Phase I polishing

[nsheff]

Phase 1 polishing

These are the things I've found that I think we should polish up to finalize phase I before moving on to phase 2.

Display cleanup

• [x] png images displayed on the splash page should be links to bigger images (PDFs)
• [x] include links to download the bed file, and to JSON metadata, on the splash page for the BED file
• [x] add a digest (md5 checksum for the bed file) to the splash page
• [x] The PNG titles should be below, not next to the plots
• [x] The display size of the PNG thumbnails should be reduced to about 75% of current size
• [x] Splash page should show the name of the BED file
• [x] Needs a logo (logo is broken)
• [x] The splash page should show the other metadata associated with the file, like cell type. this information is included in the LOLA database format, for example, and should therefore be passed on as PEP csv columns. All this information needs to be displayed on the splash page.

Features

• [x] there should be some way to locate files on the main page; like a search bar. several search components are outlined in phase I, such as searching by regions, cpg content, etc. these should have corresponding boxes on the dashboard to retrieve lists of bed file links.
• [x] implement all requested API endpoints in phase I document

Repo cleanup

• [x] Add BSD2 license files
• [ ] rename bedstat-rest-api to bedbase.

Bedstat cleanup

• [x] bedstat should NOT produce a 'file.path' file, but should record the path in the json document. Furthermore, the path should be stored as a RELATIVE path, not an absolute path. the PEP and files should be able to be moved around easily.
• [x] plot text is too small; should use smaller canvas size (not resolution) for PNGs (using width=x, height=x in the R png() and pdf() function calls, I think).
• [x] the pipeline should produce both PNGs (for thumbnails) and PDFs (vector, as the target of the links).
• [x] bedstat is currently keeping files unzipped. The specification calls for files to be stored and served in compressed format.

[nsheff]

ok looks good. but the search result links are the same:

see how the ID differs but they all link to the same thing?

ID	Link to html	Raw file	JSON
wgEncodeAwgTfbsBroadGm12878CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadDnd41CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadDnd41CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadDnd41CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadGm12878CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadH1hescRbbp5a300109aUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadHelas3CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadDnd41CtcfUniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json
wgEncodeAwgTfbsBroadH1hescEzh239875UniPk	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=html	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=bed	http://0.0.0.0:8000/regionset/?id=wgEncodeAwgTfbsBroadGm12878CtcfUniPk&format=json

[nsheff]

~Almost everything else looks great! The only issue I found is: bedstat is currently keeping files unzipped. The specification calls for files to be stored and served in compressed format. In my hands, the "raw_bedfile" linked is still unzipped. this should be only linked if it's zipped; if it's unzipped, it should be zipped into the folder. One last tiny thing: instead of naming every link as "raw_bedfile", let's use the actual name of the bedfile or something.~

Never mind, I can take care of these things. I see what's going on with the zipping now.

[nsheff]

Ok I think phase I is done once you fix those links.

[oddodaoddo]

Ok, done. That was stupid, I looked at the search results page but I did not see the details :disappointed: