Amend group-name parser to avoid forbidden forge-string characters

[stschiff]

Some of the group names that we've imported from the AADR contain symbols like < or others that trip up our parsing of forge strings.

Ideally we should have group names following alphanumeric symbols and _ and - and . or so, but not anything more exotic.

[stschiff]

I think that's something that you can work on, @dhananjaya93. Just check all group names and turn any symbols other than _, -, . or alphanumeric into underscores.

[dhananjaya93]

Got it. I will start working on it. Best , Dhananjaya

…

On Wed, 1 Dec, 2021, 5:16 pm Stephan Schiffels, ***@***.***> wrote: I think that's something that you can work on, @dhananjaya93 <https://github.com/dhananjaya93>. Just check all group names and turn any symbols other than _, -, . or alphanumeric into underscores. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <https://github.com/poseidon-framework/published_data/issues/35#issuecomment-983801667>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AWAV3NM6BMAD7TN3JD5ZA4DUOZC6VANCNFSM5HXWNQQA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

[dhananjaya93]

I checked the all the group names and identified the characters to replace. I can replace them but how can I change the individual packages that already exist? Do I need to update all the .janno files of existing packages?

[nevrome]

I guess these issues are not present in every package, right? If so, then it would boil down to writing a script that loops through the package folders, replaces the .janno files and updates the POSEIDON.yml files (trident update).

[nevrome]

@AyGhal and I have fixed the problematic groups that existed now in poseidon-framework/community-archive#132.

Another question is, whether we should introduce some mechanism to prevent the creation of packages with such "broken" group names. This could be a future feature of trident.

[stschiff]

Well, we could parse them in Haskell using a parser that only allows a certain alphabet. That would then automatically complain. Though it would be a pretty brutal move.

[nevrome]

I like it. Better a painful but clear error message than obscure failures with specific sub-commands, right?

[stschiff]

OK, I just renamed this issue to make the pending ToDo clearer.

[stschiff]

This Issue will have to be worked on right in sequence-formats and actually is connected to an issue that @TCLamnidis noticed about mismatching group names in Janno vs. Genotype data. I think we should make the parser that parses Eigenstrat and Plink data more strict.

[stschiff]

I will transfer this issue over to poseidon-hs.