temporal-text 0.1.0

Estimate how time-bound text is: 0.0 (ephemeral) to 1.0 (timeless facts)

temporal-text

Estimate how time-bound text is, from ephemeral thoughts to timeless facts.

What is Temporality?

Temporality measures how "time-bound" a piece of text is:

• Low temporality (0.0): Fleeting, immediate, ephemeral

  • "I want to go to the sauna" → ~0.15
  • "Just saw the funniest thing" → ~0.10
  • "brb getting snacks" → ~0.05

• High temporality (1.0): Timeless facts, universal truths

  • "Earth has one moon" → ~0.85
  • "A triangle has three sides" → ~0.85
  • "Water boils at 100 degrees Celsius" → ~0.85

Installation

pip install temporal-text

Usage

from temporal_text import temporality_score

# Low temporality - fleeting desire
score = temporality_score("I want to go to the sauna")
print(score)  # ~0.15

# High temporality - scientific fact
score = temporality_score("Earth has one moon")
print(score)  # ~0.85

# Medium temporality - current situation
score = temporality_score("I'm currently learning Spanish")
print(score)  # ~0.35

How It Works

The library uses a lightweight TF-IDF + Ridge Regression model trained on ~1700 labeled examples. The model learns patterns like:

High temporality indicators: "is a", "always", "contains", "equals", "measures"

Low temporality indicators: "I", "my", "feel", "just", "tomorrow", "now"

The pre-trained model (~110KB) is bundled with the package - no downloads or API calls required.

License

MIT