README

Name

ChupaText HTTP server

Description

ChupaText HTTP server is a HTTP interface for ChupaText, an extensible text extractor. You can use ChupaText via HTTP.

Overview

You can run ChupaText HTTP server by the following command lines:

% bundle install
% bin/rails server

ChupaText HTTP server supports the following formats by default:

• CSV (.csv)
• Office Open XML:
  • Document:
    • .docx
    • .docm
    • .dotx
    • .dotm
  • Presentation:
    • .pptx
    • .pptm
    • .ppsx
    • .ppsm
    • .potx
    • .potm
    • .sldx
    • .sldm
  • Workbook:
    • .xlsx
    • .xlsm
    • .xltx
    • .xltm
• OpenDocument:
  • Presentation (.odp)
  • Spreadsheet (.ods)
  • Text (.odt)
• XML (.xml)

You can enable more formats by adding ChupaText decomposer plugins to Gemfile.local. You can find ChupaText decomposer plugins at https://rubygems.org/search?query=chupa-text-decomposer- . See also Gemfile.local.example.

For example, you can use chupa-text-decomposer-pdf to add support for PDF:

# Gemfile.local
gem "chupa-text-decomposer-pdf"

Note that you need to run bundle install again when you change Gemfile.local:

$ bundle install

You can use the ChupaText HTTP server by the following command line:

% curl --form [email protected] http://127.0.0.1:3000/extraction.json

ChupaText HTTP server returns the following JSON (formatted):

{
  "mime-type": "application/pdf",
  "uri": "file:///tmp/hello.pdf",
  "path": "/tmp/hello20190328-10015-hcaahd.pdf",
  "size": 6567,
  "texts": [
    {
      "mime-type": "text/plain",
      "uri": "file:///tmp/hello.txt",
      "path": "/tmp/hello.txt",
      "size": 6,
      "created_time": "2014-01-05T06:35:43.000Z",
      "source-mime-types": [
        "application/pdf"
      ],
      "creator": "Writer",
      "producer": "LibreOffice 4.1",
      "body": "Page1\n",
      "screenshot": {
        "mime-type": "image/png",
        "data": "iVBORw0KGgoAAAANSUhEUgAAAMgAAADICAIAAAAiOjnJAAAABmJLR0QA/wD/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\n/gFl5TA2XANYHwAAAABJRU5ErkJggg==\n",
        "encoding": "base64"
      }
    }
  ]
}

You can specify the URL to be extracted:

% curl \
  --data uri=https://github.com/ranguba/chupa-text-http-server/raw/master/test/fixtures/files/hello.docx \
  http://127.0.0.1:3000/extraction.json

ChupaText HTTP server returns the following JSON (formatted):

{
  "mime-type": "application/octet-stream",
  "uri": "https://github.com/ranguba/chupa-text-http-server/raw/master/test/fixtures/files/hello.docx",
  "size": 4110,
  "texts": [
    {
      "mime-type": "text/plain",
      "uri": "https://github.com/ranguba/chupa-text-http-server/raw/master/test/fixtures/files/hello.txt",
      "size": 7,
      "title": "Hello",
      "created_time": "2017-07-10T18:00:16.000Z",
      "modified_time": "2017-07-10T18:00:43.000Z",
      "source-mime-types": [
        "application/octet-stream"
      ],
      "application": "LibreOffice/5.3.4.2$Linux_X86_64 LibreOffice_project/30m0$Build-2",
      "body": "World!\n"
    }
  ]
}

You can use ChupaText HTTP server as container. See also:

• chupa-text-docker

• chupa-text-vagrant

API

The end point URL is http://#{HOST}:#{PORT}/extraction.json.

You must specify data or uri parameter. If you want to extract text from local file, you can use data parameter. If you want to extract text from remote file, you can use uri parameter.

If you use data, you must send data as multipart/form-data or application/x-www-form-urlencoded.

If you use multipart/form-data, you must specify filename of the data too. You can also specify content-type of the data too.

If you use application/x-www-form-urlencoded, you must specify mime_type parameter too. It's MIME type of the data.

Metadata are helpful to choose correct decomposer. Decomposer is a ChupaText module to extract text from the specific data.

Here are optional parameters:

• timeout: The max seconds to extract texts from the given data.

• limit_cpu: The max amount of CPU time in seconds of external process. ChupaText may use external command such as libreoffice to extract text from data.

• limit_as: The max size of the virtual memory of external process. ChupaText may use external command such as libreoffice to extract text from data.

• max_body_size: The max body size to be extracted.

Authors

• Kouhei Sutou <[email protected]>

• Shimadzu Corporation

License

GPL 3 or later.

(Kouhei Sutou has a right to change the license including contributed patches.)