feat(report_quiz): add a leaderboard by pythonbrad · Pull Request #119 · osscameroon/global-github-actions

pythonbrad · 2024-08-23T17:25:42Z

Related issues

Simple quiz bot with leaderboard project-ideas#51

Change

removed the anonymized to be able to get user answers
restructured the code to apply the DRY principle
saved the quiz data in order to be able to get poll results

TODO

close the quiz and get user answers
fetch quiz user answers in a database
generate a leaderboard
sanction / mention user who retract their answers
log to stderr
put reusable functions like escp in utils/function.sh
announce the result (message)
Review the workflow file

Notice

I plan to use a git repo as a database. Hence, the report_quiz action will hold a private repo.

The created quiz information will be committed in the git repo
git commit -m "Added a new quiz"
The user answers will be committed in the git repo
git commit -m "Added new user answers to quiz:<quiz_id>"
The final quiz stats will be committed in the git repo
git commit -m "Closed quiz:<quiz_id> and generated result"

Useful link

A jq cheetsheet: https://gist.github.com/olih/f7437fb6962fb3ee9fe95bda8d2c8fa4
A jq playground: https://jqplay.org/
The telegram API docs: https://core.telegram.org/bots/api

In saving the quiz data, we will be able to retrieve the submitted answers and build the leaderboard

This commit introduces the following changes: - Added a function `get_user_answers` and `fetch_quiz_data` to be able to get the user quiz answers and update the quiz_data database - Restricted the quiz answers to one answer submission per user per quiz

When a quiz_data is created, the user_answers field is not present and the addition can't proceed. As change, we will use an empty value in this case

- added the possibility to generate and send the leaderboard

Stop all the active quiz, submit the result and archive the quiz data

Added a friendly announcement message. Additonnal change - moved the price calculation from `generate_leaderboard` to `compute_total_score`. - fixed issue when a quiz data contains no user answer submission during the score computation

- Added the start new competition action - Added the fetching of the quiz user answers - Added the backup of the data using git - Added the ending of a current competition

pythonbrad · 2024-08-26T19:34:04Z

@Sanix-Darker , @elhmn , this pr is ready for review.

Added a documentation for the report quiz bot

The close competition fail during the closing of a already closed quiz. Since the status of this process is not important, we will ignore it. Like that the closing of the competition can proceed.

This commit handles the case when there is no results during the closing of the competition.

Sanix-Darker

Hey there, thanks for this contribution, can you provide a video demo or workflow on your PR description please ?

Sanix-Darker · 2024-10-06T11:38:32Z

@@ -0,0 +1,9 @@
+# compute the user total score from more than one quiz_data
+compute_total_score(){
+    jq -s --argjson total_price 30000 'map((.correct_answers | map(.)) as $answers | (.user_answers // []) | map(.score=(.options_ids | map(select($answers[.] == "true")) | length)-(.options_ids | map(select($answers[.] == "false")) | length) | del(.options_ids))) | add | group_by(.first_name) | map((map(.score) | add) as $total_score | .[0] | del(.score) | .total_score=$total_score) | group_by(-.total_score) | to_entries | map(.price=$total_price/((.key+1)*2))' $@


This is too compact, even tho it's understandable, make it digest please, not in a single line

Yes, but the fact that it's a set of sequential instructions, don't affect his readability.
After a look at the source JSON, it becomes easier to understand.
https://github.com/pythonbrad/reportquiz_data/blob/main/archive/1727937431/6050905932701893008.json

Between, I haven't succeeded to split it while keeping it more readable.
I can't do shell piping since it will affect the performance.

Note that I use functional programming.

Add in the comment that is up the method a simple use case représentation of what can be an input and an output, to help us see what the method is capable of if the JQ filter can not be on multi lines.

This jq command reflects the function name and his comment.

i agree with @Sanix-Darker , this is not easy to maintain. I am pretty sure you can split it into smaller sequential and logical chunks so that other people can understand it quicker. Performance is not an issue here, since this run in a runner on a cron, no one is waiting for this answer to be sent in less than 10 milliseconds. I am more for privileging readability/maintainability over speed here.

You also could write it in multiple lines for example, add some single line comments to your script, use re-usable variables etc which I think are supported by jq.

Sanix-Darker · 2024-10-06T11:40:11Z

+source "${BASE_DIR}/leaderboard/functions.sh"
+
+# load environement variables
+source "${BASE_DIR}/env.sh"


Add a -f check for this file
And exit 1 if it's not present before continuing

Fail fast is enabled at the top of the main.sh

I prefer handling personally exeptions and explicit errors messages to have full control when am building blocks in a codebase, but okay... Let's keep it like that.

