Deduplication Should Not Receive Credit

[0x4007]

We use footnote links for /annotate and issue deduplication

Excessive credit is being given to the author of the comment with any of those generated footnotes

https://github.com/ubiquity-os-marketplace/command-wallet/issues/47#issuecomment-2661604653

Do not count deduplicated footnotes. I suppose we can target the 0 prefix in the footnote ID but it also makes me think we should think of a more robust targeting system for the future. For now regex footnotes with ids starting with that - should be fine to remove from the content during preprocessing.

In the near future, we should actually prefix the generated footnotes with something more precise like dedupe- or no-credit or even better, check the diff and only offer credit if the user wrote the link, not a bot. [^01^]

If we can identify who wrote what, then we can extremely precisely reward credit where its due.

[^01^]: ⚠ 77% possible duplicate - Crediting for unique links only

[ubiquity-os-beta[bot]]

[!NOTE] The following contributors may be suitable for this task:

gentlementlegen

80% Match ubiquity-os-marketplace/text-conversation-rewards#155

koya0

80% Match ubiquity-os-marketplace/text-vector-embeddings#67

[ubiquity-os-beta[bot]]

[!NOTE] The following contributors may be suitable for this task:

koya0

81% Match ubiquity-os-marketplace/text-vector-embeddings#67

Keyrxng

79% Match ubiquity/devpool-directory-tasks#15

gentlementlegen

77% Match ubiquity-os-marketplace/text-conversation-rewards#155

[whilefoo]

Since all deduplication notes start with ⚠ xx% possible duplicate and annotate with xx% similar to issue, we can target that to avoid removing author's footnotes (I wonder if the plugin clashes with author's footnotes).

Another option is to enclose plugin's footnotes with <!-- deduplication-footnotes-start --> and <!-- deduplication-footnotes-end -->.

[0x4007]

Possible duplicate I think we could change to say something like like possibly related.

I think the 0 prefix or something similar could be a good indicator.

I always like metadata as well.

[gentlementlegen]

Another option is to enclose plugin's footnotes with and .

I like the idea but I think links would be credited still because the link within the comment (e.g. note[1]) would probably count as a link. Actually I think that footnotes themselves are already excluded (their content is not evaluated), the rewards are because of these footnotes link in the spec itself.

https://github.com/ubiquity-os-marketplace/text-conversation-rewards/blob/b9189bd8af038e750c4e9843a6b4599f25809298/src/parser/data-purge-module.ts#L65

One idea could be to ignore links redirecting to the same url as the issue they are written on?

I wonder if the plugin clashes with author's footnotes

Probably all the footnotes will be excluded.

[gentlementlegen]

/start

[ubiquity-os-beta[bot]]

Beneficiary

0x0fC1b909ba9265A846b82CF4CE352fc3e7EeB2ED

[!TIP]

• Use /wallet 0x0000...0000 if you want to update your registered payment wallet address.
• Be sure to open a draft pull request as soon as possible to communicate updates on your progress.
• Be sure to provide timely updates to us when requested, or you will be automatically unassigned from the task.

[ubiquity-os-beta[bot]]

[!IMPORTANT]

• Be sure to link a pull-request before the first reminder to avoid disqualification.
• Reminders will be sent every 21 hours if there is no activity.
• Assignees will be disqualified after 1 day and 18 hours of inactivity.

[0x4007]

Another option is to enclose plugin's footnotes with and .

I like the idea but I think links would be credited still because the link within the comment (e.g. note[1]) would probably count as a link. Actually I think that footnotes themselves are already excluded (their content is not evaluated), the rewards are because of these footnotes link in the spec itself.

text-conversation-rewards/src/parser/data-purge-module.ts

Line 65 in b9189bd

.replace(/^###### .?[\^\d+\^][\s\S]$/gm, "") One idea could be to ignore links redirecting to the same url as the issue they are written on?

I wonder if the plugin clashes with author's footnotes

Probably all the footnotes will be excluded.

Footnotes can be distinguished against normal links surely. Lets take a step back and think about it.

The portion inside the body should be a hash link to the same page the comment is written on. The footnote may have a normal link that goes to an external source which we should reward.

If we check the revision history and see who wrote it, we can assign credit accordingly. If the bot wrote something, then no credit is generated. etc.

