GGThedandClaude Opus 5 5ecea55f55 fix(ci): rendre la chaine d'integration reellement verifiante
Les quatre jobs existaient ; aucun ne verifiait ce qu'il annoncait.

security-audit
  `pip-audit --require-hashes --no-deps || pip-audit`. L'etape
  d'installation ne posait que pip-audit, et aucune des deux formes ne
  nommait le fichier d'exigences : le repli auditait l'environnement du
  runner, qui ne contenait que pip-audit lui-meme. Le job passait au vert
  sans avoir examine une seule dependance de l'application. Remplace par
  `pip-audit -r requirements.txt`.

security-scan
  Appelait `python security_scan.py`, alors que le fichier se trouve dans
  app/supporting_scripts/. En echec a chaque execution depuis le
  deplacement du fichier. Trois autres defauts sont apparus en le faisant
  tourner :
    - la CI passe --skip-http, un argument que l'argparse du script
      n'acceptait pas : sortie en erreur 2 meme avec le bon chemin.
    - check_dependencies lisait data['dependencies'] comme la liste des
      vulnerabilites. Ce tableau liste en realite TOUTES les dependances,
      chacune portant un champ vulns vide si le paquet est sain. Les ~45
      paquets installes etaient donc signales vulnerables a chaque
      execution. Le filtrage se fait desormais sur vulns non vide.
    - check_flask_config interceptait son exception et renvoyait quand
      meme all_ok : ne pas reussir a charger l'application comptait comme
      un controle reussi. La section la plus importante du rapport n'avait
      jamais tourne. Elle renvoie desormais False, et l'import fonctionne
      grace a l'ajout de la racine du projet dans sys.path.
    - la banniere en caracteres semi-graphiques faisait planter le script
      sur une console Windows en cp1252, la plateforme meme du projet.
      Passee en ASCII.

lint
  Ruff n'avait aucun fichier de configuration : le job tournait sur le jeu
  de regles par defaut. La configuration vit maintenant dans pyproject.toml.
  `ruff format --check` est retire pour l'instant : la base n'ayant jamais
  ete formatee, il echouerait sur 62 fichiers sur 64 pour des raisons
  etrangeres a la correction. Reformatage puis application : QUA-002.

test
  Un `echo` protege par continue-on-error : le job annoncait un succes
  sans rien executer. Il lance desormais pytest avec couverture, et bloque.

permissions: contents: read au niveau du workflow, aucune etape n'ecrivant
dans le depot.

Deploiement Gitea
  actions/checkout@v7 n'existe pas (derniere majeure : v5) : le workflow
  echouait des sa premiere etape. Ramene a v4.
  Le miroir lftp poussait l'integralite de l'arbre de travail, dont
  clear_db.py -- un script qui vide toutes les tables et recree
  admin/password -- vers le noeud de production. Liste d'exclusions ajoutee.
  --delete reste volontairement absent : les contrats televerses, les
  journaux et le .env du serveur vivent sous la racine de deploiement et
  sont absents du depot ; les supprimer detruirait des donnees.

Le workflow de deploiement n'a pas pu etre execute depuis ici : la
syntaxe lftp reste a valider lors du prochain deploiement manuel.

Co-Authored-By: Claude Opus 5 <[email protected]>
2026-08-07 19:47:24 -04:00
2026-08-06 14:46:51 -04:00
2026-07-30 09:52:32 -04:00
2026-08-03 16:43:46 -04:00

Plateforme centralisée de tryouts

Security Configuration

Required Environment Variables

Before deploying, create a .env file which integrates everything in the .env.exemple. Ensure you follow the comments of the exemple if you are to use this tool in production.

Security Features Implemented

  • Rate Limiting: Login endpoint limited to 10 requests per minute to prevent brute-force attacks
  • Secure Session Cookies: HTTPSOnly, SameSite=Lax, and Secure flags enabled
  • CSRF Protection: Enabled by default on all forms
  • HTTPS Enforcement: Automatic redirect to HTTPS in production
  • Security Headers: X-Frame-Options, X-Content-Type-Options, Content-Security-Policy, HSTS
  • Open Redirect Prevention: URL validation on login redirect
  • Authorization Checks: Proper ownership validation on all sensitive operations
  • nginx: reverse-proxy and load balancer
  • Waitress WSGI: Production ready WSGI

When true in .env:

  • Forces HTTPS only
  • Forcer secure cookies

App details

Code

  • Full python backend using flask
  • statics are pure HTML and CSS
  • Some js to add logic to styling and showing certain pages/cards

Functionalities

  • User base with sign-ins: Forces users to create an account and register pertinent information for tryouts and teams. The admin can attribute them a role.
  • User-Role-Based Permissions: admin - full acces, coach/manager - access to team management, player - views what he is registered in (no management), scout - view only
  • Tryout Management: manage internal tryout teams, organise internal tryouts matches (3 formats, team vs team, PvP, scrim). Coaches can Evaluate players based on 10 criteria
  • Team Management: manage teams for the season, create matches and practices. When planning a practice there will be a calendar showing player availabitlities slots to help chose a time
  • Coach and Player Availabilities: Allow better planning for the coaches, and for players to book One on Ones with their coach.
  • Player Notes: Coaches can give notes to their players. The players will see them and there is a history which keeps the most recent notes.
  • Team Notes: Coaches can give notes to their teams, where all players from that team can see the note.
  • One on One: Players can request a One on One meeting with their coach. This sends a discord dm to the coach to accept or refuse. The player is then notified of the response.
  • Availabilities: Allow players and coach to enter the moments they are available. Allows for easier practice setup and One on One planning.

Discord Integration

The application supports sending Discord direct messages to coaches when players request One on One sessions, when matches/tryouts/practices are created and a player is in it, and the players get match reminders 24h before a match.

When sending a One on One request, the coach can accept via the platform or react to the discord message to answer the booking request. Same thing with matches and practices, the players can react or answer on the platform.

Setup Instructions

1. Create a Discord Bot (Not needed for UdeS user, the bot already exists)

  1. Go to the Discord Developer Portal
  2. Create a new application
  3. Go to the "Bot" tab and create a bot user
  4. Copy the bot token - this will be your DISCORD_BOT_TOKEN
  5. Enable the "Message Content Intent" under Privileged Gateway Intents (required for sending messages)

2. Add Bot to your server

For the bot to send DMs:

  1. Each user must have the bot added to their Discord server OR be friends with the bot
  2. Users need to add their Discord User ID to their profile:
    • Enable Developer Mode in Discord (User Settings → Advanced → Developer Mode)
    • Right-click on their profile → Copy ID
    • Enter this numeric ID in the "Discord User ID" field in their profile settings

How It Works

When a player submits a One on One request:

  1. The system checks if the coach has a Discord User ID configured
  2. If configured, a direct message is sent to the coach via the Discord bot

Message Format

The Discord DM includes:

  • Player name
  • Team name
  • Requested date and time slot
  • Discussion points (if provided)
  • Link to the application for approval/rejection
  • Two provided reactions to accept or refuse via discord
S
Description
No description provided
Readme
2.3 MiB
Languages
Python 67.9%
HTML 27%
CSS 3.5%
JavaScript 1.6%