This docker container runs spaCy behind NodeJS API allowing clients to communicate with spaCy via HTTP requests.

Prebuilt Docker images are available at https://hub.docker.com/r/jlundan/spacy-nodejs-alpine/

Supported tags and Dockerfile links

• 1.0.0, latest (1.0.0/Dockerfile) - spaCy version 1.9.0

Examples

Running

docker run -p 5000:3000 jlundan/spacy-nodejs-alpine

Get some info about the API

curl http://localhost:5000

You should receive following kind of response

{
  "title": "spaCy JSON service",
  "spacyInfo": {
    "name": "spacy",
    "version": "1.9.0",
    "website": "https://spacy.io/"
  }
}

Processing a sentence

curl -X POST --data '{"input": "This is a test."}' \
    -H "Content-Type: application/json" http://localhost:5000

You should receive following response

[{
  "text": "This is a test.",
  "len": 5,
  "tokens": [
    "This",
    "is",
    "a",
    "test",
    "."
  ],
  "noun_phrases": [
    "a test"
  ],
  "parse_tree": [
    {
      "word": "is",
      "lemma": "be",
      "NE": "",
      "POS_fine": "VBZ",
      "POS_coarse": "VERB",
      "arc": "ROOT",
      "modifiers": [
        {
          "word": "This",
          "lemma": "this",
          "NE": "",
          "POS_fine": "DT",
          "POS_coarse": "DET",
          "arc": "nsubj",
          "modifiers": []
        },
        {
          "word": "test",
          "lemma": "test",
          "NE": "",
          "POS_fine": "NN",
          "POS_coarse": "NOUN",
          "arc": "attr",
          "modifiers": [
            {
              "word": "a",
              "lemma": "a",
              "NE": "",
              "POS_fine": "DT",
              "POS_coarse": "DET",
              "arc": "det",
              "modifiers": []
            }
          ]
        },
        {
          "word": ".",
          "lemma": ".",
          "NE": "",
          "POS_fine": ".",
          "POS_coarse": "PUNCT",
          "arc": "punct",
          "modifiers": []
        }
      ]
    }
  ],
  "parse_list": [
    {
      "word": "This",
      "lemma": "this",
      "NE": "",
      "POS_fine": "DT",
      "POS_coarse": "DET"
    },
    {
      "word": "is",
      "lemma": "be",
      "NE": "",
      "POS_fine": "VBZ",
      "POS_coarse": "VERB"
    },
    {
      "word": "a",
      "lemma": "a",
      "NE": "",
      "POS_fine": "DT",
      "POS_coarse": "DET"
    },
    {
      "word": "test",
      "lemma": "test",
      "NE": "",
      "POS_fine": "NN",
      "POS_coarse": "NOUN"
    },
    {
      "word": ".",
      "lemma": ".",
      "NE": "",
      "POS_fine": ".",
      "POS_coarse": "PUNCT"
    }
  ]
}]

Processing multiple sentences

curl -X POST --data '{"input": ["This is a test.", "This is another test, it has \"quotes\" in it."]}' \
    -H "Content-Type: application/json" http://localhost:5000

The response is similar than with processing single sentence. Responses are returned for each sentence in the input array in the same order input sentences were given.

Building

docker build -t [tag]:[version] --build-arg SPACY_VERSION=[spacy version] .