Starters¶

The pulse.starters module provides convenient helpers for common tasks. Each function accepts either a list of strings or a path to a text/CSV/TSV/Excel file where the first column is treated as the text input.

Supported file types: .txt, .csv, .tsv, .xls, .xlsx.

Async Support¶

All starter functions have async equivalents in the pulse.async_starters module. Simply add _async to the function name and use await:

import asyncio
from pulse.async_starters import sentiment_analysis_async, generate_themes_async

async def main():
    # Async versions of all starter functions
    sentiment_result = await sentiment_analysis_async(["great", "bad"])
    themes_result = await generate_themes_async(["food good", "service slow"])

    # Process multiple operations concurrently
    tasks = [
        sentiment_analysis_async(["text1", "text2"]),
        generate_themes_async(["text3", "text4"])
    ]
    results = await asyncio.gather(*tasks)

asyncio.run(main())

For detailed async patterns, see the Async Patterns Guide.

Loading Inputs¶

get_strings(source: list[str] | str) -> list[str] loads strings from a list or file path.

Sentiment¶

from pulse.starters import sentiment_analysis

resp = sentiment_analysis(["great", "bad"])  # or a file path
print([r.sentiment for r in resp.results])

Parameters: - input_data: list[str] | str - version: str | None = None - await_job_result: bool = True – If False returns a Job. - auth: pulse.auth._BaseOAuth2Auth | None = None – Optional auth. - client: CoreClient | None = None – Existing client; else one is created.

Notes: - Automatically sets fast=True when len(texts) <= 200.

Theme Allocation¶

from pulse.starters import theme_allocation

res = theme_allocation(["food good, service slow"], themes=["Food", "Service"])
print(res.assign_single())

Parameters: - input_data: list[str] | str - themes: list[str] | None – If None, the helper runs a ThemeGeneration step automatically. - auth: pulse.auth._BaseOAuth2Auth | None = None - client: CoreClient | None = None

Returns: ThemeAllocationResult with helpers like assign_single(), assign_multi(), heatmap(), and to_dataframe().

Clustering¶

from pulse.starters import cluster_analysis

resp = cluster_analysis(["good", "bad", "ok"], k=2)
print(resp.clusters)

Parameters: - input_data: list[str] | str - k: int – Number of clusters. - algorithm: str | None = None - await_job_result: bool = True - auth: pulse.auth._BaseOAuth2Auth | None = None - client: CoreClient | None = None

Summaries¶

from pulse.starters import summarize

resp = summarize(["Great food, slow service"], question="What do diners say?", length="short")
print(resp.summary)

Parameters: - input_data: list[str] | str - question: str - length: str | None = None – One of bullet-points, short, medium, long. - preset: str | None = None – One of five-point, ten-point, one-tweet, three-tweets, one-para, exec, two-pager, one-pager. - await_job_result: bool = True - auth: pulse.auth._BaseOAuth2Auth | None = None - client: CoreClient | None = None