Nannotax Enabling nannofossil research

My workflow for adding plates from publications

A prime way of expanding content on the Nannotax website is to upload images from key publications. These notes explain the workflow I have developed for doing this, it is not a beginners guide but may be useful for people wanting to know how the site is being developed or are contemplating getting seriously involved. The objective is to end up with a set of imges which are tagged with, the publication they came from, the plate and image number in that publication and the species identification. The images from each publication will end up (a) being visible together in gallery pages as well as (b) appearing on the relevant species pages. If you have got high quality plates of images from publications of yours which you would like to see added to the Nannotax please do feel free to contact us. 

The programs I use are:
FileJuicer - a Mac shareware program for extracting images and other content from complex files.
ImageJ - a freeware, platform independent, Java image analysis program.
iView - an image cataloguing and databasing programme (was available for Mac and Windows, latest incarnation is as Image Pro from Phase One).
   
1. Obtain plates - scans or original files.

There is a substantial amount of work involved in getting images into the system so the priority is to get publications from papers where (a)  we can get the original images, not rely on scanning the printed versions and (b) the images were first class and fully documented in the first place. Also enter the publication reference into the biblio module

2. Split into sub images and add scalebars
If the original plate files were produced in Claris Draw, Illustrator, Word, PowerPoint or similar you can use  FileJuicer to retrieve the subimages; then use imageJ to add scalebars.
If tif images or similar use rename macro to save separate sub-images and add scalebars. 2µm scale bars are prefered, add a label stating the bar length. NB I have written ImageJ macros to facilitate this.  
If a sub-image has two or more separate taxa in it then split it into sub-sub-images.

3. Annotate in iView
a. Create an iView catalog file and import the images
b. Use Edit/Custom_fields to add species-f custom field and set views to show this field + title + status field
c. Add image numbers in status field (as plate-image eg 02-05) be sure to add leading zeros so sorting works. For strange image names use ~ to sort an image to end of sequence - e.g 54~AA sorts after 54Z
d. Add species identification in species-f field (NB to allow autoentry and error checking be sure that on the computer being used the list of taxa is loaded as "species-f" as in library/application support/iView/Plug-ins/Vocaublary/Default/)
Use cf NOT ? and sp not sp. (periods and question marks will cause problems during upload to Nannotax)
e. remember to sync annotations (save from catalog -> images) periodically.

4. Use iView applescript to rename sub images in format "03-04 Watznaueria barnesiae.tif"
NB The idea here is to retain the original numbering of the plates so that the image details can readily be accessed from the reference and so that they are sorted into a useful sequence (hence adding the leading zeroes is useful). I have written are various applescripts to automate this and if the script window is open it can be custom edited then run without the need to separately save every variant.

5. Use iView to create jpg image copies NB If there are tif originals it is best to keep them as well as jpegs and leave jpeg creation till end of editing so that tiff images have the same annotations etc. The nannotax system, however, only really likes jpeg images.

--------------------

NB Steps 1-5 above have ended up with a collection of jpeg image files, each illustrating a single specimen, with scale bars on each image, and with file names that include both the original plate number and the taxon name. There are many different ways to get here and no special reason to follow the workflow outlined above. The next steps, however, deal with getting the images into Nannotax and so should be followed more closely. 

------------------

6. Create galleries using taxonomy module. Name the overall gallery with paper name and sub galleries with plate number or other cunning scheme. Galleries can hold up to 100 images on one page, generally 20- to 50 images seems to be a sensible number. To create gallery go to taxonomy module and select the image galleries editor (or ask me to do this for you). 

7. Batch upload plates into galleries. NB Direct drag and drop from iView into the batch upload screen works fine.
Bulk editing - select all by using checkbox at top left; add image gallery; add biblio ref; remove spurious tags from taxonomies not wanted; check names for possible loss of special characters etc.

8. Review and enjoy. The full set of images uploaded will be on the relevant gallery page(s) and the individual sub-images wil apear on the taxon pages.

 
NB It would be good to also upload PDFs of papers into the nannotax system if allowed by the publisher, and/or add a DOI link to the paper if available elsewhere. This is not a perfect workflow since data on the samples the specimens came  from and their age is not included, yet. The data is however accessible from the original  publications.

Jeremy Young March 2011