To go deeper, the source will fails with something like "env.sh No such file or directory" (I guess)... But in an another composes scenario, that error is not going to reflect a user error meaningful enought to fix without relooking what the script does

Sanix-Darker · 2024-10-06T11:43:29Z

@@ -0,0 +1,84 @@
+MAX_LENGTH=100


MAX_LENGTH_OF... ???

pythonbrad · 2024-10-08T19:00:02Z

🗳️ To have at the end a human-readable data, at the level of the archiving should I convert the JSON to YAML using yq ?
Note:

yq is able to read the JSON as input.
yq is not fully jq query compatible.

pythonbrad · 2024-10-15T13:01:43Z

@Sanix-Darker, feel free to apply your updates in this pr

rakici

Thanks a lot for submitting this PR.

I went half way through the submission, and I decided to stop because I think that the complexity of this solution, does no go well with the programming languages that chosen (bash).

This is how I think we can do this leaderboard with bash or python using a simpler approach.

1- We will move to an unanonymous poll
2- We will add a new cron job that will request, answers from our users, and update 2 leaderboards.
- An all time leaderboard that aggregates scores for each user since the beginning of time
- A monthly leaderboard that gives us who leads and won for a single month. And is reinitialised each months.
3- That 2nd cronjob could detect that we are at the end of the month, reinitialise the monthly leaderboard and send a message of who is the leader. This could also be done by a 3rd job instead.

This solution seem way simpler and more practical. If we want to do something more complex I would want us to create an telegram app, but I don't think that's needed as our little quizz can work well, with a very small set of features.

rakici · 2024-11-10T10:07:18Z

+            git init
+            git remote add origin https://${{ secrets.REPORTQUIZ_DATA_REPO_TOKEN }}@github.com/osscameroon/reportquiz_data.git
+            git pull origin main
+            git switch -c main


This is going to cause a bunch of issues, as the global-github-actions main branch could conflict and create some confusion with the reportquiz_data repo. I recommend using actions/checkout@v4 and clone that repo on a separate path https://github.com/actions/checkout?tab=readme-ov-file#checkout-multiple-repos-side-by-side

Another question, does the reportquiz_data repo already exist ? If not can we create it, before this PR is merged ?

This is going to cause a bunch of issues, as the global-github-actions main branch could conflict and create some confusion with the reportquiz_data repo.

The repo is initialized at the level of the bot folder. I don't think that it can create confusions. 🤔

Another question, does the reportquiz_data repo already exist ? If not can we create it, before this PR is merged ?

It require a dedicated repo for data submissions.

The idea behind is to make the quiz data open source and available for anyone interested to play with it.

And there is also the fact that, I prefer the database and the code stay separate.

rakici · 2024-11-10T10:10:48Z

+          source main.sh
+          # update the quiz database
+          fetch_quiz_user_answers


We seem to be using main.sh as a library here and in several other places. Can we a separate script for each of the function here, or rename main.sh lib.sh, I don't like that last approach but if it that's what you prefer go for it.

