replaCy
replaCy copied to clipboard
spaCy match and replace, maintaining conjugation
Bumps [certifi](https://github.com/certifi/python-certifi) from 2021.10.8 to 2022.12.7. Commits 9e9e840 2022.12.07 b81bdb2 2022.09.24 939a28f 2022.09.14 aca828a 2022.06.15.2 de0eae1 Only use importlib.resources's new files() / Traversable API on Python ≥3.11 ... b8eb5e9 2022.06.15.1...
In a project using replaCy, when we autoload the match_dict, it loads `./resources/match_dict.json` .... in replaCy ie site-packages Need to load relative to the file instantiating replaCy, or remove smart-loading
KenLMScorer is fantastic. Just so useful. However, it isn't core to replaCy and should be a custom pipeline component (that we expect most people to use... think like `en_core_web_sm` is...
Hi, From the readme docs it seems clear that verb substitution with presented view is preserved. Is it possible to take an input text, and alter point of view? E.G....
Once #85 is done, we should build an alternate scorer that uses MLM's like BERT. [this code](https://github.com/awslabs/mlm-scoring) exposes a perplexity-like score for MLMs, it really seems not bad. This wouldn't...
I think this should be done before #3. [Poetry](https://python-poetry.org/) has provided a standardized way to manage dependencies, build, and publish. Finally... it only took python like 30 years to figure...
replaCy extensions can be added... but not removed or renamed. We also don't constrain component names to be unique. So to match the spaCy API (and just have reasonable functionality),...
Honestly not sure if this is a great idea... We'd want to duplicate the match `len(dict.items())` times, one for each entry. An example usage would be: (`resources/variables/lookups.json`) ```json { "opposites":...
In our newest project we are using a wrapped version of replacy to support list variables in match_dicts, like so ```python import json import os from typing import List from...
[spaczz](https://github.com/gandersen101/spaczz) is a nice fuzzy matching library for spaCy, and could be a nice addition to replaCy. We should consider if we want to solve the general problem of "allow...