#Modules for photoflow code#

This code and plugin seem much more at home in a Flow situation, where they will always be allowed to run their course. Although this could be worked out, if the workflow shown in the example app is really how you wanted to do things, but I didn't bother with that here. This could easily result in data arriving out of order, and any given page then displaying the wrong info. There these processes will not die when you leave the page, and will send their information back through the channel whenever they do complete. One issue not really addressed in the example app is the processing deadlock you can get into if you start looking at lots of pages without waiting for the color clustering to complete. The additional image processing actually added little time to the overall run time of the flow as the rate limited processes were still the deciding factor. %Result = item) doĬolor_three: Enum.at(an_data.dominant, 2) Proc_params = convert_params(opts) ++ Ĭase Porcelain.exec(executable_path(), proc_params, out: :iodata, async_in: true) do It then retrieves the result for us, and hopefully all thats left to do is strip off the length byte from the front of the results, and then let :erlang.binary_to_term converts that into a convenient Elixir Map for us. (Python module for ImageMagic) to devel task Add abr2gbr to photography. I just convert the parameters to a set of flags and Porcelain handles booting up the C instance and sending it the flags. remove WNPP info Photoflow now in Salsa - Add Vcs info And photoflow. So I turned to the Porcelain gem to help me get that rolling which made things very easy. It didn't have a large effect in the example app but was a big improvement for the PhotoFlow implementation. I found it better to boot an instance, process one request and then end the instance, this puts the threading control into the calling applications hands, if needed allowing letting several processes analyze images at once. I did orginally try starting a single instance of OpenCV code and sending messages back and forth, but the length of time the analyses took was a huge bottleneck and left me unable to process images concurrently. There are plenty of ways to use your own C code in Elixir, in this case the simplest and safest way turned out to be the best. Shown below is the display of a simple sample app (available in this github repository ). The ocv_photo_analyzer lets me generate histograms and find the dominant colors inside images through elixirs Port module. In this post I'll extend my elixir image analysis abilities with the C++ based OpenCV library, and the ocv_photo_analyzer wrapper package source code on Github | hex. For Ubuntu systems, Photoflow is available via some third party PPA, so installing it on your Ubuntu 15.04 Vivid Vervet, Ubuntu 14.04 Trusty Tahr and derivative systems like Linux Mint 17.2 Rafaela, Linux Mint 17.1 Rebecca, Linux Mint 17 Qiana, Pinguy OS 14.04, Elementary OS 0.3 Freya, Deepin 2014, Peppermint 6, Peppermint 5, LXLE 14.04 and.

But in that post the operations performed were all pretty simple. In my previous blog post I discussed using using Elixir's Flow module to process photos for uptake into a database.

Modules for photoflow