Releases: imgflo/imgflo
imgflo 0.2.0
server
Moved to separate git repository jonnor/imgflo-server
runtime
Registration as Flowhub.io runtime now done in main imgflo
executable,
use environment variables FLOWHUB_USER_ID
and IMGFLO_RUNTIME_ID
to specify.
Support for running on Heroku out-of-the-box, just set above envvars
and push git repo to an heroku app and connect from Flowhub.
Added annotations for many more port types; including number, booleans, enums, colors.
This lets Flowhub bring up more suitable UIs than the general string input.
Added support for default values for IIPs.
Multiple graphs are now supported, allowing to switch
between graphs in a Flowhub project without wierd issues.
Fix an issue with Flowhub 0.2.0+ due to not setting runtime capabilities correctly.
The Processor
component now checks that bounds have a sane value,
and disables processing if 0x0 or clips if above max_size
.
NB: Currently max_size limit not configurable.
Default GEGL build has been updated to recent git master version,
and now includes the 'workshop' operations.
imgflo 0.1.0
imgflo now consists of two complimentary parts:
a Flowhub-compatible runtime for interactively building image processing graphs,
and an image processing server for image processing on the web.
The runtime combined with the Flowhub IDE allows to visually create image
processing pipelines in a node-based manner, similar to tools like the Blender node compositor.
Live image output is displayed in the preview area in Flowhub, and will
update automatically when changing the graph.
The server provides a HTTP API for processing an input image with an imgflo graph.
GET /graph/mygraph?input=urlencode(http://example.com/input-image.jpeg)¶m1=foo¶m2=bar
The input and processed image result will be cached on disk.
On later requests to the same URL, the image will be served from cache.
The server can be deployed to Heroku with zero setup, just push the git repository to an Heroku app.
The operations used in imgflo are provided by GEGL, and new operations can be added using the C API.
A (somewhat outdated) list of operations can be seen here: http://gegl.org/operations.html
Blogpost: http://www.jonnor.com/2014/04/imgflo-0-1-an-image-processing-server-and-flowhub-runtime