This addon provides an action to extract information from images in Alfresco by using Google Vision API. This Google Service is not free but it's very cheap: please consider reading Google Vision API pricing before using this addon.
It includes following data on every image:
- Labels categorized by Google
- Logo text (if detected)
- Landmark text (if detected)
- Text (by using OCR)
- Safe Search information (adult, medical, spoof and violence)
License The plugin is licensed under the LGPL v3.0.
State Current addon release 1.0.0 is BETA
Compatibility The current version has been developed using Alfresco 5.1 and Alfresco SDK 2.2.0, although it should run in Alfresco 5.0.d and Alfresco 5.0.c
Browser compatibility: 100% supported
Languages Currently only provided in English but it supports (at least) Spanish.
No original Alfresco resources have been overwritten
The binary distribution is made of one AMP file to be deployed in Alfresco and one AMP file to be deployed in Share:
You can install them by using standard Alfresco deployment tools
If you are new to Alfresco and the Alfresco Maven SDK, you should start by reading Jeff Potts' tutorial on the subject.
You can build the artifacts from source code using maven
$ mvn clean package
After installation, following properties must be included in alfresco-global.properties
# Google vision
google.vision.application.name=keensoft-Alfresco/1.0
# Credentials path
google.vision.credentials.json=/tmp/key.json
google.vision.max.results=3
# Google translate (empty value for non translation)
google.translate.language=es
google.translate.api.key=XXXXXXXXXXXXXXXXX
# Safe search configuration, possible values:
# UNKNOWN, VERY_UNLIKELY, UNLIKELY, POSSIBLE, LIKELY, VERY_LIKELY
google.safe.search.adult=LIKELY
google.safe.search.medical=
google.safe.search.spoof=
google.safe.search.violence=POSSIBLE
# Background process
google.pool.core.size=1
google.pool.maximum.size=1
google.pool.thread.priority=5
- Including a rule on a folder by selecting google-vision-action
- Every dropped image on this folder will be sent to Google Vision in order to get labels as Alfresco Tags, landmark and logo as Alfresco Custom Aspect and OCR text as Alfresco Description.
- If a language is selected for Google Translate in alfresco-global.properties, labels, landmark and text is translated before including values in Alfresco.
- This operation is performed asynchronously, so results are not available just after uploading the image.
Including missing functionalities from Google Vision API
Build a new metadata extractor including Google Vision labels (idea provided by Ray Gauss)
Following features should be also included:
- Provide a new Share action with params:
- Number of results
- Language selection
- Detecting explicit content
- Detecting OCR
- Retrieving image attributes
- Face detection
- Develop a repo Service for Google Vision