sjcjohnston
commented
9 years ago First: Sorry for taking so long to respond. Second: I would suggest a separate file for each traced image (preferably in a .zip). The reason for this is that oftentimes (at least with the Gaelic data) the set of images that are traced at once don't necessarily belong together, ie. they are images of different sounds for different studies, and (in my opinion) they are easier to manipulate separately.
Of course this is under the framework/mindset that we have had in the past. If it would be easier for Mohsen's database to keep track of and load in the trace files if they were all within one CSV file that might be better in the long run. The only potential issue with putting them all within a CSV is if the metadata (the sound and word that the trace belongs to) would be accessible or not. If I understand correctly, Mohsen's database reads in the trace files individually, which currently have all of that information in them; I don't know how the database would obtain that information, unless there is a header line within the CSV containing all of the required information for each trace.
Long story short: For the short term, I would say individual trace files would be easier for the analysis code that I have written for Gaelic to run, otherwise I would have to write a script separating the information.
For the long term, I think it might be cleaner (less clutter) if there is one CSV file with all required information that the database can load and store. When we need individual trace files, then we could obtain them from the database, which (I assume) would extract the information into separate trace files for us to use.
Final answer: Both? Unless the change to the database to accept this kind of CSV file is trivial, in which case we could start using the database now in order to extract the individual traces we need.
Should we download separate files for each traced image or a csv-like file for a set of images?