Python API


# connect to the API
from sentinelsat import SentinelAPI, read_geojson, geojson_to_wkt
from datetime import date

api = SentinelAPI('user', 'password', '')

# download single scene by known product id<product_id>)

# search by polygon, time, and SciHub query keywords
footprint = geojson_to_wkt(read_geojson('/path/to/map.geojson'))
products = api.query(footprint,
                     date=('20151219', date(2015, 12, 29)),
                     cloudcoverpercentage=(0, 30))

# download all results from the search

# convert to Pandas DataFrame
products_df = api.to_dataframe(products)

# GeoJSON FeatureCollection containing footprints and metadata of the scenes

# GeoPandas GeoDataFrame with the metadata of the scenes and the footprints as geometries

# Get basic information about the product: its title, file size, MD5 sum, date, footprint and
# its download url

# Get the product's full metadata available on the server
api.get_product_odata(<product_id>, full=True)

Valid search query keywords can be found at the Copernicus Open Access Hub documentation.


The Copernicus Open Access Hub and probably most Data Hubs require authentication. You can provide your credentials with SentinelAPI(<your username>, <your password>). Alternatively, you can use SentinelAPI(None, None) and enter your credentials in a file .netrc in your user home directory in the following form:

login <your username>
password <your password>

Either way, if you get an error 401 Unauthorized, your credentials were wrong or not yet active for the endpoint you are contacting.

Sorting & Filtering

In addition to the search query keywords sentinelsat allows filtering and sorting of search results before download. To simplify these operations sentinelsat offers the convenience functions to_geojson(), to_dataframe() and to_geodataframe() which return the search results as a GeoJSON object, Pandas DataFrame or a GeoPandas GeoDataFrame, respectively. to_dataframe() and to_geodataframe() require pandas and geopandas to be installed, respectively.

In this example we query Sentinel-2 scenes over a location and convert the query results to a Pandas DataFrame. The DataFrame is then sorted by cloud cover and ingestion date. We limit the query to first 5 results within our timespan and download them, starting with the least cloudy scene. Filtering can be done with all data types, as long as you pass the id to the download function.

# connect to the API
from sentinelsat import SentinelAPI, read_geojson, geojson_to_wkt
from datetime import date

api = SentinelAPI('user', 'password', '')

# search by polygon, time, and SciHub query keywords
footprint = geojson_to_wkt(read_geojson('map.geojson'))
products = api.query(footprint,
                     date=('20151219', date(2015, 12, 29)),

# convert to Pandas DataFrame
products_df = api.to_dataframe(products)

# sort and limit to first 5 sorted products
products_df_sorted = products_df.sort_values(['cloudcoverpercentage', 'ingestiondate'], ascending=[True, True])
products_df_sorted = products_df_sorted.head(5)

# download sorted and reduced products

Getting Product Metadata

Sentinelsat provides two methods for retrieving product metadata from the server, one for each API offered by the Copernicus Open Access Hub:

  • query() for OpenSearch (Solr), which supports filtering products by their attributes and returns metadata for all matched products at once.
  • get_product_odata() for OData, which can be queried one product at a time but provides the full metadata available for each product, as well as information about the product file such as the file size and checksum, which are not available from OpenSearch.

Both methods return a dictionary containing the metadata items. More specifically, query() returns a dictionary with an entry for each returned product with its ID as the key and the attributes’ dictionary as the value.

All of the attributes returned by the OpenSearch API have a corresponding but differently named attribute in the OData’s full metadata response. See the DataHubSystem’s metadata definition files to find the exact mapping between them (OpenSearch attributes have a <solrField> tag added):

