nlp-cache-sebastian-stigler

NLP Cache

A caching component for Doc classes in spacy.

Installation

pip install nlp_cache_sebastian_stigler

Usage

import spacy

nlp = spacy.load("de_core_news_sm")
text = "Ich wünsche Ihnen einen guten Morgen!"
doc_1 = nlp(text)  # mean loading time about 4.5 ms
doc_2 = nlp(text)  # recurrent mean loading time about 4.5 ms

To speed up loading a recurrent text you can use this caching utility, to lookup known texts from the cache stored in the cache directory (relative to the current directory).

import spacy
from nlp_cache import NlpCache

nlp = NlpCache(uncached_nlp=spacy.load("de_core_news_sm"), path="cache")
text = "Ich wünsche Ihnen einen guten Morgen!"
doc_1 = nlp(text)  # mean loading time about 4.5 ms
doc_2 = nlp(text)  # recurrent mean loading time about 840 µs