rakici · 2024-11-10T10:17:41Z

+5. Setting up your bash terminal
+
+   ```bash
+   source main.sh


It would be nice to avoid loading this in people's local terminal. That's why I recommend having a script for each of the function you want to use.

rakici · 2024-11-11T18:05:00Z

+   . start_quiz_competition
+   # Fetch user answers of the qizzes.
+   . fetch_quiz_users_answers
+   # Stop the current quiz competition and submit resutls.
+   . stop_quiz_competition


Again that doesn't look like a great way to run scripts, on a local terminal session.

Can we a separate script for each of the function here, or rename main.sh lib.sh

@rakici
It's to be used in a script.
I used it directly in the github action, because I don't see the utility to create a script to run one command

Should I rename it lib.sh ?

rakici · 2024-11-11T18:12:31Z

@@ -0,0 +1,9 @@
+# compute the user total score from more than one quiz_data
+compute_total_score(){
+    jq -s --argjson total_price 30000 'map((.correct_answers | map(.)) as $answers | (.user_answers // []) | map(.score=(.options_ids | map(select($answers[.] == "true")) | length)-(.options_ids | map(select($answers[.] == "false")) | length) | del(.options_ids))) | add | group_by(.first_name) | map((map(.score) | add) as $total_score | .[0] | del(.score) | .total_score=$total_score) | group_by(-.total_score) | to_entries | map(.price=$total_price/((.key+1)*2))' $@


i agree with @Sanix-Darker , this is not easy to maintain. I am pretty sure you can split it into smaller sequential and logical chunks so that other people can understand it quicker. Performance is not an issue here, since this run in a runner on a cron, no one is waiting for this answer to be sent in less than 10 milliseconds. I am more for privileging readability/maintainability over speed here.

You also could write it in multiple lines for example, add some single line comments to your script, use re-usable variables etc which I think are supported by jq.

rakici · 2024-11-11T18:23:11Z

+
+# genarate a textual leaderboard with price
+generate_leaderboard(){
+    jq -r '["1st", "2nd", "3rd", "4th", "5th"] as $rangs | ["🥇", "🥈", "🥉", "🎖️"] as $badges | map(.key as $key | [([$rangs[.key], "PLACE", "-", .price, "yotas"] | join(" ")), (.value | map([$badges[$key], .first_name, "("+"@"+.username+")"] | join(" ")) | sort)] | flatten | join("\\n")) | join("\\n\\n")' $1


rakici · 2024-11-11T18:37:41Z

+      - name: Check the status of the competition
+        shell: bash
+        run: |
+          source main.sh
+          # start a new competition each monday morning
+          if [ $(date +%a_%p) == Mon_AM ]; then start_quiz_competition; fi
+          # end the competition each sunday evening
+          if [ $(date +%a_%p) == Sun_PM ]; then
+            stop_quiz_competition
+            touch /tmp/skip_send_quiz
+          fi


The competition thing is nice, but is not necessary for this initial version, and I think it makes the bot a little more complicated than it should have been.

it makes the bot a little more complicated than it should have been.

I don't understand, could you explain?

rakici · 2024-11-11T18:38:14Z

+# update quiz data based on user answers submission
+fetch_quiz_user_answers(){
+    # require a ongoing competition
+    has_ongoing_competition || return 0


We don't need that.

I don't understand, you means has_ongoing_competition || return 0 ?

pythonbrad · 2024-11-11T22:03:38Z

Based on your reviews, i begin to think that my code looks complex and not easy to read and maintain.

In brief, I think bash was not a good choice for this project.

Maybe another scripting programming language like python will be great. 🤔

Since, i won't maintain it anymore, i will simplify close the pr. Feel free to take inspiration from it for your next proposal bot. 🙂

pythonbrad · 2025-03-15T08:01:24Z

1- We will move to an unanonymous poll

2- We will add a new cron job that will request, answers from our users, and update 2 leaderboards.
   - An all time leaderboard that aggregates scores for each user since the beginning of time

   - A monthly leaderboard that gives us who leads and won for a single month. And is reinitialised each months. 
3- That 2nd cronjob could detect that we are at the end of the month, reinitialise the monthly leaderboard and send a message of who is the leader. This could also be done by a 3rd job instead.

@elhmn, it's what the bot is doing.

Please, Can you check the pr again? Or clarify what is missing?

pythonbrad and others added 3 commits August 22, 2024 21:03

refactor(report_quiz): structure the code

91c5c1f

patch(report_quiz): save the quiz data

ab20a34

In saving the quiz data, we will be able to retrieve the submitted answers and build the leaderboard

pythonbrad force-pushed the reportquiz_leaderboard branch 2 times, most recently from 3280748 to 48e548f Compare August 24, 2024 10:34

chore(report_quiz): prepare next tasks

d3cc3da

pythonbrad force-pushed the reportquiz_leaderboard branch from 48e548f to d3cc3da Compare August 24, 2024 10:53

