bhoonidhi-downloader

A CLI and Python SDK for searching, saving, and downloading satellite imagery from ISRO's Bhoonidhi Earth Observation Portal.

Features

• Search by bounding box — filter by satellite, sensor, and date range.
• Named, persistent queries — every search is saved under a short slug (misty-falcon), so you can come back to it later, refresh it for new scenes, or download from it without re-querying the portal.
• Concurrent, verified downloads — fetches multiple scenes in parallel, verifies each with a SHA256, and skips anything already downloaded.
• Browse the archive — list every satellite/sensor Bhoonidhi currently supports, live from the portal.
• Session management — login once, refresh your token when it goes stale, no need to keep re-entering credentials.
• Usable as a library — every command has a matching Python function, so you can script searches/downloads directly instead of shelling out.

Installation

pip install bhoonidhi-downloader

This installs two equivalent commands — bhoonidhi-downloader and the shorter bhd. They're the same tool; use whichever you'd rather type.

Quickstart

Log in, run a search, and download what you find — three commands, start to finish:

# 1. Authenticate (prompts for username/password if omitted)
bhd auth login

# 2. Search a bounding box + date range, save the results as a named query
bhd query create 91.77 92 25.496 25.695 2023-12-01 2023-12-30 --sat Sentinel-2A --sen MSI

# 3. Download everything the query found
bhd query download misty-falcon --out ./downloads

query create prints the scenes it found in a table and tells you what slug it saved them under (here, misty-falcon — yours will be different). You don't have to download right away: come back anytime with bhd query show misty-falcon, or bhd query refresh misty-falcon to check for newly published scenes in the same area.

Command reference

Every command supports --help for its full option list. This is the short version; see the full documentation for every flag and example.

auth — session management

archive — browse what's available

query — search, save, and download

Using it as a library

Every CLI command is a thin wrapper around a plain Python function — nothing about the underlying logic depends on being invoked from a terminal. If you're scripting a bulk ingestion pipeline or wiring this into a notebook, call into bhoonidhi_downloader.core directly instead of shelling out:

from bhoonidhi_downloader.core.archive import ArchiveManager
from bhoonidhi_downloader.core.query.command import run_query_create, run_query_download
from bhoonidhi_downloader.logger import get_console
from datetime import datetime

console = get_console()

# Browse the archive programmatically
manifest = ArchiveManager().build_manifest()
print(manifest["Sentinel-2A"].keys())  # -> dict_keys(['MSI'])

# Search + save a query (same thing `bhd query create` does)
query = run_query_create(
    console,
    minx=91.77, maxx=92, miny=25.496, maxy=25.695,
    start_date=datetime(2023, 12, 1), end_date=datetime(2023, 12, 30),
    satellite="Sentinel-2A", sensor="MSI",
)

# Download everything it found
run_query_download(console, slug=query.slug, out="./downloads")

Command handlers live under core/<domain>/command.py (auth, archive, search, query, download) — each one takes a rich.console.Console (get one via bhoonidhi_downloader.logger.get_console()) and returns plain data (a QuerySchema, a list of scenes, a bool) rather than printing-and-exiting like a CLI would. The render.py modules alongside them are what turn that data into the tables you see on screen — you can skip them entirely and just work with the returned objects.

What's actually downloadable

Bhoonidhi's archive covers far more than what this tool can fetch directly. Every scene is searchable and shows up in results — but only scenes marked OpenData_DirectDownload (typically the more recent ones) can actually be pulled by query download. Anything priced or on-order shows up as metadata/planning-only for now, since there's no cart/order flow implemented yet.

The satellite/sensor list itself isn't hardcoded anywhere in this tool — bhd archive list fetches it live from the portal every time, so it's always current. Run it to see exactly what's searchable today rather than relying on a list here that would just go stale.

A couple of Bhoonidhi-specific quirks worth knowing about going in:

• No resumable downloads. The portal doesn't honor HTTP Range requests, so an interrupted download restarts from byte 0 rather than picking up where it left off. query download reports this explicitly (↺ restarted from scratch) when it happens.
• Cold storage. Scenes older than roughly a year often fail with a 404 error on direct download — Bhoonidhi's archiving policy isn't publicly documented, but this age threshold is a consistent pattern. The download report flags these as cold_storage rather than a generic failure. This CLI can only fetch OpenData_DirectDownload scenes; a 404'd scene can't be retrieved through it at all. You can request the scene directly on the Bhoonidhi Browse & Order Portal — cart/order support is planned for this CLI but not yet implemented.
• Already have it somewhere else? query download checks if a scene's already downloaded and SHA-verified in a different folder before re-fetching it — if it finds one, it'll tell you and ask before wasting bandwidth. --force skips the check and downloads anyway.

Limitations

• Search is bounding-box only — no point-coordinate or shapefile-based search yet.
• Only OpenData_DirectDownload scenes can be fetched directly; priced/on-order scenes show up in search results but are skipped on download (see above).
• No cart/order integration, so scenes that need to be requested first aren't reachable through this tool at all.

Development

git clone https://github.com/geovicco-dev/bhoonidhi-downloader.git
cd bhoonidhi-downloader
uv sync --group dev

Before committing, run the same checks CI runs:

uv run ruff check src/
uv run ruff format --check src/

Changelog

See CHANGELOG.md for what's changed release to release.

License

MIT