Site Weight Journey Runner

This project measures the network weight of a real user journey on modern SPA + SSR websites (like Nuxt). It was built to address a limitation of GreenIT-Analysis: in SPA/SSR setups, forcing full SSR reloads on every page re-downloads JS/CSS and skews results. Instead, this tool captures the true cost of client-side navigation.

Goals:

• Reproduce a realistic user journey with configurable actions (clicks, waits, multi-action steps).
• Measure request count and transfer sizes per step within a controlled time window.
• Provide two CSV outputs: a human-friendly summary and a technical report, plus a JSON + schema for analysis.
• Stay compatible with SPA navigation (URL changes, late async calls, prefetch behavior).

This repo contains a small Playwright-based CLI that measures network weight for a user journey. It executes a journey described in a journey.js file you provide.

Install

npm install

Usage

Pass either a folder (the CLI will look for journey.js inside it) or a direct path to a journey.js file.

npx site-weight ./journey.js

npx site-weight ./example.com/

Journey Config (journey.js)

Export a default object. Use steps for detailed journeys. If steps is empty or missing, the script uses urls instead.

Full steps example

export default {
  name: "Example journey",
  startUrl: "https://www.example.com/",
  mode: "per_step",
  settle: {
    waitUntil: "networkidle",
    extraMs: 1200,
  },
  scroll: {
    enabled: true,
    mode: "page",
    stepPx: 800,
    pauseMs: 250,
    maxMs: 12000,
    backToTop: false,
  },
  steps: [
    {
      name: "Accueil SSR",
      action: "noop",
    },
    {
      name: "Aller a la programmation",
      action: "click",
      selector: "a[href*='programmation'], a[href*='spectacles']",
    },
    {
      name: "Ouvrir une fiche spectacle",
      action: "click",
      selector: "main a[href*='spectacle'], main a[href*='spectacles']",
    },
  ],
};

Simple urls example

urls can contain absolute or relative URLs. When urls is used, the script tries to click a link whose href matches the target URL (absolute match first, then relative, with a trailing slash tolerance). If no link is found, it falls back to window.location.href and logs a warning. In this mode, startUrl is ignored and each URL becomes a step named after the URL. Relative URLs require startUrl to be set.

export default {
  name: "Simple journey",
  urls: [
    "https://www.example.com/",
    "https://www.example.com/programme/",
  ],
};

Multi-actions step example

Use actions to run multiple actions in a single step. Actions run in order, with no settle between them; settle happens once at the end of the step. The report action field concatenates the action names (e.g. click>click) and page_url captures the final URL after the step finishes.

export default {
  name: "Example journey",
  startUrl: "https://www.example.com/",
  steps: [
    {
      name: "Aller a la programmation",
      actions: [
        {
          action: "click",
          selector: "button[aria-controls='menu-id-api-menus-25']",
        },
        {
          action: "click",
          selector: "a[href='/programme/saison-25-26']",
        },
      ],
    },
  ],
};

Output

Reports are written to a timestamped folder inside the journey.js directory: /<journey-folder>/reports-<timestamp>/. Filenames include a run id derived from the journey name and timestamp. request_count is computed within a step window (start to end), and window_duration_ms reports that window length. Each row also includes page_url, the current page URL after the step finishes. Two CSV files are produced: <runId>.csv (human-friendly) and technical_<runId>.csv (full technical report).

EcoIndex sources

We compute EcoIndex using the official quantiles and formula from the EcoIndex Python project, and the public methodology description:

• Quantiles: https://raw.githubusercontent.com/cnumr/EcoIndex_python/main/components/ecoindex/data/quantiles.py
• Grades: https://raw.githubusercontent.com/cnumr/EcoIndex_python/main/components/ecoindex/data/grades.py
• Formula: https://raw.githubusercontent.com/cnumr/EcoIndex_python/main/components/ecoindex/compute/ecoindex.py
• Methodology: https://www.ecoindex.fr/comment-ca-marche/

Official references:

• EcoIndex: https://www.ecoindex.fr/
• GreenIT-Analysis: https://github.com/cnumr/GreenIT-Analysis
• GreenIT DOM count method: https://raw.githubusercontent.com/cnumr/GreenIT-Analysis/master/script/analyseFrame.js