API Reference

tesseract-streamlit.parse

Parsing module for tesseract-streamlit.

Utilities for parsing and formatting inputs to be injected into the Streamlit template.

Routines here handle both the OpenAPI Specification generated for a Tesseract, and functions passed into the user-defined functions module. Each are then parsed and formatted into string descriptions, and structured with lists and dictionaries, for easy injection in the Streamlit Jinja template.

The object to be passed to the Jinja template is an instance of TemplateData, which may be created using the only entry-point to this module: extract_template_data().

All other members are private or TypedDict definitions.

class tesseract_streamlit.parse.FuncDescription

Represents a brief summary of a Python function.

Stores the function’s name, the first line of its docstring as the title, and the remaining docstring content as documentation.

This is used in the Streamlit template to inject the functions into the app, and annotate their plot outputs with the title and docs the user provides.

backend identifies which plotting library is supported. “builtin” indicates that support is provided by Streamlit natively.

backend: Literal['builtin', 'pyvista']

docs: str

name: str

title: str

class tesseract_streamlit.parse.JinjaField

Input field schema preprocessed for Jinja template compatibilityping.

Replaces “ancestors” from _InputField with “parent_container”, “container”, “uid”, “stem”, and “key”.

container: str

default: NotRequired[Any]

description: str

key: str

parent_container: str

stem: str

title: str

type: str

uid: str

class tesseract_streamlit.parse.TemplateData

Schema for passing data directly to Jinja template for injection.

The following key value pairs are:

metadata: background information on the Tesseract instance. schema: structure of the input form. url: URI of the running Tesseract instance. needs_pyvista: whether to enable PyVista support for the UDFs. udf_defs: source code for the UDF definitions. udfs: register of the UDFs, sorted by their input parameters.

metadata: TesseractMetadata

needs_pyvista: bool

schema: list[JinjaField]

udf_defs: NotRequired[str]

udfs: NotRequired[UdfRegister]

url: str

class tesseract_streamlit.parse.TesseractMetadata

Basic info about the Tesseract being interfaced.

Title, description, and version information to be injected into the header of the Streamlit template.

description: str

title: str

version: str

class tesseract_streamlit.parse.UdfRegister

Stores the user-defined function descriptions as a register.

Each kind of function is stored in its own list, based on what parameter it takes. If it takes Tesseract inputs as a parameter, it is stored in the list under the ‘inputs’ key. Likewise for Tesseract outputs. If it takes both inputs and outputs, it’s stored under ‘both’.

This register is passed for injection into the Streamlit app template. Storing the functions by the parameters they take enables them to be passed the correct parameters in the generated app.

both: list[FuncDescription]

inputs: list[FuncDescription]

outputs: list[FuncDescription]

exception tesseract_streamlit.parse.UserDefinedFunctionError(message)

Exception for invalid user defined functions.

exception tesseract_streamlit.parse.UserDefinedFunctionWarning(message)

Warning for user defined functions.

tesseract_streamlit.parse.extract_template_data(url, user_code, pretty_headings)

Formats Tesseract and user-defined function inputs for template.

Retrieves and parses the OpenAPI specification of the running Tesseract. User code is also analysed. These inputs are then formatted and pre-processed according to the TemplateData schema to provide all data required to statically render the Streamlit app.

Parameters:

• url (str) – URI of the running Tesseract instance.

• user_code (Path | None) – path of the user-defined plotting function module.

• pretty_headings (bool) – whether to format parameters names as headings.

Returns:

Preprocessed data describing the Streamlit app based on the InputSchema, ready for injection into the app template.

Return type:

TemplateData