Business Class Seatmap Generator w/ Amadeus API

CLI utility for browsing Amadeus seat-map availability across a set of long-haul trips. It pulls live data from the Amadeus Self-Service APIs and caches each day's responses under data/<YYYYMMDD> so you can re-render the output without hitting the API again.

Highlights

• Builds ASCII seat maps with wide-character awareness so layouts stay aligned even when using emoji markers.
• Highlights the lowest fare per route with a green border so deals stand out immediately.
• Caches availability, pricing, and seatmap responses by run date (data/<YYYYMMDD>) for quick offline replays.
• Prints a week-by-week grid plus per-day window-seat summaries with optional pricing pulled from flight offers.
• Generates calendar heatmaps (per route and round-trip) to make fare trends obvious when scanning many dates.

Requirements

• Python 3.11+ (tested locally with 3.11)
• pip install -r requirements.txt
• Amadeus Self-Service API credentials (test and/or production)
• Pillow (for PNG export; included in requirements.txt so GitHub Actions can commit PNGs)

Local setup

python -m venv .venv
source .venv/bin/activate
pip install --upgrade pip
pip install -r requirements.txt
cp .env.template .env

Fill in the .env file with the credentials from your Amadeus developer account:

TEST_AMADEUS_CLIENT_ID=...
TEST_AMADEUS_CLIENT_SECRET=...
AMADEUS_CLIENT_ID=...
AMADEUS_CLIENT_SECRET=...

The script reads these via python-dotenv, so the .env file only needs to exist in the repo root.

Configuration

Most runtime knobs live in config.py, so you rarely need to edit flight_search.py directly. Notable settings:

• ENVIRONMENT picks the Amadeus host and must be either production or test for the fetch scripts.
• TRAVEL_WINDOWS lists the routes/date ranges that should be fetched and rendered.
• FLIGHT_SEARCH_FILTERS stores the request arguments (travel_class, non_stop, included_airline_codes, etc.) passed to the Amadeus flight_offers_search endpoint.
• Visual toggles such as SEATMAP_OUTPUT_STYLE, SHOW_SEATMAP_PRICE, HEATMAP_EMPHASIS_STYLES, STATUS_SYMBOLS, and WINDOW_AVAILABLE_SYMBOL control how the ASCII/emoji output looks.
• Currency/price decoration (CURRENCY_SYMBOLS, BORDER_COLORS, etc.) reference the semantic tokens defined in colors.py; extend colors.TOKEN_MAP if you need custom ANSI sequences.

Automation & publishing

• Seatmap pipeline – Manually runnable via GitHub Actions (Run Seatmap Pipeline). It installs deps, runs run_pipeline.py, and commits the refreshed PNGs in docs/ back to the repo.
• Static site deploy – Deploy static content to Pages publishes the contents of docs/ to GitHub Pages at https://robertzaufall.github.io/amadeus-seatmap-bc/. It triggers on master when files in docs/**/*.png or docs/**/*.html change, and can also be dispatched manually.
• Local runs still work the same way; the automation just keeps the hosted artifacts current.

Travel windows & filters

Define the routes and date ranges you care about by editing the TRAVEL_WINDOWS list inside config.py. Each window is inclusive, so the scripts request seat maps for every day in the range. The first two windows drive the combined round-trip heatmaps (outbound vs. return).

Flight-offer search filters live in config.FLIGHT_SEARCH_FILTERS, so you can tweak cabin class, airline, or connection rules without touching the main script. Any key/value pairs in this dict get splatted into the SeatMaps.fetch call.

Workflow

1. Configure config.py (set ENVIRONMENT to test or production, define TRAVEL_WINDOWS, adjust FLIGHT_SEARCH_FILTERS).
2. Load environment variables from .env (see above).
3. Fetch and cache data for today's run:

   python get_availability.py       # builds travel_dates.json + availability_responses.json
   python get_prices_oneway.py      # per-leg pricing -> pricing_responses_*oneway*.json
   python get_prices_return.py      # round-trip pricing -> pricing_responses*.json
   python get_seatmaps.py           # seat-map payloads -> seatmap_responses.json

   Each script reuses the existing files for today's date unless you flip refresh_data = True near the top of the file.
4. Render the reports from the cached data:

   python flight_search.py

One-off seatmap lookup

Use flight_offer_seatmaps.py to fetch seatmaps for a single flight offer without running the full pipeline. It searches /v2/shopping/flight-offers, requests seatmaps for the result, and caches everything under data/flights/<ENV>-<FROM>-<TO>-<DATE>-<TIME>-<CLASS>-<AIRLINE>-<CURRENCY>/.

Example (queries business class; omit --class to fetch both cabins):

python flight_offer_seatmaps.py \
  --date 2024-12-05 \
  --time 22:25 \
  --from BKK \
  --to FRA \
  --airline TG \
  --class BUSINESS \
  --currency EUR

Notes:

• Seatmaps print to the terminal and save as seatmaps.png plus JSON files (flight_offers.json, seatmaps.json, metadata.json) in the cache folder.
• Seatmaps are fetched when the flight-offer search yields exactly one matching offer per requested cabin; tweak --max-offers or specify --class if you get multiple hits.
• Cached responses stay fresh for 4 hours by default (--cache-ttl-hours 0 or --force-refresh to bypass); diffs for seatmaps write to seatmaps.diff when content changes.
• Use --environment test|production to select the Amadeus host; credentials are read from .env just like the main pipeline.

Data directory layout

All generated API responses are timestamped by run date: data/<YYYYMMDD>/. Common files include:

• travel_dates.json: cleaned list of dates per travel window (from availability checks).
• availability_responses.json/unavailable_flights.json: raw availability responses and filtered-out flights.
• pricing_responses_oneway.json and pricing_responses_simple_oneway.json: leg-level pricing used for labels and price lookups.
• pricing_responses.json and pricing_responses_simple.json: round-trip pricing used to build seatmap requests.
• seatmap_responses.json: raw seatmap payloads consumed by flight_search.py.

If you want to re-render a prior run, copy its dated folder to today's date inside data/ so flight_search.py can find the JSON files.

Output details

• Weekly seat-map grid – One block per day, grouped by week. Missing data shows a NO DATA placeholder, and the absolute lowest fare per route gets a green border plus a rounded price label at the bottom of the block.
• Route availability boxes – After the grid, every route receives a bordered box that lists the available window seats per date, prefixed with a relative fare symbol and paired with a mini calendar heatmap for additional visual context.
• Round-trip price heatmaps – If you provide at least two travel windows (outbound + inbound), the script prints two combined matrices: one based on window-seat fares only and another that considers any price returned by the offer search.

Contributing

No special tooling is required beyond standard linting/formatting for Python scripts. Please keep fixtures anonymized and avoid committing sensitive traveler data.

License

This project is released under the MIT License. See LICENSE for the full text.