FlagImage Error: a float is required

Dear All,

I am trying to select witch images I will analyze in CellProfiler by using the FlagImage Module. I have a column in my metadata file that contains numeric values (0 for those images I what to skip and 1 for those I want to run). I select to Skip the image if it is flagged and choose a minimum value (0.5, but 1 also does not work) to flag the image if it is below. Flag is based in Whole-image measurement.
I get the error message below. It asks for a a float, but even if I use 0.1 and 1.1 (instead of 0 and 1) in the column in my Metadata file, the error keeps happening. I am using CellProfiler Version: 2.2.0rc2 in a 64bit Windows 10 machine.
Is there an alternative way to do this ?

Thank you very much!!



Traceback (most recent call last):
File “cellprofiler\gui\pipelinecontroller.pyc”, line 2826, in do_step
File “cellprofiler\pipeline.pyc”, line 2067, in run_module
File “cellprofiler\modules\flagimage.pyc”, line 364, in run
File “cellprofiler\modules\flagimage.pyc”, line 478, in run_flag
File “cellprofiler\modules\flagimage.pyc”, line 514, in eval_measurement
TypeError: a float is required

Sorry all,

I just solved the problem!! I was supposed to choose the data type, separately for each metadata entry, in the last drop-box setting of the Metadata Input module (the default is to consider all text). The help is clear about it (see below)

If anyone has the same problem, here it is.



Metadata can be stored as either a text or numeric value:

Text: Save all metadata item as text.

Choose for each: Choose the data type separately for each metadata entry. An example of when this approach would be necessary would be if a whole filename is captured as metadata but the file name is numeric, e.g., “0001101”. In this situation, if the file name needs to be used for an arithmetic calculation or index, the name would need to be converted to a number and you would select “Integer” as the data type. On the other hand, if it important that the leading zeroes be retained, setting it to an integer would them upon conversion to a number. In this case, storing the metadata values as “Text” would be more appropriate.

Just to add something. I got the error message only when in test mode. When I tried to run the pipiline before checking, CellProfiler just stopped responding and I had to kill it.