Not sure if this is currently possible -- add support if not.

Various printouts to distribute during the competition:

• per-person schedule sheets (sorted by staff team, fixed height)
• Master sheet for stage leads

not for 2023.

e.g. multi blind scrambling, registration

Does WCIF and competition-groups support this?

with a different status (not "pending").

Needs wcif and wca website updates.

There are going to be lots of null checks, most of which are just going to be early returns.

Add a nullable field to args. If not set to true, pass the null upward. If true, the function should handle it differently.

JobCountScorer prefers giving staff jobs to people who have done less work so far. There are a few issues:

• Not all jobs are the same -- a 2x2 group that lasts 10 minutes should be weighted less than judging Multi for an hour
• Some credit should be given to people competing, not just staffing. While ideally everyone would get the same amount of work, regardless of how many events they do, in practice we don't want to burn out staff who have a lot of events. Competing should probably count for approximately 1/4 the weight of judging, but this can be decided by the calling script
• If we are considering competing, we should only consider jobs that have already happened. Otherwise, if groups are assigned first, and then judging, then people doing a lot of events would have no jobs at the beginning of the competition, then switch to a full workload.

Figure out how to do scorecards. Fork groupifier scorecard code?

Currently you need to do (1 + 2), rather than just 1 + 2. See if this can be fixed without too much PEMDAS pain.

The error from peggyjs says what line + character has the error.

Rather than a regex, it should use a grammar to make sure it extracts the right number of angle braces.

e.g. right now if expectedValue is Array<Tuple<Number, String>> then genericValue is Tuple<Number, String>>.

Make a linear programming model for staff assignments on each group for a given round.

Not sure whether to treat this as a bug in Cluster() or my scripts for Nats23, but a staff member who dropped out received staff assignments late in the competition. Make sure this doesn't happen again.

Currently passing [] as an argument makes it an Array<$T>, which fails to resolve in whatever calling function expects an Array. Figure out a way to make this work.

In /script, we're currently just logging when there's an error in the WCA API.

Compscript is complicated enough that it would probably be worthwhile to add unit tests.

Add an option to run natshelper using WCA staging website or production.

There are a limited number of cases where it would be helpful to implicitly cast, e.g. Group -> Round, Round -> Event.

Be careful with extending this too far, e.g. Number -> String could break Add().

for example:
-group start time in a range
-round that another person is competing in

Group assignment for finals should be implemented differently (even / odd, station numbers)

It's weird to have e.g. a function that returns Boolean(Person) and takes an Activity as an arg, and an equivalent Boolean(Activity) that takes Person as an arg.

Refactor arg matching + selection in driver.js to allow some function args to be external (i.e. provided when resolving, not in the textbox).

Type inference is certainly not perfect. Consider allowing the input to specify types, e.g.

Function<TypeA, TypeB>(args)

The parser should recognize Event and Round differently. Event and Round can continue to be represented by an ActivityCode.

The parser does not need to recognize Group; this should be an internal-only type. The underlying representation should be a WCIF Activity. ActivityCodes for Groups are ambiguous after #17, so ActivityCode is no longer sufficient.

There should be conversion methods between these three types.

• It's stored as a string in wcif, apparently
• (NumberProperty("staff-team") == NumberProperty("staff-team", 2008CLEM01)) works
• (StringProperty("staff-team") == "1") doesn't work
• (NumberProperty("staff-team") == 1) doesn't work

??

At the moment this is more a question for which the answer should likely be public: is there any basic example out there of a basic spreadsheet format and the result it provides in person's properties when imported?

I've been shared some private example usage of compscript, but the data in spreadsheets have been appropriately omitted, and now I'm left with a few questions on the format I should target (because importing data dynamically sounds like a great idea for Euros ;)).

Let's take for instance the following data (csv, but separated by ';' intentionally, because of how I think values are parsed in the code):

identifier:string:wca_id;identifier:integer:user_id;property:list:scramble-events;property:string:team-kind;ignore:Some nice thing
2008VIRO01;;3x3,2x2;teams;titi
2008PIAU01;;5x5,2x2;teams;toto
;61698;;teams;tutu
2018GODI01;;;data-entry;tata

Reading from the source code, I would guess this results in adding the properties "scramble-events" and "team-kind", respectively holding an array of string and a string, that I would be able to query using respectively ArrayProperty("scramble-events") and StringProperty("team-kind")?
The "Some nice thing" column would simply be ignored.

Could you please confirm using this format is enough to identify whom to attach the data to based on the wca_id/user_id fields? (and does it matter if both of them are filled?)

