Spellcheck microservice

This is a microservice designed to check the spelling of words. Based on pyenchant. Exposes a REST API.
Current available languages are: ru_RU, en_US, es_ES, fr_FR, de_DE, pt_PT.
It runs blazingly fast due to the use of pychant in its kernel, LRU cache usage and pypy.
Also it supports feature called «user dictionaries» — user can add his own word-exceptions to personal dictionary.

Quickstart

• docker run -p 10113:10113 -t --mount source=spellcheck-dicts,target=/data/ xfenix/spellcheck-microservice:latest
• check http://localhost:10113/docs/ for full REST documentation
• main REST endpoint you will be needed is http://localhost:10113/api/check/ (this will be available without authorization)

Configuration

Config options

You can change config of the service by changing the environment variables. Here is a list of them:

• SPELLCHECK_SENTRY_DSN Sentry DSN for integration. Empty field disables integration. Default value is empty string.
• SPELLCHECK_API_KEY define api key for users dictionaries mostly. Please, provide, if you want to enable user dictionaries API. Default value is empty string.
• SPELLCHECK_ENABLE_CORS enable CORS for all endpoints. In docker container this option is disabled. Default value is True.
• SPELLCHECK_STRUCTURED_LOGGING enables structured (json) logging. Default value is True.
• SPELLCHECK_WORKERS define application server workers count. If you plan to use k8s and only scale with replica sets, you might want to reduce this value to 1. Default value is 8. Restrictions: Gt(gt=0), Lt(lt=301)
• SPELLCHECK_SERVER_ADDRESS binding address, default value suitable for docker. Default value is 0.0.0.0.
• SPELLCHECK_PORT binding port. Default value is 10113. Restrictions: Gt(gt=1023), Lt(lt=65536)
• SPELLCHECK_CACHE_SIZE define LRU cache size for misspelled word/suggestions cache. Any value less than 1 makes the cache size unlimited, so be careful with this option. Default value is 10000.
• SPELLCHECK_API_PREFIX define all API's URL prefix. Default value is /api/.
• SPELLCHECK_DOCS_URL define documentation (swagger) URL prefix. Default value is /docs/.
• SPELLCHECK_MAX_SUGGESTIONS defines how many maximum suggestions for each word will be available. 0 means unlimitied. Default value is 0. Restrictions: Ge(ge=0)
• SPELLCHECK_DICTIONARIES_PATH define directory where user dicts is stored. This is inner directory in the docker image, please map it to volume as it shown in the quickstart part of this readme. Default value is /data.
• SPELLCHECK_DICTIONARIES_STORAGE_PROVIDER define wich engine will store user dictionaries. Default value is StorageProviders.FILE.
• SPELLCHECK_DICTIONARIES_DISABLED switches off user dictionaries API no matter what. Default value is False.
• SPELLCHECK_USERNAME_MIN_LENGTH minimum length of username. Default value is 3.
• SPELLCHECK_USERNAME_MAX_LENGTH maximum length of username. Default value is 60.
• SPELLCHECK_EXCLUSION_WORDS_STR String with list of words which will be ignored in /api/check endpoint each request. Example: 'foo, bar'. Default value is empty string.

Deployment

Note: all docker & docker-compose variants use named volumes to store user dictionaries.

Plain docker

docker run -p 10113:10113 -t --mount source=spellcheck-dicts,target=/data/ xfenix/spellcheck-microservice:latest

Docker-compose

• Save this example configuration as docker-compose.yml:

version: "3.9"
services:
  spellcheck:
    image: xfenix/spellcheck-microservice:latest
    ports:
      - "10113:10113"
    volumes:
      - spellcheck-dicts:/data/

volumes:
  spellcheck-dicts:

• Then run docker-compose up

Changelog

You cand find it here https://github.com/xfenix/spellcheck-microservice/releases

Development

Quickstart

• Clone this repo
• For MacOS X brew install enchant
• For Debian/Ubuntu apt-get install -y enchant-2 hunspell-ru
• uv sync --group dev
• source .venv/bin/activate
• Execute make command to run local development server
• README is automatically updated in the CI pipeline for each commit

Notes

Default api-key for local development is debug (you will need this to work with user dictionaries API).

Please check ./Makefile for more details

Troubleshooting

For MacOS X on Apple Silicon add PYENCHANT_LIBRARY_PATH=/opt/homebrew/lib/libenchant-2.dylib to your .zprofile