Is it possible to show fixed comments first?

[raulsperoni]

Hi, maybe this has been answered but I haven't found it.

I know that comments show up in a semi random manner, as explained here: https://compdemocracy.org/comment-routing/

But in the experiment that I'm planning we would like to have some demographic variables about the participant, or to have some "first" comment that may work as a target feature in a posterior analysis.

Even if pol.is as is doesn't allow to do this, If I were to embed the participation interface in another website, is there a way to use cookies or something similar to associate the person answering pol.is survey, with the person answering some question in the website where the iframe is loaded?

I hope I explained myself.
Thank you!!

[metasoarous]

Hi @raulsperoni. Thanks for your question!

You can absolutely do this. I'd recommend taking a look at https://github.com/compdemocracy/polis-embed-example. That shows you how you can use our xid system to track identity per user with a cookie (localStorage parameter actually) which you can tie to whatever questions you ask the person prior to them participating in the polis survey/conversation itself. You can then merge the data in post-conv analysis as you like.

Please let me know if this makes sense or if I can provide further clarification.

Thanks again!

[raulsperoni]

Hi @metasoarous,

So, I'm trying to achieve what you suggested, I can do it embedding from pol.is, but votes can't be casted to a custom pol.is installation in other domain (I've done a number of changes, and set env variables to make this installation work).

I'm getting this in the logs:

{
  lang: 'es',
  weight: 0,
  vote: 0,
  tid: 5,
  pid: 'mypid',
  conversation_id: '3cc5bz2bfc',
  agid: 1,
  xid: 'hola'
}
{
  host: 'cuestionario.luc.uy',
  'user-agent': 'Mozilla/5.0 (X11; Linux x86_64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/92.0.4515.159 Safari/537.36',
  'content-length': '108',
  accept: 'application/json, text/javascript, */*; q=0.01',
  'accept-encoding': 'gzip, deflate, br',
  'accept-language': 'es',
  'cache-control': 'max-age=0',
  'content-type': 'application/json; charset=UTF-8',
  origin: 'https://cuestionario.luc.uy',
  referer: 'https://cuestionario.luc.uy/3cc5bz2bfc?parent_url=http%3A%2F%2Flocalhost%3A3000%2F&referrer=&xid=hola',
  'sec-ch-ua': '"Chromium";v="92", " Not A;Brand";v="99", "Google Chrome";v="92"',
  'sec-ch-ua-mobile': '?0',
  'sec-fetch-dest': 'empty',
  'sec-fetch-mode': 'cors',
  'sec-fetch-site': 'same-origin',
  'x-forwarded-for': '186.48.2.119',
  'x-forwarded-host': 'cuestionario.luc.uy',
  'x-forwarded-port': '443',
  'x-forwarded-proto': 'https',
  'x-forwarded-server': '5dbff774ad8d',
  'x-real-ip': '186.48.2.119',
  'x-requested-with': 'XMLHttpRequest'
}
authtype doXidConversationIdAuth
no xInfo yet
created dummy
Error: there is no unique or exclusion constraint matching the ON CONFLICT specification
    at module.exports.Client._emitResult (/app/node_modules/pg-native/index.js:173:26)
    at module.exports.Client._read (/app/node_modules/pg-native/index.js:215:31)
    at PQ.emit (events.js:315:20)
    at PQ.EventEmitter.emit (domain.js:486:12)
    at PQ.callbackTrampoline (internal/async_hooks.js:129:14) {
  severity: 'ERROR',
  code: '42P10',
  file: 'plancat.c',
  line: '835',
  routine: 'infer_arbiter_indexes'
}

This is what's reported here: #287

But I'm a bit lost on how to fix it.

Thank you for your help!

[patcon]

My quick guess is that the one or two on conflict postgres queries involve an xid value, that is not being passed through for some reason. I'd search for where ON CONFLICT is used, then add debug conde to work back where the null/empty value is coming from. It might be something in the embed.js code that is called for embeds on the frontend of the embedding site... Maybe it's looking for a match to pol.is or your custom domain, and it's not passing that regex test for some reason.

Maybe helpful context (though resolved if you're at latest commit): #758

(Ps, FACTTIC is amazing. Goooo co-operatives!)

[raulsperoni]

(go co ops!)

Ok you are right, I was sending something that was not a string as xid. Thank you!

Next question (I don't want to be abusive with questions but I'm really excited about what we are trying to do with pol.is)

If I'm using data-xid, comments doesn't require social login even if that option is checked in the admin. I understand how maybe this is intentional, but even if we can moderate comments we would like to also have the "see you and your friends" in the groups if people want to.

I could implement this on my side and send the data right? this would be done using other parameters like data-xid in the iframe parent div?

I also wanted to say that I am willing to contribute in documentation or anything else that maybe useful (maybe on the tweaks necessary to deploy polis in other domain, but I'm wouldn't know where to start)

Thanks!

[metasoarous]

If I'm using data-xid, comments doesn't require social login even if that option is checked in the admin. I understand how maybe this is intentional, but even if we can moderate comments we would like to also have the "see you and your friends" in the groups if people want to.

This does seem like a bug to me. Would you please open up an issue for this?

I could implement this on my side and send the data right? this would be done using other parameters like data-xid in the iframe parent div?

I don't think this is correct; You could in the parent app/page require users to login to facebook/twitter before letting them participate, but I don't think that login would be reflected in the iframe necessarily (I could be wrong though).

I also wanted to say that I am willing to contribute in documentation or anything else that maybe useful (maybe on the tweaks necessary to deploy polis in other domain, but I'm wouldn't know where to start)

That's great! Thank you so much! We're always happy to accept PRs which improve the documentation or deployability of the software. You're always welcome to review our open issues for things that have already come up but haven't been fully resolved yet.

Thanks again!

(and yes, go co-ops 🙌)

[raulsperoni]

This does seem like a bug to me. Would you please open up an issue for this?

Yes: #1139

I don't think this is correct; You could in the parent app/page require users to login to facebook/twitter before letting them participate, but I don't think that login would be reflected in the iframe necessarily (I could be wrong though).

Yeah, I tried sending more stuff, but It doesn't get picked up

<div
       id="polisDiv"
       className="polis"
       data-conversation_id={conversationId}
       data-xid={"hola"}
       data-x_name={user?.name}
       data-x_profile_image_url={user?.picture}
     ></div>

That's great! Thank you so much! We're always happy to accept PRs which improve the documentation or deployability of the software. You're always welcome to review our open issues for things that have already come up but haven't been fully resolved yet.

Will do, thanks a lot for your help!