search

davidpng / FCS_Database

Program to scrape an FCS directory of metadata
GNU General Public License v3.0
3 stars 2 forks source link

What does add_FCSstats_db.py do? #13

Closed davidpng closed 9 years ago

davidpng commented 9 years ago

Moving this to here: When we run add_FCSstats what files should it use? - What is the goal behavior?

My proposal is that add_FCSstats should only operate on a prebuilt database from make_FCSmeta. Ideally, it should be able to load in all the files (from the directory structure) in the database without having to reference the filesearch functionality. I'm was a little confused as to why add_FCSstats needs to have a filelist search capability replicated from make_FCSmeta.

  1. You could use a filelist in directly as the args.dir (built in capabilities but need to be more explicitly documented.
  2. You could pull file directory from the database.

Then this brings up the question of why not just make the database with the meta and stats information. My objection is that we might want to update certain stats information over time so if we have a fixed metainfo database, we can always keeps a non-stat added version to update the thing. Also, it may turn out that we only need to extract stats from a subset of cases found by a metainfo-database search.

hermands commented 9 years ago

add_FCSstats searches for files because we I didn't have the option of pulling in a file list when I wrote it.

I think we agree then. add_stats should query database to get a file/case list and run on that. so it needs args.dir and it needs a db. Will need to change query functions so that default query result is all files/cases in db.

davidpng commented 9 years ago

Hi Dan, I think I screwed up the file list creation by making the list absolute rather than relative. I can fix that pretty easily though.

if you pass a *.txt file into args.dir instead of a directory, it will switch to load text file mode and read in the files into a list.

On Wed, Nov 12, 2014 at 6:57 PM, Daniel Herman notifications@github.com wrote:

add_FCSstats searches for files because we I didn't have the option of pulling in a file list when I wrote it.

I think we agree then. add_stats should query database to get a file/case list and run on that. so it needs args.dir and it needs a db. Will need to change query functions so that default query result is all files/cases in db.

— Reply to this email directly or view it on GitHub https://github.com/davidpng/FCS_Database/issues/13#issuecomment-62833547 .

hermands commented 9 years ago

Oh. I didn't realize that. nice.

On Nov 12, 2014, at 10:28 PM, David Ng notifications@github.com wrote:

Hi Dan, I think I screwed up the file list creation by making the list absolute rather than relative. I can fix that pretty easily though.

if you pass a *.txt file into args.dir instead of a directory, it will switch to load text file mode and read in the files into a list.

On Wed, Nov 12, 2014 at 6:57 PM, Daniel Herman notifications@github.com wrote:

add_FCSstats searches for files because we I didn't have the option of pulling in a file list when I wrote it.

I think we agree then. add_stats should query database to get a file/case list and run on that. so it needs args.dir and it needs a db. Will need to change query functions so that default query result is all files/cases in db.

— Reply to this email directly or view it on GitHub https://github.com/davidpng/FCS_Database/issues/13#issuecomment-62833547 .

— Reply to this email directly or view it on GitHub.

Daniel Herman, M.D., Ph.D. Laboratory Medicine Resident, PGY2 Office: (206) 598-0254 Cell: (914) 960-1004 hermands@uw.edu