OpenSearch example

>>> api.query(date=('NOW-8HOURS', 'NOW'), producttype='SLC')
              {'acquisitiontype': 'NOMINAL',
               'beginposition': datetime.datetime(2017, 4, 25, 15, 56, 12, 814000),
               'endposition': datetime.datetime(2017, 4, 25, 15, 56, 39, 758000),
               'filename': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF.SAFE',
               'footprint': 'POLYGON ((34.322010 0.401648,36.540989 0.876987,36.884121 -0.747357,34.664474 -1.227940,34.322010 0.401648))',
               'format': 'SAFE',
               'gmlfootprint': '<gml:Polygon srsName="" xmlns:gml="">\n   <gml:outerBoundaryIs>\n      <gml:LinearRing>\n         <gml:coordinates>0.401648,34.322010 0.876987,36.540989 -0.747357,36.884121 -1.227940,34.664474 0.401648,34.322010</gml:coordinates>\n      </gml:LinearRing>\n   </gml:outerBoundaryIs>\n</gml:Polygon>',
               'identifier': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF',
               'ingestiondate': datetime.datetime(2017, 4, 25, 19, 23, 45, 956000),
               'instrumentname': 'Synthetic Aperture Radar (C-band)',
               'instrumentshortname': 'SAR-C SAR',
               'lastorbitnumber': 16302,
               'lastrelativeorbitnumber': 130,
               'link': "'04548172-c64a-418f-8e83-7a4d148adf1e')/$value",
               'link_alternative': "'04548172-c64a-418f-8e83-7a4d148adf1e')/",
               'link_icon': "'04548172-c64a-418f-8e83-7a4d148adf1e')/Products('Quicklook')/$value",
               'missiondatatakeid': 110481,
               'orbitdirection': 'ASCENDING',
               'orbitnumber': 16302,
               'platformidentifier': '2014-016A',
               'platformname': 'Sentinel-1',
               'polarisationmode': 'VV VH',
               'productclass': 'S',
               'producttype': 'SLC',
               'relativeorbitnumber': 130,
               'sensoroperationalmode': 'IW',
               'size': '7.1 GB',
               'slicenumber': 8,
               'status': 'ARCHIVED',
               'summary': 'Date: 2017-04-25T15:56:12.814Z, Instrument: SAR-C SAR, Mode: VV VH, Satellite: Sentinel-1, Size: 7.1 GB',
               'swathidentifier': 'IW1 IW2 IW3',
               'title': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF',
               'uuid': '04548172-c64a-418f-8e83-7a4d148adf1e'}),

OData example

Only the most basic information available from the OData API is returned by default, if full=True is not set. The full metadata query response is quite large and not always required, so it is not requested by default.

>>> api.get_product_odata('04548172-c64a-418f-8e83-7a4d148adf1e')
{'date': datetime.datetime(2017, 4, 25, 15, 56, 12, 814000),
 'footprint': 'POLYGON((34.322010 0.401648,36.540989 0.876987,36.884121 -0.747357,34.664474 -1.227940,34.322010 0.401648))',
 'id': '04548172-c64a-418f-8e83-7a4d148adf1e',
 'md5': 'E5855D1C974171D33EE4BC08B9D221AE',
 'size': 4633501134,
 'title': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF',
 'url': "'04548172-c64a-418f-8e83-7a4d148adf1e')/$value"}

With full=True we receive the full metadata available for the product.

