drivenbyentropy / aptasuite

A full-featured bioinformatics software collection for the comprehensive analysis of aptamers in HT-SELEX experiments.
https://drivenbyentropy.github.io/
GNU General Public License v3.0
24 stars 11 forks source link

GUI - Aptamer Family Analysis not working #14

Closed PJpb closed 6 years ago

PJpb commented 6 years ago

Hi, After running the clustering, I get a table of clusters with their IDs and counts on the left of the window. image I can sort the clusters by size in the table. image

When I press any of the clusters on the left, the app calculates the data for a moment, and next the middle window stays blank but shows some columns. No charts on the right. image

After I click one of the other clusters on the left, the mid table dissapears and nothing changes on the right. image

LOG: log_2018-02-16_16-26-08.txt

PJpb commented 6 years ago

So after the changes (v0.5.7 now) the bug persists (even when using local instance and files). Here are the source data file and the logs, including the PowerShell console log: 1302_merged_utf8_1000seq_txt.txt log_2018-02-17_16-46-18.txt configuration_txt.txt console_log.txt

drivenbyentropy commented 6 years ago

Thanks for the detailed report. I can reproduce this here with any dataset as well and I am looking into it.

drivenbyentropy commented 6 years ago

Should be fixed in v0.5.8.

PJpb commented 6 years ago

The bug persists. Nothing is displayed in the middle window. The middle window (table columns' captions) goes blank after changing the cluster. And actually aptasuite halted totally ("Brak odpowiedzi" in the window name means "Not answering" ;) ) when I changed the clusters again, I had to force-close it. (App started from PowerShell in Administrator mode) image 1302_merged_utf8_1000seq_txt.txt console_log.txt log_2018-02-18_12-38-05.txt configuration_txt.txt

edit: Also, please check that, it was calculating infinitely, I had to close down the app: image console_log_Cluster_error_at_the_end.txt log_2018-02-18_13-10-38.txt configuration_txt.txt

drivenbyentropy commented 6 years ago

This one was hidden deep, but I think I fixed it. Please confirm if you have the opportunity. Thanks!

New version is v0.6.0

PS: You might need to create a new experiment as v0.6.0 is not compatible with previous versions in some aspects.

PJpb commented 6 years ago

Sorry, no luck. (GUI v0.6.0 started from CLI with administrator's privileges) image

Also the re-sorting (Div vs Size) doesn't work, it makes the table on the left disappear. Re-running the AptaCluster doesn't make it appear again. image

1302_m2k.fastq_txt.txt console_log.txt log_2018-02-19_15-04-35.txt configuration_txt.txt

drivenbyentropy commented 6 years ago

Wow, this is crazy.

At first I could not reproduce your results for anything. Then I looked again at your console log and and realized that the error was related to parsing a string which was formatted using Javas DecimalFormat class to three decimal points.

DecimalFormat df_count = new DecimalFormat("#.###");

What I failed to realize before, is that DecimalFormat is subject to localization. This means that

double count = 123.456789;
String count_string = df_count.format(count);
System.out.println(count_string);

Would print 123.456 in the US, but 123,456 in other countries which use comma as the decimal separator. Hence, any table in AptaSuite showing the sequences would fail to load when creating the cycle counts columns due to

Number count_number = Double.ParseDouble(count_string);

which expects a dot as decimal separator.

I have now added a fixed locale to all relevant sections in the code. Please let me know if this finally fixes the issue. It should also fix #18.

New version is v0.6.1

PJpb commented 6 years ago

Was this the main error all the time, or have you actually fixed other stuff as well thanks to me being a pain in the ass?;)

It works and it's beautiful. Great job, thank you :)

(More bugs to come over the week when I get down to the juicy stuff and start to analyze real data ;) )

drivenbyentropy commented 6 years ago

It was a mixture of smaller issues and this one, the latter being responsible for most of the confusion/non-reproducibility ;-). I am glad it is finally fixed!

Thanks so much for your persistence.