Why does Autoquant reject xml <SourceFile> input (for implentation of AutoQuant Connect)?
Hi everyone,
I am attempting to automate batch processing (3D deconvolution) using Autoquant Connect, but I have run into a snag and would really appreciate any advice.
Some details:
I am trying to use Autoquant Connect to send files to Autoquant for processing. I originally used some sample files from previous tests done a by a colleague and managed to get the process working. By process I mean
- Write xml file containing data required for Autoquant to perform deconvolution
- Use command prompt to send xml data to Autoquant to perform the processing.
We now have our own 2 photon microscope up and running and I have acquired some images to process by using the process and xml file structure above. For the image acquisition I use Prairie View, and for a test batch I have acquired:
- 3 areas/regions
- Of 14 slices each
- Each region repeated 4 times
The folder structure output from Prairie View is
- Main folder with all regions
o Folder containing Region 1
§ Z stack 1 (14 slices)
§ Z stack 2 (14 slices)
§ Z stack 3 (14 slices)
§ Z stack 4 (14 slices)
o Folder containing Region 2
§ Z stack 1 (14 slices)
§ Z stack 2 (14 slices)
§ Z stack 3 (14 slices)
§ Z stack 4 (14 slices)
o Folder containing Region 3
§ Z stack 1 (14 slices)
§ Z stack 2 (14 slices)
§ Z stack 3 (14 slices)
§ Z stack 4 (14 slices)
For the moment I use only one channel but in the future I plan to use at least 2.
The problem I am having now with the different structure is that Autoquant does not seem to like the specification of the Source File I put into the xml (for Autoquant Connect step).
I have tried:
- Using the *.xml file output from Prairie View, there is one for each region
- Specifying each .tif individually by specifying channel and frame as attributes in the source file tag
Both
of these have not worked. The Result file ErrorDescription tag states: Error
preparing deconvolution; code: 1. Unfortunately I have no way of knowing what
code 1 means and so I am unable to debug logically without trying every random
possibility of Source File input.
Essentially in the end I would like to send the entire main
folder (containing all region folders containing all stacks and repetitions) to
Autoquant to deconvolve in succession.
Thoughts anyone?
I would greatly appreciate any help/advice on this as it is really bottlenecking the subsequent processing I am working on. I can send example files of the generated xml as well as the code I have written if this would help.
Thanks,
Gabi
Answers
If possible, I would very much like to see the generated XML to see what might look off about it.
Thanks,
Andy
The files are attached in the email David CC'd you in on.
Please let me know if you received them or if I should send again.
Thanks,
Gabi
For anyone else that may be having issues with an AutoQuant Connect file, there were a couple of pointers that were relevant in this case.
First, when specifying the SourceFile location attributes (TimePoint, Channel, and Frame), keep in mind that the locations are zero-indexed. You may have files with one-indexed names, like:
c:\images\myexperiment_t1_ch1_z1.tif
c:\images\myexperiment_t1_ch1_z2.tif
but for the purposes of the SourceFile node, those will need to map to set locations starting from zero, like this:
<SourceFile TimePoint="0" Channel="0" Frame="0">c:\images\myexperiment_t1_ch1_z1.tif</SourceFile>
<SourceFile TimePoint="0" Channel="0" Frame="1">c:\images\myexperiment_t1_ch1_z2.tif</SourceFile>
Second, the dimension nodes - ImageWidth, ImageHeight, FrameCount, ChannelCount, TimePointCount (in particular the last two) - are optional in cases where they can be taken from the SourceFile metadata, or are 1 in value. But they become necessary otherwise. When in doubt, include them - having them be redundant won't hurt anything.