>>> api.get_product_odata('04548172-c64a-418f-8e83-7a4d148adf1e', full=True)
{'Acquisition Type': 'NOMINAL',
 'Carrier rocket': 'Soyuz',
 'Cycle number': 107,
 'Date': datetime.datetime(2017, 4, 25, 15, 56, 12, 814000),
 'Filename': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF.SAFE',
 'Footprint': '<gml:Polygon srsName="" xmlns:gml="">\n   <gml:outerBoundaryIs>\n      <gml:LinearRing>\n         <gml:coordinates>0.401648,34.322010 0.876987,36.540989 -0.747357,36.884121 -1.227940,34.664474 0.401648,34.322010</gml:coordinates>\n      </gml:LinearRing>\n   </gml:outerBoundaryIs>\n</gml:Polygon>',
 'Format': 'SAFE',
 'Identifier': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF',
 'Ingestion Date': datetime.datetime(2017, 4, 25, 19, 23, 45, 956000),
 'Instrument': 'SAR-C',
 'Instrument abbreviation': 'SAR-C SAR',
 'Instrument description': '<a target="_blank" href=""></a>',
 'Instrument description text': 'The SAR Antenna Subsystem (SAS) is developed and build by AstriumGmbH. It is a large foldable planar phased array antenna, which isformed by a centre panel and two antenna side wings. In deployedconfiguration the antenna has an overall aperture of 12.3 x 0.84 m.The antenna provides a fast electronic scanning capability inazimuth and elevation and is based on low loss and highly stablewaveguide radiators build in carbon fibre technology, which arealready successfully used by the TerraSAR-X radar imaging mission.The SAR Electronic Subsystem (SES) is developed and build byAstrium Ltd. It provides all radar control, IF/ RF signalgeneration and receive data handling functions for the SARInstrument. The fully redundant SES is based on a channelisedarchitecture with one transmit and two receive chains, providing amodular approach to the generation and reception of wide-bandsignals and the handling of multi-polarisation modes. One keyfeature is the implementation of the Flexible Dynamic BlockAdaptive Quantisation (FD-BAQ) data compression concept, whichallows an efficient use of on-board storage resources and minimisesdownlink times.',
 'Instrument mode': 'IW',
 'Instrument name': 'Synthetic Aperture Radar (C-band)',
 'Instrument swath': 'IW1 IW2 IW3',
 'JTS footprint': 'POLYGON ((34.322010 0.401648,36.540989 0.876987,36.884121 -0.747357,34.664474 -1.227940,34.322010 0.401648))',
 'Launch date': 'April 3rd, 2014',
 'Mission datatake id': 110481,
 'Mission type': 'Earth observation',
 'Mode': 'IW',
 'NSSDC identifier': '2014-016A',
 'Operator': 'European Space Agency',
 'Orbit number (start)': 16302,
 'Orbit number (stop)': 16302,
 'Pass direction': 'ASCENDING',
 'Phase identifier': 1,
 'Polarisation': 'VV VH',
 'Product class': 'S',
 'Product class description': 'SAR Standard L1 Product',
 'Product composition': 'Slice',
 'Product level': 'L1',
 'Product type': 'SLC',
 'Relative orbit (start)': 130,
 'Relative orbit (stop)': 130,
 'Satellite': 'Sentinel-1',
 'Satellite description': '<a target="_blank" href=""></a>',
 'Satellite name': 'Sentinel-1',
 'Satellite number': 'A',
 'Sensing start': datetime.datetime(2017, 4, 25, 15, 56, 12, 814000),
 'Sensing stop': datetime.datetime(2017, 4, 25, 15, 56, 39, 758000),
 'Size': '7.1 GB',
 'Slice number': 8,
 'Start relative orbit number': 130,
 'Status': 'ARCHIVED',
 'Stop relative orbit number': 130,
 'Timeliness Category': 'Fast-24h',
 'date': datetime.datetime(2017, 4, 25, 15, 56, 12, 814000),
 'footprint': 'POLYGON((34.322010 0.401648,36.540989 0.876987,36.884121 -0.747357,34.664474 -1.227940,34.322010 0.401648))',
 'id': '04548172-c64a-418f-8e83-7a4d148adf1e',
 'md5': 'E5855D1C974171D33EE4BC08B9D221AE',
 'size': 4633501134,
 'title': 'S1A_IW_SLC__1SDV_20170425T155612_20170425T155639_016302_01AF91_46FF',
 'url': "'04548172-c64a-418f-8e83-7a4d148adf1e')/$value"}


Copernicus Open Access Hub no longer stores all products online for immediate retrieval. Offline products can be requested from the Long Term Archive (LTA) and should become available within 24 hours. Copernicus Open Access Hub’s quota currently permits users to request an offline product every 30 minutes.

A product’s availability can be checked with a regular OData query by evaluating the Online property value or by using the is_online() convenience method.

product_info = api.get_product_odata(<product_id>)
is_online = product_info['Online']
# or
is_online = api.is_online(<product_id>)

if is_online:
    print('Product {} is online. Starting download.'.format(<product_id>))<product_id>)
    print('Product {} is not online.'.format(<product_id>))

When trying to download an offline product with download() it will trigger its retrieval from the LTA.

Given a list of offline and online products, download_all() will download online products, while concurrently triggering the retrieval of offline products from the LTA. Offline products that become online while downloading will be added to the download queue. download_all() terminates when the download queue is empty, even if not all products were retrieved from the LTA. We suggest repeatedly calling download_all() to download all products, either manually or using a third-party library, e.g. tenacity.

from sentinelsat import SentinelAPI
import tenacity

api = SentinelAPI('user', 'password')

@tenacity.retry(stop=tenacity.stop_after_attempt(3), wait=tenacity.wait_fixed(3600))
def download_all(*args, **kwargs):
    return api.download_all(*args, **kwargs)

downloaded, triggered, failed = download_all(<product_ids>)


Sentinelsat logs to sentinelsat and the API to sentinelsat.SentinelAPI.

