search

fiji / Colocalisation_Analysis

Fiji's plugin for colocalization analysis
http://imagej.net/Coloc_2
GNU General Public License v3.0
25 stars 18 forks source link

Results: results appear in different order in SingleWindowDisplay output and log window #25

Closed chalkie666 closed 7 years ago

chalkie666 commented 9 years ago

eg the result text threshold regression bisection appears first (log) or in the middle (PDF and SingleWindowDisplay) in the order of results.... order should be the same in every results output case.

(Also warnings in the PDF output appear in the middle, instead of the top of the list... See #53)

etadobson commented 7 years ago

It seems that in the log window... 3 extra, redundant lines are printed at the beginning as compared to the results window, which include:

  1. Coloc_Job_Name
  2. Mask Type Used
  3. Threshold regression

Otherwise, the order along with the results & log windows are the same. And the only difference between these and the PDF appears to be the warnings being inserted between 'Mask ID Used' and 'm (slope)'. Is this correct @chalkie666 ?

chalkie666 commented 7 years ago

@etarena yes we should stop the redundant lines from being output into the log. I expect that was my fault. It's something to do with some part of the code being run twice for some reason I could not figure out.

etadobson commented 7 years ago

@chalkie666 ok. i'll work on this one next... but might split it out from the PDF warning issue. I might file that one separately and work on this piecewise. is that ok?

chalkie666 commented 7 years ago

Sure!

On 10 Jan 2017 22:09, "Ellen T Arena" notifications@github.com wrote:

@chalkie666 https://github.com/chalkie666 ok. i'll work on this one next... but might split it out from the PDF warning issue. I might file that one separately and work on this piecewise. is that ok?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fiji/Colocalisation_Analysis/issues/25#issuecomment-271698626, or mute the thread https://github.com/notifications/unsubscribe-auth/AAP2MhXz7XdmpPFFqxtnHX91DmtuS7nuks5rQ_NygaJpZM4Fl6Ad .

etadobson commented 7 years ago

@chalkie666 - it seems this exception of 3 extra, redundant lines all call the handleValue(String name, String value) method... they are the only values that are represented by two Strings (all the other guys have numerical values as well). still investigating ... but seems they get printed in the log window before being printed from 'valueResults' List - in line 344 in SingleWindowDisplay.

chalkie666 commented 7 years ago

@etarena yes it might be that. I remember trying to find where things were being done twice but didn't find it. The complexity of the Plugin design doesn't help here. Johannes suggested to rip out some or all of the results handler and data container complexity and use builtin imagej gadgets... And I agree. For the reasons you now face.

etadobson commented 7 years ago

@chalkie666 good news... i found where in the code these extra lines are being printed to the log window... in Coloc_2 class... lines 490-494: constitutes a loop that 'let the algorithms feed their results to the handler' ... so - we have progress!

chalkie666 commented 7 years ago

@etarena Good catch!

On 18 January 2017 at 20:11, Ellen T Arena notifications@github.com wrote:

@chalkie666 https://github.com/chalkie666 good news... i found where in the code these extra lines are being printed to the log window... in Coloc_2 class... lines 490-494: constitutes a loop that 'let the algorithms feed their results to the handler' ... so - we have progress!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/fiji/Colocalisation_Analysis/issues/25#issuecomment-273570706, or mute the thread https://github.com/notifications/unsubscribe-auth/AAP2Mt92ohe3jA4tSo0_pxt2Uw7Kwow5ks5rTmPngaJpZM4Fl6Ad .

etadobson commented 7 years ago

@chalkie666 - I think I found it... It seems to be in the PDFWriter class... line 118 - there is code there to "send (output parameter name, value) to IJ.log for scraping batch results". If I comment out this line of code IJ.log(name + ", " + value);, the extra log printing is no longer an issue and 'solves' the problem without any apparent, adverse effects... just let me know your thoughts on this when you are able.

chalkie666 commented 7 years ago

@etarena ok, that was my fault then....

i added those IJ.log lines allover the place. if you see it prevents the duplication, but the results are still printed correctly to ij log, then all good! ;-)