cfjjb-planning

Build a per-academy Excel fight-planning for any CFJJB competition. The scraper drives the CFJJB site with a headless Chromium browser, harvests the participants / planning / brackets tabs straight from the rendered DOM, and joins the three views on (category, page, mat) so every athlete ends up with their actual date, time, tatami and N/M tableau number.

Output

The builder produces one .xlsx per academy with these columns:

Participant	Équipe	Catégorie	Date	Heure	Tapis	Tableaux	Note
Jane DOE	Example Academy	BLANCHE - ADULTE - FEMME - LEVE	2026-04-26	11:25	Tatamis 6	7/9
John DOE	Example Academy	MARRON - MASTER 3/4 - HOMME - PESADISSIMO	2026-04-25	14:17	Tatamis 5	1/1	P.Seule

Rows are colour-tinted by belt, sorted by date → time → mat, and saved to output/planning_<academy>.xlsx.

Install

A Python 3.11+ virtual environment is recommended.

git clone https://github.com/SebastienGaspar/cfjjb-planning.git
cd cfjjb-planning

python3 -m venv .venv
source .venv/bin/activate            # on Windows PowerShell: .venv\Scripts\Activate.ps1

pip install --upgrade pip
pip install -r requirements.txt
playwright install chromium          # one-time: grabs the bundled browser

Run

Two steps: extract, then build. The extractor writes all intermediate JSON into output/; the builder reads from there.

1. Extract

--competition is required and accepts either the numeric id or any cfjjb.com URL that carries it:

# Bare id — writes to output/941/
python3 extract_cfjjb.py --competition 941

# Or any of these URL forms
python3 extract_cfjjb.py --competition https://cfjjb.com/competitions/signup/info/941
python3 extract_cfjjb.py --competition "https://cfjjb.com/competition-group/470?tab=participants&id=941"

# Only refresh one tab
python3 extract_cfjjb.py --competition 941 --tabs brackets

# Watch the browser work
python3 extract_cfjjb.py --competition 941 --headed

# Override the output location (default: output/<competition-id>/)
python3 extract_cfjjb.py --competition 941 --out-dir output/orleans_gi

Each run writes into its own output/<competition-id>/ folder so that multiple competitions of the same event (Gi, No-Gi, Kids Gi, Kids No-Gi) can be extracted without overwriting each other.

After this step, output/<competition-id>/ contains:

File	Source
participants_by_team.json	the "Liste des Participants" tab, ?by_team
planning.json	the "Le Planning" tab
brackets_by_category.json	the "Tableaux de Combats" tab (aggregated per category)
brackets_sections_raw.json	one record per bracket section, with outer_html for debugging
meta.json	the competition id, its page-visible name, and a derived short_label (GI, NO GI, Kids GI, Kids NO GI)
<tab>.html and <tab>.png	rendered snapshot of each tab for debugging

2. Build the planning for your academy

# xlsx (default) — reads from output/941/
python3 build_planning_xlsx.py --academy "Example Academy" --input-dir output/941

# Same thing as a PDF
python3 build_planning_xlsx.py --academy "Example Academy" --input-dir output/941 --format pdf

# Pick the output basename (extension added automatically)
python3 build_planning_xlsx.py --academy "Example Academy" --input-dir output/941 --name orleans_gi

The academy match is a substring, accent- and case-insensitive, so "example" or "EXAMPLE ACADEMY" work equally well. Output lands in <first-input-dir>/planning_<academy>.<xlsx|pdf> by default; override with --name <stem> (placed under the first input dir) or --out <path>.

A --debug mode prints a one-line trace per athlete showing which priority matched and why, with --only "name-substring" to focus it:

python3 build_planning_xlsx.py --academy "Example Academy" --input-dir output/941 --debug --only "DOE"

2b. Merge several competitions of the same event

Events on CFJJB are commonly split across four competition ids — Gi, No-Gi, Kids Gi, Kids No-Gi. After extracting each one into its own folder, pass them all to the builder in a single run:

python3 extract_cfjjb.py --competition 941      # Gi        -> output/941/
python3 extract_cfjjb.py --competition 942      # No-Gi     -> output/942/
python3 extract_cfjjb.py --competition 943      # Kids Gi   -> output/943/
python3 extract_cfjjb.py --competition 944      # Kids No-Gi -> output/944/

python3 build_planning_xlsx.py \
    --academy "Example Academy" \
    --input-dir output/941 output/942 output/943 output/944 \
    --format pdf --name orleans_2026

The join runs independently per directory (category/page/mat keys only make sense within one competition), then every row is merged and re-sorted by date → time → mat. A Compétition column is added to the output whenever more than one --input-dir is given.

The label shown in that column is derived, in order of precedence:

1. An explicit override via --input-dir path=LABEL, e.g. --input-dir output/941=GI output/942="NO GI".
2. The short_label written into meta.json by the extractor (derived from the page-visible competition name: Kids prefix if "KIDS" appears, suffix NO GI if "NO GI" / "NOGI" appears, else GI).
3. The directory basename (fallback for legacy extractions without meta.json).

Helpers

Two tiny inspectors for the extracted JSON, handy for debugging or sanity-checking a run:

python3 check_planning.py                           # list every category + slot count
python3 check_planning.py "Bleue - Adulte - Homme - Medio"

python3 check_brackets.py                           # summary by category
python3 check_brackets.py "Jane DOE"                # find a fighter
python3 check_brackets.py "Bleue - Adulte"          # find a category

How it works (short version)

CFJJB is a Vue SPA with two particularly fun behaviours:

1. Lazy mount — bracket pages and planning rows only render when scrolled into view.
2. Lazy unmount — once they leave the viewport, Vue replaces them with a placeholder again.

A one-shot scrape therefore only sees what's currently on screen. This project instead installs a MutationObserver inside the page that caches every <section class="wpage"> and <li data-cate-fight> the instant it mounts (keyed so the entry is kept forever), then scrolls the whole page top-to-bottom to force every lazy element to briefly materialise. The cache is read at the end, guaranteeing complete data.

Files

extract_cfjjb.py             # playwright scraper — writes the JSONs under output/
build_planning_xlsx.py       # reads the JSONs, writes the Excel
check_planning.py            # inspect output/planning.json
check_brackets.py            # inspect output/brackets_by_category.json
requirements.txt             # playwright + openpyxl

License

MIT — see LICENSE.