Monday 9/2

Spent the day starting to put the image manager into JSP. Nobody's here since it's Labor Day, so most of it has to wait until tomorrow.

To do:
  1. Ask Chris for a new copy of the image database, change location in code, and get it to work.
  2. Talk to Jim about adding images to the database, and about new server issues.
  3. Keep putting image manager into JSP.


Tuesday 9/3

To run Rati's program (from sulcus): java -classpath .:/home/ratis/mm.mysql-2.0.4-bin.jar file1 Chris says there isn't really a reason to make a new copy of the database, so I'll just use the one that's there. I had to change the classpath that Tomcat was using to include that jar file in the JSP version, and change 'localhost' to 'sulcus.biostr.washington.edu' in the code. The connections are being made and the interface seems to work, but no pictures are coming up yet.

To add images to the database, the program would have to generate an IML file. I may or may not have time to do this. I'll put it at the end of the list, and if I get to it, I get to it. Jim says: "One way would be to create a unique color for each distinct object in the scene (ie each primitive), render the scene with these colors, then do image processing to find the borders of all these rendered objects. Linda is of course an expert at this, but basically, for each color (which is now uniquely associated with a named object), find the connected components for that color, find the border of each connected component - and write these borders out in the IML file."

This kind of coding has already been done elsewhere, except giving each primitive a distinct color.

To do:
  1. Check error messages in Action to get first image up there.
  2. Finish JSP version and put in FME.


Wednesday 9/4

It was close to working, except that the images were coming from this location:

/usr/people/ratis/var/repos/image_repo/fsa/default

...and they can't actually be displayed unless they're in a public_html folder. So I asked Rati if that folder could be moved. In the meantime, I worked out the FME interface and got a little more done on the website. All that's left is to make sure the images and captions get displayed right.

The server seemed to be working this morning, but it isn't now. It freezes whenever I ask a real question (other than get-api or quit). I asked JJ about it again.

The report: basically I should write a summary of what I've been doing, with pictures and descriptions. I should leave them a version in both html and word form. Later, they can use it to write a journal paper.

To do:
  1. Get image folders moved and fix the details.
  2. Investigate FM stuff.


Thursday 9/5

Until JJ or Rati gets back to me, I can't do much. (This seems to be a common problem.) I worked on the website and thought of report topics, in the meantime.

Stuff to write about: Rati says that the folder can't really be moved, because another program interacts with it directly. So... I hacked around the problem again with a cgi script, moveImg.cgi, which just copies the image from her folder to my public_html/manager folder. So the page can get the image. The next/previous buttons work somewhat, but there are still some bugs.

To do:
  1. Finish fixing bugs.
  2. Start writing the paper.


Friday 9/6

As far as I can tell, the JSP image manager now works. There are still some strange things that happen, but they're caused by things deep in the image-processing part, and I don't think it's worthwhile for me to try to figure them out now.

I started writing the report this afternoon. The FM server finally went up this afternoon, so there's a chance to work on it now, too.

To do:
  1. Finish report.
  2. Investigate FM server.