What happens if one of the columns' header doesn't have the ignore:/identifier:/property: prefix?

For the record I still intend to run a test competition with a simplified example of basically everything, that could serve as a publicly available showcase of compscript (and for which I could make the spreadsheet used public, hence "fixing" this "issue".

Make it only depend on the times. TBD whether it should be startTime, endTime, and/or date.

Currently group names are of the form "Red1". This will break

Other tools like competitiongroups.com display the room name (Red) next to the group number, so this isn't necessary. It also will break code like Groupifier's parseActivityCode.

Switch to assuming group names are numbers.

This is the table in the readme:

Table(
  Persons(And(Registered(), (FirstName() == "Luke"))),
  [Column("Name", Name()),
   Column("WCA ID", WcaId(), WcaLink()),
   Column("Average", PersonalBest(_333)),
   Column("Single", PersonalBest(_333, "single")),
   Column("psych sheet ranking", PsychSheetPosition(_333))],
  PersonalBest(_333, "average"))

It does not seem to work "as-is", and I think it's because of recent changes in the sort functions.

Removing the last arguments makes it work, but loses the sort capabilities: would you have an updated example to sort based on an attribute of the elements?

I've started playing a bit more on real data: creating and assigning groups emits a lot of output; just providing a sample below:

Added group 7x7x7 Cube Round 1 Blue 1 from 2024-07-25T10:00:00.000+02:00 to 2024-07-25T10:35:00.000+02:00
Added group 7x7x7 Cube Round 1 Blue 2 from 2024-07-25T10:35:00.000+02:00 to 2024-07-25T11:10:00.000+02:00
Added group 7x7x7 Cube Round 1 Red 1 from 2024-07-25T10:00:00.000+02:00 to 2024-07-25T10:35:00.000+02:00
Added group 7x7x7 Cube Round 1 Red 2 from 2024-07-25T10:35:00.000+02:00 to 2024-07-25T11:10:00.000+02:00
...
3x3x3 Fewest Moves Round 1 Groups
Side 1 (368)
... followed by the table of the group

It's absolutely great to have that output, but the concatenation of everything makes it a bit hard to read.
It would be awesome to be able to "opt in" or "opt out" of these outputs (ideally I'd love to select a couple of interesting outputs, but not everything).

Is there an option/parameter I missed?
Would you be open to have a way to "silence" the output for some calls?

Reference: thewca/worldcubeassociation.org#7492

After #37 there's no need to load the WCIF before parsing the script.

-Not enough breaks before competing -- increase the weight of this scorer or reset?
-Not enough anchoring to prior group job -- maybe change the way this works, to decay based on the number of previous consecutive groups (basically try to avoid singleton jobs)

Add support for tagging specific Activities (or Persons) as being on the live stream.

Figure out a generic way of mapping columns to properties.

Not just a hard cutoff of eligibility -- add a scorer for PR.

Add support for for() loops.

That will allow version controlling individual comps' natsscript.

Mysql TEXT fields have a 65k limit, which all of the UDFs in a single field are exceeding.

Some exceptions crash the program. Expose these better.

Editing groupifier extensions on groupifier (e.g. page size, scorecard layout) seems to require setting a number of groups. I'm not sure what side effects that will have. Add a way to write the extensions we need directly from natshelper.

So that a few staff can request to be on the same team.

I think it's intended, but I'll create an issue just in case: AddPerson only add the wcaUserId to the person's detail in the WCIF.
I read a bit the activity on the WCA website repository related to non competing staff, and it looks like to me that the intended workflow is the following:

• compscript adds the person's wcaUserId to the WCIF
• executing the change on the WCA website makes it create a registration for that user
• upon subsequent execution in compscript, the person's details actually show up now that the registration exists

Am I correct in how that function is supposed to be used?

I think there are a few actions that can be taken:

• document the function's behavior, and mention that forgetting to push to the WCA website will lead to incomplete information about persons, and crashes in some cases.
• optionally compscript could query the WCA website in such cases, to include the user's data right away in the WCIF (which should be fine even if we got them wrong, because they would be erased anyway when the website creates the actual registration).
• if we don't to query the website for that, we may also include dummy data until the registration is properly create (eg "Imported Person XYZ" with a fake dob); just so that the rest of the script doesn't crash when in "dry-run" mode.

I can definitely take care of 1, but what do you think about doing 2?
If you think it's a corner case I'd definitely be willing to do 3, as I do a lot of "dry-run" things before actually patching the WCIF!

AssignGroups needs to support FMC and Multi, where all competitors are assigned to all attempts.

Reference: thewca/worldcubeassociation.org#7491