pythonbrad added 6 commits August 24, 2024 11:19

fix(report_quiz): fix problem of null user_answers

6573803

When a quiz_data is created, the user_answers field is not present and the addition can't proceed. As change, we will use an empty value in this case

feat(report_quiz): generate leaderboard for a round

d0dc642

- added the possibility to generate and send the leaderboard

patch(report_quiz): stop a quiz round and submit result

e6549f3

Stop all the active quiz, submit the result and archive the quiz data

feat(report_quiz): announce the winners

b8d6fca

Added a friendly announcement message. Additonnal change - moved the price calculation from `generate_leaderboard` to `compute_total_score`. - fixed issue when a quiz data contains no user answer submission during the score computation

feat(report_quiz): announce a new quiz competition

f755558

ci(report_quiz): update the workflow

f861494

- Added the start new competition action - Added the fetching of the quiz user answers - Added the backup of the data using git - Added the ending of a current competition

pythonbrad marked this pull request as ready for review August 26, 2024 19:32

HanslettTheDev and others added 10 commits October 1, 2024 12:52

doc(report_quiz): added a README.md (#1)

e360a2e

Added a documentation for the report quiz bot

fix(report_quiz): fixed the access to file

c277770

chore(report_quiz): improved error management

0546ad2

ci(reportquiz): update ci

7430c4a

chore(reportquiz): minor update

41ec5da

fix(reportquiz): error when attempt to close a closed quiz

76e1c25

The close competition fail during the closing of a already closed quiz. Since the status of this process is not important, we will ignore it. Like that the closing of the competition can proceed.

ci(reportquiz): update the ci

44b06e3

fix(reportquiz): stop competition fail when no results

29dd9e5

This commit handles the case when there is no results during the closing of the competition.

ci(reportquiz): fix update database

ab5f784

chore(reportquiz): fix leaderboard formatting

5816b5b

Sanix-Darker reviewed Oct 6, 2024

View reviewed changes

pythonbrad changed the title ~~feat(report_bot): add a leaderboard~~ feat(report_quiz): add a leaderboard Nov 7, 2024

rakici requested changes Nov 11, 2024

View reviewed changes

pythonbrad added 2 commits November 11, 2024 22:43

Fixed typos in functions.sh

954c1e3

Update action name in report_a_quiz.yaml

0d4b74d

pythonbrad closed this Nov 11, 2024

pythonbrad reopened this Jul 19, 2026

pythonbrad added 7 commits July 20, 2026 03:51

fix compatibility issue with the current quizapi version

b53dcae

Update secret name in workflow

1944497

Update report_a_quiz.yaml

539d417

Add workflow_dispatch trigger to report_a_quiz.yaml

2d70bd5

Enable debugging in main.sh

6f84609

Update environment variable handling and git config

b98aa5c

fix jq code

1ef97c0

pythonbrad force-pushed the reportquiz_leaderboard branch from eba596d to 1ef97c0 Compare July 20, 2026 12:11

pythonbrad added 2 commits July 20, 2026 13:13

Remove DEBUG variable from main.sh

c68dfb8

update quiz topics

4339c58

	jq -r '["1st", "2nd", "3rd", "4th", "5th"] as $rangs \| ["🥇", "🥈", "🥉", "🎖️"] as $badges \| map(.key as $key \| [([$rangs[.key], "PLACE", "-", .price, "yotas"] \| join(" ")), (.value \| map([$badges[$key], .first_name, "("+"@"+.username+")"] \| join(" ")) \| sort)] \| flatten \| join("\\n")) \| join("\\n\\n")' $1
	jq -r '["1st", "2nd", "3rd", "4th", "5th"] as $ranks \| ["🥇", "🥈", "🥉", "🎖️"] as $badges \| map(.key as $key \| [([$ranks[.key], "PLACE", "-", .price, "yotas"] \| join(" ")), (.value \| map([$badges[$key], .first_name, "("+"@"+.username+")"] \| join(" ")) \| sort)] \| flatten \| join("\\n")) \| join("\\n\\n")' $1

Uh oh!

Conversation

pythonbrad commented Aug 23, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Related issues

Change

TODO

Notice

Useful link

Uh oh!

pythonbrad commented Aug 26, 2024

Uh oh!

Sanix-Darker left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pythonbrad Oct 6, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Sanix-Darker Oct 6, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pythonbrad commented Oct 8, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

pythonbrad commented Oct 15, 2024

Uh oh!

rakici left a comment • edited by elhmn Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pythonbrad Nov 11, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pythonbrad Nov 11, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

pythonbrad Nov 11, 2024 • edited Loading Uh oh! There was an error while loading. Please reload this page.

pythonbrad commented Aug 23, 2024 •

edited

Loading

pythonbrad Oct 6, 2024 •

edited

Loading

Sanix-Darker Oct 6, 2024 •

edited

Loading

pythonbrad commented Oct 8, 2024 •

edited

Loading

rakici left a comment •

edited by elhmn

Loading

pythonbrad Nov 11, 2024 •

edited

Loading

pythonbrad Nov 11, 2024 •

edited

Loading

pythonbrad Nov 11, 2024 •

edited

Loading