Truly I think that seeing who wrote what is the most robust solution. We should definitely be doing this.

[gentlementlegen]

A footnote in MD form is text[^1] and [^1]: footnote. Transformed into html it is a [^01^]: ⚠ 67% possible duplicate - <a href="https://www.github.com/ubiquity-os-marketplace/command-wallet/issues/20#20">no docs for populating supabase</a> . I believe GitHub parses it and transforms it on its side, no this is not visible in html either, not distinguishable.

We could eventually rely on the revision. But this also will collide with the rewrite plugin, so issue specs won't be rewarded anymore if the user triggered a rewriting because the changes will be authored by the bot.

[0x4007]

But this also will collide with the rewrite plugin, so issue specs won't be rewarded anymore if the user triggered a rewriting because the changes will be authored by the bot.

If this is the only exception lets allow it to be a problem but only map it to /rewrite instead of including the time label change. Alternatively we can adjust the prompt to make absolutely minimal changes to the spec, which I think is a good idea either way. If it leaves some of the original authored content then this is perfectly valid!

---

Footnotes in the source code when I write them looks like this:

My footnote[^1^]

[^1^]: details

[0x4007]

test ^1

[0x4007]

You taught me something new, I didn't realize that we only need a single ^

[gentlementlegen]

Okay then maybe let's use regex as an immediate fix, and forward let's solve https://github.com/ubiquity-os-marketplace/text-conversation-rewards/issues/201 so we have a precise evaluation of footnotes based on the comment history.

[ubiquity-os-beta[bot]]

 [ 108.12 WXDAI ] 

@gentlementlegen

Contributions Overview

View	Contribution	Count	Reward
Issue	Task	1	100
Issue	Comment	1	8.12

Conversation Incentives

Comment	Formatting	Relevance	Priority	Reward
I like the idea but I think links would be credited still becaus…	9.06 content: content: p: score: 0 elementCount: 1 a: score: 5 elementCount: 1 result: 5 regex: wordCount: 78 wordValue: 0.1 result: 4.06	0.7	4	8.12

 [ 11.28 WXDAI ] 

@whilefoo

Contributions Overview

View	Contribution	Count	Reward
Issue	Comment	1	11.28

Conversation Incentives

Comment	Formatting	Relevance	Priority	Reward
Since all deduplication notes start with `⚠ xx% possible dup…	2.2 content: content: p: score: 0 elementCount: 1 result: 0 regex: wordCount: 38 wordValue: 0.1 result: 2.2	1	4	11.28

 [ 86.704 WXDAI ] 

@0x4007

Contributions Overview

View	Contribution	Count	Reward
Review	Base Review for #308	1	25
Review	Code Review	1	1.32
Issue	Specification	1	14.44
Issue	Comment	5	45.944

Review Details for #308

Changes	Priority	Reward
+28 -5	4	1.32

Conversation Incentives

Comment	Formatting	Relevance	Priority	Reward
We use footnote links for `/annotate` and issue deduplic…	11.3 content: content: p: score: 0 elementCount: 1 br: score: 0 elementCount: 2 a: score: 5 elementCount: 1 result: 5 regex: wordCount: 131 wordValue: 0.1 result: 6.3	1	1	14.44
Possible duplicate I think we could change to say something like…	1.95 content: content: p: score: 0 elementCount: 1 result: 0 regex: wordCount: 33 wordValue: 0.1 result: 1.95	0.7	4	7.056
Footnotes can be distinguished against normal links surely. Lets…	4.97 content: content: p: score: 0 elementCount: 1 result: 0 regex: wordCount: 99 wordValue: 0.1 result: 4.97	1	4	24.88
If this is the only exception lets allow it to be a problem but …	4.88 content: content: h2: score: 1 elementCount: 1 p: score: 0 elementCount: 1 result: 1 regex: wordCount: 74 wordValue: 0.1 result: 3.88	0.5	4	12.52
test	0.1 content: content: p: score: 0 elementCount: 1 result: 0 regex: wordCount: 1 wordValue: 0.1 result: 0.1	0	4	0
You taught me something new, I didn't realize that we only need …	1 content: content: p: score: 0 elementCount: 1 result: 0 regex: wordCount: 15 wordValue: 0.1 result: 1	0.3	4	1.488