Skip to content

Extract textual information using the Apache Tika library from JSON streams

License

Notifications You must be signed in to change notification settings

sesam-community/content-extractor

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

24 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Content extraction microservice

A Java microservice for transforming a JSON entity stream. This service is designed to be used as a microservice system for the HTTP transform in a Sesam service instance.

Build Status

It decodes binary content or downloads files via URLs, and then extracts textual information using the Apache Tika library (http://tika.apache.org/). The resulting text is inserted as a string property into the entities and returned to the client.

You will need a working Java 8 dev environment and Maven 3.x to build this service.

Running in Docker

cd content-extraction-service
mvn clean install
docker build -t sesam/content-extraction-service:latest .
docker run -it -p 4567:4567 sesam/content-extraction-service:latest

Get the IP from docker:

docker inspect -f '{{.Name}} - {{.NetworkSettings.IPAddress }}' content-extraction-service

Example

JSON entities can be posted to 'http://localhost:4567/transform'. The result is streamed back to the client. Exchange "localhost" with the Docker IP if running in Docker.

$ curl -s -XPOST 'http://localhost:4567/transform' -H "Content-type: application/json" -d '[{ "_id": "jane", "url": "http://some-url-to-a-file/file.pdf"}]' | jq -S .
[
  {
    "_id": "jane",
    "url": "http://some-url-to-a-file/file.pdf",
    "_content": "The content in text form"
  }
]

Note the example uses curl to send the request and jq prettify the response.

Configuration

You can configure the service with the following environment variables:

Variable Description Default
SOURCE_PROPERTY The name of the property holding the source value. Note that the URL must be encoded. Supported URL schemes are: 'http' and 'https'. The values must be URLs (e.g. http://example.org/my.doc), Transit-encoded URLs (e.g. ~rhttp://example.org/my.doc) or Transit-encoded Base64-encoded bytes (e.g. ~bYWJj). "url"
TARGET_PROPERTY The name of the property that will hold the extracted content. "_content"
USERNAME Username for authentication needed to download the file represented by the SOURCE_PROPERTY. If not set or null, no authentication will be attempted. Not set
PASSWORD The password for the USERNAME. Only applicable if USERNAME is set. Not set
AUTH_TYPE The authentication method to use if USERNAME and PASSWORD is both set. This is an enum with the following valid values (case sensitive): "basic", "digest" and "ntlm". If "ntlm" is specified, you can provide additional information in WORKSTATION and/or DOMAIN. "basic"
WORKSTATION Used in conjuction with "ntlm" as AUTH_TYPE Not set
DOMAIN Used in conjuction with "ntlm" as AUTH_TYPE Not set
SOCKET_TIMEOUT TCP socket timeout value in seconds. "120"
CONNECTION_TIMEOUT Connection timout value in seconds. "10"
TRUST_EVERYTHING Disable TLS certificate and hostname verification. Note that this is insecure. "false"
THREADS The number of worker threads to use for content download and extraction. Note that increasing this value will also increase memory and CPU usage. If the value set increases resource usage beyond the capactity of the machine running the service the service process may be terminated. "8"
LOGLEVEL Sets the log level to use. All logs are written to stdout. Must be one of "trace", "debug", "info", "warn", or "error") "info"

When running in Docker you can either specify this in a file (see https://docs.docker.com/compose/env-file/) or on the command line with "docker run .. -e VAR1=VAL1 -e VAR2=VAL2 .."

About

Extract textual information using the Apache Tika library from JSON streams

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Java 98.9%
  • Shell 1.1%