APLReader - parsing of mtb file

[dgault]

Issue was raised in the following thread - ​https://www.openmicroscopy.org/community/viewtopic.php?f=13&t=8013

The example Olympus APL database was provided in QA-17057

The reader is able to read in the column names and rows of the database. However when trying to load a list of image files it appears to be reading from the wrong table value.

Requires further investigation, preferably with more sample data in order to determined the correct values to be read when searching for the filenames and paths from the mtb table.

Update from Jan 2017

Further APL issues have been found in forum thread ​https://www.openmicroscopy.org/community/viewtopic.php?f=13&t=8200&p=17869#p17869
Associated file sets can be found in QA-17509

Main issues here:

• Naming of imported files would be much easier with Experiment Name being used
• Incorrect dimension parsing, Z planes being read as time stacks
• Some of the associated files are listed with the full file paths rather than a suitable name

Initial debugging suggests that different orders of column headings are used for different files

[dgault]

Additional sample files provided in QA-17107

[dgault]

PR has been opened for a number of reported issues -​https://github.com/openmicroscopy/bioformats/pull/2304

The sample file procided ExampleDatabase? still has unresolved issues. In it there are files which for example are 768 x 576, 1 channel, time point, 1 z point, 8 bits per data, we are expecting to have a dataset of size 442368, but instead the loaded data is larger than expected (66977). When the tif is opened on its own, either through Bio-Formats or otherwise it appears to have 3 channels. It is worth noting that when opened on its own the tiff also does not appear to display correctly. The following is the data available for the image from the database:
ID 2
Flags 0
Record Name Wafer
File Name 435VG8IO_F00000002.tif
Path C:\Exampledatabase\435VG8IO_DocumentFiles\435VG8IO_F00000002.tif
File Size 66977
Image Name Wafer
Bits per Pixel 8
X-Resolution 768
Y-Resolution 576
Image Type Mono
Document Creation Date 11/26/1992 0:0:0
Document Creation Time 12/13/1901 20:45:52
Channel ADDA 1
Magnification 15000
Width 6692.81045751634
Height 5137.123745819398
Calibration Unit nm
Document Insertion Date 11/8/2002 0:0:0
Document Insertion Time 12/13/1901 20:45:52
Volume ID 1
Volume Label Exampledatabase_435VG8IO_0_0
Parent ID 1
RecordType? 2
Date 11/8/2002 0:0:0
Time 12/13/1901 20:45:52
Fixed 0
Frames 1
Audio File 0
Example Wafer

[dgault]

Further testing shows that there are additional issues regarding populating metadata for APL datasets.

From QA-17107, the LifeScienceDemo? dataset, only a single image from a large number of tiffs are being recognised. Reading the mtb file is returning only 3 rows of data, 1 with the row headings and 1 with a single image listed. This fileset should contain further additional image files which are not being read correctly.

Additionally the "example2" dataset is returning incorrect metadata, listing 33 metadata records for 11 images in many instances, other times info is missing or displaying incorrectly. At least one record name is missing completely (for picture 9, should be "PGP").

[szleo]

Replying to dgault:

From QA-17107, the LifeScienceDemo? dataset, only a single image from a large number of tiffs are being recognised. Reading the mtb file is returning only 3 rows of data, 1 with the row headings and 1 with a single image listed. This fileset should contain further additional image files which are not being read correctly.

Not sure if this is related to the bug, but two of the TIFF files are hidden:

._4X56O858_F00000085.tif
._4X56O858_F00000088.tif