|
_serialize_logical_plan_helper
|
fenic.core._serde.proto.plan_serde._serialize_logical_plan_helper
|
Serialize a logical plan to a wrapper LogicalPlanProto (plan oneof set).
Plan-specific modules register implementations for this helper to return the
LogicalPlanProto with the appropriate oneof field set. Common fields (schema/cache_info)
are added by serialize_logical_plan().
|
site-packages/fenic/core/_serde/proto/plan_serde.py
| false | true | 21 | 35 | null |
LogicalPlanProto
|
[
"logical_plan",
"context"
] | null | null | null |
Type: function
Member Name: _serialize_logical_plan_helper
Qualified Name: fenic.core._serde.proto.plan_serde._serialize_logical_plan_helper
Docstring: Serialize a logical plan to a wrapper LogicalPlanProto (plan oneof set).
Plan-specific modules register implementations for this helper to return the
LogicalPlanProto with the appropriate oneof field set. Common fields (schema/cache_info)
are added by serialize_logical_plan().
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["logical_plan", "context"]
Returns: LogicalPlanProto
Parent Class: none
|
function
|
serialize_logical_plan
|
fenic.core._serde.proto.plan_serde.serialize_logical_plan
|
Serialize a logical plan to the wrapper LogicalPlanProto.
This calls the helper to get a wrapper with the correct oneof set, then adds
common fields like schema and cache_info.
|
site-packages/fenic/core/_serde/proto/plan_serde.py
| true | false | 38 | 58 | null |
LogicalPlanProto
|
[
"logical_plan",
"context"
] | null | null | null |
Type: function
Member Name: serialize_logical_plan
Qualified Name: fenic.core._serde.proto.plan_serde.serialize_logical_plan
Docstring: Serialize a logical plan to the wrapper LogicalPlanProto.
This calls the helper to get a wrapper with the correct oneof set, then adds
common fields like schema and cache_info.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["logical_plan", "context"]
Returns: LogicalPlanProto
Parent Class: none
|
function
|
deserialize_logical_plan
|
fenic.core._serde.proto.plan_serde.deserialize_logical_plan
|
Deserialize a logical plan from protobuf format.
This function determines which oneof field is set in the LogicalPlanProto
and delegates to the appropriate deserialization helper function.
Args:
logical_plan_proto: The protobuf representation to deserialize.
context: The serde context for error reporting and path tracking.
Returns:
LogicalPlan: The deserialized logical plan, or None if empty.
Raises:
DeserializationError: If the protobuf is invalid or deserialization fails.
|
site-packages/fenic/core/_serde/proto/plan_serde.py
| true | false | 61 | 99 | null |
Optional[LogicalPlan]
|
[
"logical_plan_proto",
"context"
] | null | null | null |
Type: function
Member Name: deserialize_logical_plan
Qualified Name: fenic.core._serde.proto.plan_serde.deserialize_logical_plan
Docstring: Deserialize a logical plan from protobuf format.
This function determines which oneof field is set in the LogicalPlanProto
and delegates to the appropriate deserialization helper function.
Args:
logical_plan_proto: The protobuf representation to deserialize.
context: The serde context for error reporting and path tracking.
Returns:
LogicalPlan: The deserialized logical plan, or None if empty.
Raises:
DeserializationError: If the protobuf is invalid or deserialization fails.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: ["logical_plan_proto", "context"]
Returns: Optional[LogicalPlan]
Parent Class: none
|
function
|
_deserialize_logical_plan_helper
|
fenic.core._serde.proto.plan_serde._deserialize_logical_plan_helper
|
Deserialize a logical plan.
|
site-packages/fenic/core/_serde/proto/plan_serde.py
| false | true | 102 | 113 | null |
Optional[LogicalPlan]
|
[
"underlying_proto",
"context",
"schema"
] | null | null | null |
Type: function
Member Name: _deserialize_logical_plan_helper
Qualified Name: fenic.core._serde.proto.plan_serde._deserialize_logical_plan_helper
Docstring: Deserialize a logical plan.
Value: none
Annotation: none
is Public? : false
is Private? : true
Parameters: ["underlying_proto", "context", "schema"]
Returns: Optional[LogicalPlan]
Parent Class: none
|
module
|
serde_context
|
fenic.core._serde.proto.serde_context
|
Helper utilities for path tracking in serde operations.
|
site-packages/fenic/core/_serde/proto/serde_context.py
| true | false | null | null | null | null | null | null | null | null |
Type: module
Member Name: serde_context
Qualified Name: fenic.core._serde.proto.serde_context
Docstring: Helper utilities for path tracking in serde operations.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
attribute
|
EnumType
|
fenic.core._serde.proto.serde_context.EnumType
| null |
site-packages/fenic/core/_serde/proto/serde_context.py
| true | false | 57 | 57 | null | null | null | null |
TypeVar('EnumType', bound=Enum)
| null |
Type: attribute
Member Name: EnumType
Qualified Name: fenic.core._serde.proto.serde_context.EnumType
Docstring: none
Value: TypeVar('EnumType', bound=Enum)
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
class
|
SerdeContext
|
fenic.core._serde.proto.serde_context.SerdeContext
|
Context for managing serialization/deserialization state and path tracking.
Provides centralized error handling, path tracking, and field-level serde operations
for protobuf serialization/deserialization. All serde operations should use this
context to ensure consistent error reporting and path information.
|
site-packages/fenic/core/_serde/proto/serde_context.py
| true | false | 60 | 953 | null | null | null | null | null |
[] |
Type: class
Member Name: SerdeContext
Qualified Name: fenic.core._serde.proto.serde_context.SerdeContext
Docstring: Context for managing serialization/deserialization state and path tracking.
Provides centralized error handling, path tracking, and field-level serde operations
for protobuf serialization/deserialization. All serde operations should use this
context to ensure consistent error reporting and path information.
Value: none
Annotation: none
is Public? : true
is Private? : false
Parameters: none
Returns: none
Parent Class: none
|
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.