Cacher is the code snippet organizer for pro developers

1. サイトAからサイトBに対して、https://siteb.hoge?id=fuga123のようなパラメータをつけて遷移させる
　
2. サイトB側で、URLパラメータからid=を抽出してCookieに保存
　
3. サイトB側で動的にimg要素(src属性値にサイトAの計測APIのURLにパラメータを付与したもの)を生成する
　
44. 手順3の時点でサイトAに対してリクエストが発生し、サイトA経由でサイトBに着地した人数を計測できる

# dbt seed — Guide pédagogique

## Ce que `dbt seed` fait

1. **Parse** les fichiers YAML pour connaître la configuration (schema, types, colonnes)
2. **Pour chaque fichier CSV** dans `seeds/` :
   - Supprime la table si elle existe (DROP)
   - Crée la table avec les types définis (CREATE)
   - Insère les données du CSV (INSERT)

## Ce que `dbt seed` ne fait PAS

| Action | Commande |
|--------|----------|
| Exécuter les models | `dbt run` |
| Exécuter les tests | `dbt test` |
| Compiler le SQL 

# Cascade de configuration dbt

## Principe

dbt applique les configurations par héritage : le niveau le plus spécifique gagne toujours.

## Niveaux de priorité

```
┌─────────────────────────────────────────────────────────────┐
│  1. config() dans le fichier SQL        ← Plus spécifique   │
├─────────────────────────────────────────────────────────────┤
│  2. Fichier YAML (_seeds.yml, _models.yml)                  │
├─────────────────────────────────────────────────────────────┤
│  3. dbt_proj

# Guidelines de documentation dbt

## Structure standard d'une description

```yaml
description: |
  # Titre du model/seed
  
  Description courte en une phrase.
  
  ## Grain
  Une ligne par [entité].
  
  ## Sources
  - source_1
  - source_2
  
  ## Notes
  Informations importantes, cas particuliers, hypothèses.
```

## Métadonnées recommandées

```yaml
meta:
  owner: data-team@company.com     # Responsable du model
  tier: gold                       # Niveau de qualité (bronze/silver/gold)
``

# Structure type : _seeds.yml

> 📖 Pour les conventions de documentation (descriptions, métadonnées, checklist), voir **dbt-documentation-guidelines.md**

## Syntaxe complète

```yaml
version: 2

seeds:
  - name: nom_du_seed               # Doit correspondre au nom du fichier CSV (sans .csv)
    description: "Description"      # Documentation (supporte le Markdown; pour du multiligne, utiliser le pipe: | ou la syntaxe YAML avancée (voir fichier doc guidelines))
    
    meta:                    

# dbt CLI Cheatsheet

## Configuration & Debug

| Commande | Effet |
|----------|-------|
| `dbt debug` | Valide la connexion et la configuration |
| `dbt deps` | Installe les packages du `packages.yml` |
| `dbt clean` | Supprime `target/`, `dbt_packages/`, `logs/` |

## Exécution

| Commande | Effet |
|----------|-------|
| `dbt seed` | Charge les fichiers CSV dans le warehouse |
| `dbt run` | Exécute tous les modèles |
| `dbt test` | Exécute tous les tests |
| `dbt build` | Fait `seed` + `run`

/**
 * @param {number[][]} grid
 * @param {number} k
 * @return {number}
 *
 * This solution uses a DP-by-value technique:
 * - Normal moves (right/down) cost grid[i][j]
 * - Teleports cost 0 but can only go to cells with value <= current cell
 * - We simulate using t = 0..k teleports
 *
 * Key idea:
 *   For each teleport count t, we:
 *     1. Start from the previous DP layer (t-1 teleports)
 *     2. Apply a "value sweep" that propagates the best reachable cost
 *        to all cells with <= 

java --version

so this is as good start but i need to be able to injest and parse, html,markdown,json, plaintext, css,js python, jsx,tsx,ts, pdf, webp, jpeg, jpg, png , i need to be able have my notes injested from notion and obsidian, i need to store code snippets and web docs, i need to injest timestamp, vidoeos, so mp3/mp4s , and whateber else is a common codec videos are stored in.  we need to be able to scrape a sningle webpage, we need to give the  app a list of urls, or have it generate a sitemap then s

@charset "UTF-8";
/*
Thema Name: XXXXX株式会社
Description: XXXX株式会社様用のテンプレートです
Version: 1.0
Author: H.Yamada
*/

1. Create a Compute Instance:
    - Log into OCI Console
    - Navigate to Compute → Instances → Create Instance
    - Choose Ubuntu or Oracle Linux
    - Configure SSH keys for access
    - Open ports 80 (HTTP) and 443 (HTTPS) in security lists
  2. Connect and Setup Server:
  ssh -i your-key.pem ubuntu@<instance-ip>

  # Install Node.js
  curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
  sudo apt-get install -y nodejs

  # Install PM2 (process manager)
  sudo npm install -g p

/**
 * @param {number} n
 * @param {number[][]} edges
 * @return {number}
 *
 * NOTE:
 * This code performs a Dijkstra-like search using a custom MinHeap.
 * It treats every directed edge u→v with cost w as:
 *   - forward:  u → v with cost w
 *   - reverse:  v → u with cost 2w
 *
 * This is NOT the correct logic for the original problem,
 * but this cleaned version preserves your structure exactly.
 */

class MinHeap2 {
    constructor() {
        this.data = [];
    }

    size() {
        ret

<NavigationSet_Add>
    <Code><![CDATA[primary_navigation]]></Code>
    <Name><![CDATA[Primary Site Navigation Bar]]></Name>
    <Description><![CDATA[]]></Description>
    <Template>![CDATA[]]</Template>
    <Notes>#Set_Current_Time#</Notes>
    <Layout>Horizontal Drop-Down</Layout>
    <Items>
        <Item>
            <Active>true</Active>
            <Name><![CDATA[Shop All]]></Name>
            <Link type="Page" target="_self"><![CDATA[CTLG]]></Link>
        </Item>
        <I

Lecture 1 - Decisiones financieras de hogares y empresas.

Ejemplo de un proyecto = campaña de marketing.
Se asemeja a una inversion.

Conclusion una decision de negocio se trata de realmente valuar un projecto o su flujo de caja.
Esto es realmente un problema financiero.
Segundo, la valuacion no es un ejercicio subjetivo.

Finanza es la herramienta basica para resolver esas actividades/negocios (campaña de marketing, R&D project, expansion de una linea de prodccion, etc)
y tomar buenas desicion

# Your rule content
When creating code in JavaScript use functional programing best practices esp when creating new code, so it should look like this: you are given a task, you then create a scratchpad.md and use that for you thought process and computations, use it to work through the tasks in a logical manor in order unless you have a good reason to do so . after you code for a task is complete, test said code, if it passes , go to next task and repseat if not loop around and refactor , test a

from qgis.core import QgsGeometry, QgsPoint, QgsProject

layer = QgsProject.instance().mapLayersByName('test')[0]

layer.startEditing()

for f in layer.getFeatures():
    geom = f.geometry()

    # sécurité : seulement LineString
    if geom.isMultipart():
        continue

    pts = geom.constGet().points()
    dist = 0.0
    new_pts = []

    for i, p in enumerate(pts):
        if i > 0:
            dist += pts[i-1].distance(p)
        new_pts.append(QgsPoint(p.x(), p.y(),