Skip to main content

ReplaCy = spaCy Matcher + pyInflect. Create rules, correct sentences.

Project description

replaCy: match & replace with spaCy

We found that in multiple projects we had duplicate code for using spaCy’s blazing fast matcher to do the same thing: Match-Replace-Grammaticalize. So we wrote replaCy!

  • Match - spaCy’s matcher is great, and lets you match on text, shape, POS, dependency parse, and other features. We extended this with “match hooks”, predicates that get used in the callback function to further refine a match.
  • Replace - Not built into spaCy’s matcher syntax, but easily added. You often want to replace a matched word with some other term.
  • Grammaticalize - If you match on ”LEMMA”: “dance”, and replace with suggestions: ["sing"], but the actual match is danced, you need to conjugate “sing” appropriately. This is the “killer feature” of replaCy

spaCy pypi Version Code style: black

Requirements

  • spacy >= 2.0 (not installed by default, but replaCy needs to be instantiated with an nlp object)

Installation

pip install replacy

Quick start

from replacy import ReplaceMatcher
from replacy.db import load_json
import spacy


match_dict = load_json('/path/to/your/match/dict.json')
# load nlp spacy model of your choice
nlp = spacy.load("en_core_web_sm")

rmatcher = ReplaceMatcher(nlp, match_dict=match_dict)

# get inflected suggestions
# look up the first suggestion
span = rmatcher("She extracts revenge.")[0]
span._.suggestions
# >>> ['exacts']

Input

ReplaceMatcher accepts both text and spaCy doc.

# text is ok
span = r_matcher("She extracts revenge.")[0]

# doc is ok too
doc = nlp("She extracts revenge.")
span = r_matcher(doc)[0]

match_dict.json format

Here is a minimal match_dict.json:

{
  "extract-revenge": {
    "patterns": [
      {
        "LEMMA": "extract",
        "TEMPLATE_ID": 1
      }
    ],
    "suggestions": [
      [
        {
          "TEXT": "exact",
          "FROM_TEMPLATE_ID": 1
        }
      ]
    ],
    "match_hook": [
      {
        "name": "succeeded_by_phrase",
        "args": "revenge",
        "match_if_predicate_is": true
      }
    ],
    "test": {
      "positive": [
        "And at the same time extract revenge on those he so despises?",
        "Watch as Tampa Bay extracts revenge against his former Los Angeles Rams team."
      ],
      "negative": ["Mother flavours her custards with lemon extract."]
    }
  }
}

For more information how to compose match_dict see our wiki:

Citing

If you use replaCy in your research, please cite with the following BibText

@misc{havens2019replacy,
    title  = {SpaCy match and replace, maintaining conjugation},
    author = {Sam Havens, Aneta Stal, and Manhal Daaboul},
    url    = {https://github.com/Qordobacode/replaCy},
    year   = {2019}
}

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

replacy-3.6.1.tar.gz (27.4 kB view hashes)

Uploaded source

Built Distribution

replacy-3.6.1-py3-none-any.whl (30.1 kB view hashes)

Uploaded py3

Supported by

AWS AWS Cloud computing Datadog Datadog Monitoring Facebook / Instagram Facebook / Instagram PSF Sponsor Fastly Fastly CDN Google Google Object Storage and Download Analytics Huawei Huawei PSF Sponsor Microsoft Microsoft PSF Sponsor NVIDIA NVIDIA PSF Sponsor Pingdom Pingdom Monitoring Salesforce Salesforce PSF Sponsor Sentry Sentry Error logging StatusPage StatusPage Status page