This repository is a copy of jddeal/python_cmr which is no longer maintained. It has been copied here with the permission of the original author for the purpose of continuing to develop a python library that can be used for CMR access.

Python CMR

Python CMR is an easy to use wrapper to the NASA EOSDIS Common Metadata Repository API. This package aims to make querying the API intuitive and less error-prone by providing methods that will preemptively check for invalid input and handle the URL encoding the CMR API expects.

Getting access to NASA's earth science metadata is as simple as this:

from cmr import CollectionQuery, GranuleQuery, ToolQuery, ServiceQuery, VariableQuery

api = CollectionQuery()
collections = api.archive_center("LP DAAC").keyword("AST_L1*").get(5)

print("Collections:")
for collection in collections:
  print(collection["short_name"])

api = GranuleQuery()
granules = api.short_name("AST_L1T").point(-112.73, 42.5).get(3)

print("Granule Titles:")
for granule in granules:
  print(granule["title"])

Collections:
AST_L1A
AST_L1AE
AST_L1T
Granule Titles:
SC:AST_L1T.003:2149105822
SC:AST_L1T.003:2149105820
SC:AST_L1T.003:2149155037

Installation

To install from pypi:

$ pip install python-cmr

To install from github, perhaps to try out the dev branch:

$ git clone https://github.com/nasa/python_cmr
$ cd python-cmr
$ pip install .

Examples

This library is broken into two classes, CollectionQuery and GranuleQuery. Each of these classes provide a large set of methods used to build a query for CMR. Not all parameters provided by the CMR API are covered by this version of python-cmr.

The following methods are available to both collection and granule queries:

# search for granules matching a specific product/short_name
api.short_name("AST_L1T")

# search for granules matching a specific version
api.version("006")

# search for granules at a specific longitude and latitude
api.point(-112.73, 42.5)

# search for granules in an area bound by a box (lower left lon/lat, upper right lon/lat)
api.bounding_box(-112.70, 42.5, -110, 44.5)

# search for granules in a polygon (these need to be in counter clockwise order and the
# last coordinate must match the first in order to close the polygon)
api.polygon([(-100, 40), (-110, 40), (-105, 38), (-100, 40)])

# search for granules in a line
api.line([(-100, 40), (-90, 40), (-95, 38)])

# search for granules in an open or closed date range
api.temporal("2016-10-10T01:02:00Z", "2016-10-12T00:00:30Z")
api.temporal("2016-10-10T01:02:00Z", None)
api.temporal(datetime(2016, 10, 10, 1, 2, 0), datetime.now())

# only include granules available for download
api.downloadable()

# only include granules that are unavailable for download
api.online_only()

# search for collections/granules associated with or identified by concept IDs
# note: often the ECHO collection ID can be used here as well
# note: when using CollectionQuery, only collection concept IDs can be passed
# note: when uses GranuleQuery, passing a collection's concept ID will filter by granules associated
#       with that particular collection.
api.concept_id("C1299783579-LPDAAC_ECS")
api.concept_id(["G1327299284-LPDAAC_ECS", "G1326330014-LPDAAC_ECS"])

# search by provider
api.provider('POCLOUD')

# search non-ops CMR environment
from cmr import CMR_UAT
api.mode(CMR_UAT)

Granule searches support these methods (in addition to the shared methods above):

# search for a granule by its unique ID
api.granule_ur("SC:AST_L1T.003:2150315169")
# search for granules from a specific orbit
api.orbit_number(5000)
# search for a granule by name
api.short_name("MOD09GA").readable_granule_name(["*h32v08*","*h30v13*"])

# filter by the day/night flag
api.day_night_flag("day")

# filter by cloud cover percentage range
api.cloud_cover(25, 75)

# filter by specific instrument or platform
api.instrument("MODIS")
api.platform("Terra")

# filter by a sort_key note: sort_keys are require some other fields to find some existing granules before they can be sorted

api.parameters(short_name="OMNO2", version="003", provider='GES_DISC', sort_key='-start_date')

Collection searches support these methods (in addition to the shared methods above):

# search for collections from a specific archive center
api.archive_center("LP DAAC")

# case insensitive, wildcard enabled text search through most collection fields
api.keyword("M*D09")

# search by native_id
api.native_id('native_id')

# filter by tool concept id
api.tool_concept_id('TL2092786348-POCLOUD')

# filter by service concept id
api.service_concept_id('S1962070864-POCLOUD')

Service searches support the following methods

# Search via provider
api = ServiceQuery()
api.provider('POCLOUD')

# Search via native_id
api.native_id('POCLOUD_podaac_l2_cloud_subsetter')

# Search via name
api.name('PODAAC L2 Cloud Subsetter')

# Search via concept_id
api.concept_id('S1962070864-POCLOUD')

Tool searches support the following methods

# Search via provider
api = ToolQuery()
api.provider('POCLOUD')

# Search via native_id
api.native_id('POCLOUD_hitide')

# Search via name
api.name('hitide')

# Search via concept_id
api.concept_id('TL2092786348-POCLOUD')

Variable searches support the following methods

# Search via provider
api = VariableQuery()
api.provider('POCLOUD')

# Search via native_id
api.native_id('JASON_CS_S6A_L2_AMR_RAD_STATIC_CALIBRATION-AMR_Side_1-acc_lat')

# Search via name
api.name('/AMR_Side_1/acc_lat')

# Search via concept_id
api.concept_id('V2112019824-POCLOUD')

As an alternative to chaining methods together to set the parameters of your query, a method exists to allow you to pass your parameters as keyword arguments:

# search for AST_L1T version 003 granules at latitude 42, longitude -100
api.parameters(
    short_name="AST_L1T",
    version="003",
    point=(-100, 42)
)

Note: the kwarg key should match the name of a method from the above examples, and the value should be a tuple if it's a parameter that requires multiple values.

To inspect and retrieve results from the API, the following methods are available:

# inspect the number of results the query will return without downloading the results
print(api.hits())

# retrieve 100 granules
granules = api.get(100)

# retrieve 25,000 granules
granules = api.get(25000)

# retrieve all the granules possible for the query
granules = api.get_all()  # this is a shortcut for api.get(api.hits())

By default the responses will return as json and be accessible as a list of python dictionaries. Other formats can be specified before making the request:

granules = api.format("echo10").get(100)

We can add token to the api calls by setting headers using the following functions:

# Use token function for EDL echo-token or launchpad token
api.token(token)

# Use bearer token function for EDL bearer tokens
api.bearer_token(token)

The following formats are supported for both granule and collection queries:

json (default)
xml
echo10
iso
iso19115
csv
atom
kml
native

Collection queries also support the following formats:

dif
dif10
opendata
umm_json
umm_json_vX_Y (ex: umm_json_v1_9)

Developing

python-cmr uses the poetry build system. Download and install poetry before starting development

Install Dependencies

With dev dependencies:

poetry install

Without dev dependencies:

poetry install --no-dev

Update Dependencies

poetry update

Add new Dependency

poetry add requests

Development-only dependency:

poetry add --dev pytest

Build project

poetry build

Lint project

poetry run flake8

Run Tests

poetry run pytest

jaybarra/python_cmr