There is no predefined logging handler, so in order to have your script print the log messages, either use logging.baseConfig

import logging

logging.basicConfig(format='%(message)s', level='INFO')

or add a custom handler for sentinelsat (as implemented in

import logging

logger = logging.getLogger('sentinelsat')

h = logging.StreamHandler()
fmt = logging.Formatter('%(message)s')

Product node API

The product node API is implemented in the sentinelsat.products module and, in addition to all the features provided by the standard API, allows to download only some of the files included in a Sentinel product exploiting the node selection feature provided by the OData Web API.

from sentinelsat import SentinelProductsAPI, make_path_filter

# define the filter function to select files (to be excluded in this case)
nodefilter = make_path_filter("*measurement/*", exclude=True)

# connect to the API
api = SentinelProductAPI("user", "password")

# download a single product excluding measurement files<product_id>, nodefilter=nodefilter)

Of course it also works for multiple products:

# download a multiple products excluding measurement files
api.download_all(<products>, nodefilter=nodefilter)

The example above downloads all files in each of the requested products only excluding (large) measurements files. This can be useful for analyses exclusively based on product annotations (including calibration annotations) or, e.g., to download the KML preview included in the product.

The file selection is implemented by specifying a nodefilter function that is called for each file (only excluding the manifest which is always downloaded) in the requested products and returns True if the file have to be downloaded, False otherwise.

The nodefilter function has the following signature:

def nodefilter(node_info: dict) -> bool:

The node_info parameter is a dictionary containing (at least) the following keys:

url:the URL to download the product file node
node_path:the file path within the product (e.g. “./preview/map-overlay.kml”)
size:the file size in bytes (int)
md5:the file md5 checksum

In the example above it has been used an helper function (sentinelsat.products.make_path_filter()), provided by the sentinelsat.products module, that generates nodefilter functions based on glob expressions applied to the node_path value.

The following code:

nodefilter = make_path_filter("*measurement/*", exclude=True)

is more or less equivalent to:

import fnmatch

def node_filter(node_info):
  if not fnmatch.fnmatch(node_info["node_path"].lower(), pattern):
    return True
    return False

The sentinelsat product node API also provides:

Of course the user can write its own nodefilter functions if necessary.

More Examples

Search Sentinel-2 L1C by tile

To search for recent Sentinel-2 L1C imagery by MGRS tile, you can use the tileid parameter:

from collections import OrderedDict
from sentinelsat import SentinelAPI

api = SentinelAPI('user', 'password')

tiles = ['33VUC', '33UUB']

query_kwargs = {
        'platformname': 'Sentinel-2',
        'producttype': 'S2MSI1C',
        'date': ('NOW-14DAYS', 'NOW')}

products = OrderedDict()
for tile in tiles:
    kw = query_kwargs.copy()
    kw['tileid'] = tile
    pp = api.query(**kw)


NB: Older products may not be found with the tileid parameter. On the Copernicus Open Access Hub, it seems to be available for most L1C products (product type S2MSI1C) from recent years, but this differs by region, too. To be on the safe side, combine the tileid search with a filename pattern search:

kw = query_kwargs.copy()
kw['raw'] = f'tileid:{tile} OR filename:*_T{tile}_*'
pp = api.query(**kw)

Download only some of the channels of a Sentinel-1 product

In some cases the user may be interested only to a specific sub-swath and/or polarization of Sentinel1 SLC products (e.g. for an interferometric analysis).

The following nodefilter function only downloads “HH” polarization measurement files for sub-swaths “EW1” and “EW2”:

nodefilter = make_path_filter("*s1?-ew[12]-slc-hh-*.tiff")

Considering that e.g. a Dual Pol Extended Wide Swath Sentinel-1 product includes 2 measurement files for each of the 5 sub-swath the above filter allows to reduce consistently the amount of data to be downloaded (form 10 to 2 TIFF files approx 700MB each).

Of course the nodefilter function have to be used to initialize the product node API sentinelsat.products.SentinelProductsAPI as explained above.

api.download_all(<products>, nodefilter=nodefilter)