Peter, Dan, Julie, Nicole Notes

[DoctorBud]

General TE Features

• AC menus should scroll and be positioned so that they are visible. There are issues with AC menus at the lower and right edge of the window.

Ontology Editing Mode (TermGenie)

• (HIGH) Prepare for GOC meeting and TermGenie replacement demo. Use CSVs and Patterns from GO. Need GO CSVs to be generated by Chris or Jim, or will need to build dosdp-scala and figure it out myself.
• (HIGH) Generate new IRI via a UUID mechanism
• (HIGH) Implement a mechanism to allow a PR or Issue to be created at the end of an editing session. This will require dealing with GitHub auth. I might use AWS Lambda for the first attempt, since I think there are good examples of this. There may be a way similar to the latest Noctua GitHub integration that does not require OAuth and instead just creates links to be executed by the user.
• (NICE) Generalize the iri label computation that is currently hardcoded from the BEER example. This computation needs to use the template in the pattern file.

NiceToHave, Next Steps, Technical Debt

• Pre-size columns to be more reasonable (e.g., Sex can be thinner than Description)
• PUB is freetext, but might be extended to a known list of PUBs later
• Description is free text for now

HPO Features

The new inline ontology feature lets us specify small lists of terms in the config.yaml file. This is currently being applied to evidence codes, but should be adapted to address the issues below:

• Support an inline micro-ontology (or GOLR if possible) for terms under Frequency http://compbio.charite.de/hpoweb/showterm?id=HP:0040279. Since Frequency only has a single column, we will use the canonical name for HPO, rather than ID. See http://compbio.charite.de/hpoweb/showterm?id=HP:0040284 where 'very rare' should be stored.
• Peter will look into what values are appropriate for SexID and Sex.
• For fields like 'Negation', we can either choose from a micro-ontology or we can just have a warning/filter that guarantees emptystring or NOT.
• Date should be prepopulated
• The 'Add Row' button should be enabled in patternless mode, so that new HPO files can be created.
• The solution I adopted for HPO phenotype restriction was to avoid using Monarch's GOLR and to use OLS instead. That hints at problems with our GOLR configuration (or my understanding of how to query it). We really should make it so we can query our GOLR instead of OLS. Even better, make it so BioLink can be used.
• Date should be reformatted as per HPO
• (Suppressed Feature) OMIM should be MIM
• Note that the OMIM Shoutcase names are legacy, and likely the Disease Name is unused. In the fullness of time, we should use @cmungall's https://github.com/cmungall/capital-offence

Fixed

• Test and fix issues where New Row is quirky after doing a Sort By. cf Sorting the table by phenotype name column causes the add new row feature to be quirky #25
• Add a Delete Row capability.
• Add Row should scroll to editable new row
• Note that .tab doesn't seem to work with DragAndDrop
• Ensure that Phenotypes are restricted to HPO. Current plan is to do filtering in the browser.
• Fix the HPO mode so that the 4 evidence shorthand codes are used. Current plan is to use an inline micro-ontology (a list of terms) in the hpo/config.yaml.
• Fix bug where two AC menus can be open at the same time
• Verify that Firefox and Chrome work properly. Peter said that resize column didn't work, and I had load errors on Mac.
• Debug the cell-saving code, which is currently too sensitive to window scrolling, resulting in premature selection of an AC item.

[jmcmurry]

Notes from today's meeting:

• Add row anywhere (not just at top)
• Sorting should go into testing regime
• Autopopulate “assigned by”
• Make use of PrefixCommons to validate prefixes
• Autopopulate date
• Support creation of brand new files (not just editing existing ones)
• Validate using Travis (dependent on HPO annotations repo becoming public)

[pnrobinson]

Note that this (Debug the cell-saving code, which is currently too sensitive to window scrolling, resulting in premature selection of an AC item.) is described more fully in #25