Getting Started with the AρρEEARS API: Submitting and Downloading an Area Request¶
This tutorial demonstrates how to use Python to connect to the AρρEEARS API¶
The Application for Extracting and Exploring Analysis Ready Samples (AρρEEARS) offers a simple and efficient way to access and transform geospatial data from a variety of federal data archives in an easy-to-use web application interface. AρρEEARS enables users to subset geospatial data spatially, temporally, and by band/layer for point and area samples. AρρEEARS returns not only the requested data, but also the associated quality values, and offers interactive visualizations with summary statistics in the web interface. The AρρEEARS API offers users programmatic access to all features available in AρρEEARS, with the exception of visualizations. The API features are demonstrated in this notebook.
Example: Submit an area request using a U.S. National Park boundary as the region of interest for extracting elevation, vegetation and land surface temperature data¶
Connect to the AρρEEARS API, query the list of available products, submit an area sample request, download the request, become familiar with the AρρEEARS Quality API, and import the results into Python for visualization. AρρEEARS area sample requests allow users to subset their desired data by spatial area via vector polygons (shapefiles or GeoJSONs). Users can also reproject and reformat the output data. AρρEEARS returns the valid data from the parameters defined within the sample request.
Data Used in the Example:¶
- Data layers:
- NASA MEaSUREs Shuttle Radar Topography Mission (SRTM) Version 3 Digital Elevation Model
- SRTMGL1_NC.003, 30m, static: 'SRTM_DEM'
- Combined MODIS Leaf Area Index (LAI)
- MCD15A3H.006, 500m, 4 day: 'Lai_500m'
- Terra MODIS Land Surface Temperature
- MOD11A2.006, 1000m, 8 day: 'LST_Day_1km', 'LST_Night_1km'
- NASA MEaSUREs Shuttle Radar Topography Mission (SRTM) Version 3 Digital Elevation Model
Topics Covered:¶
- Getting Started
1a. Set Up the Working Environment
1b. Login [Login] - Query Available Products [Product API]
2a. Search and Explore Available Products [List Products]
2b. Search and Explore Available Layers [List Layers] - Submit an Area Request [Tasks]
3a. Import a Shapefile
3b. Search and Explore Available Projections [Spatial API]
3c. Compile a JSON [Task Object]
3d. Submit a Task Request [Submit Task]
3e. Retrieve Task Status [Retrieve Task] - Download a Request [Bundle API]
4a. Explore Files in Request Output [List Files]
4b. Download Files in a Request (Automation) [Download File] - Explore AρρEEARS Quality Service [Quality API]
5a. List Quality Layers [List Quality Layers]
5b. Show Quality Values [List Quality Values]
5c. Decode Quality Values [Decode Quality Values] - BONUS: Import Request Output and Visualize
6a. Import a GeoTIFF
6b. Plot a GeoTIFF
Dependencies:¶
- This tutorial was tested using Python 3.6.1.
- A NASA Earthdata Login account is required to complete this tutorial. You can create an account at the link provided.
- To execute section 6, the Geospatial Data Abstraction Library (GDAL) is required.
AρρEEARS Information:¶
To access AρρEEARS, visit: https://lpdaacsvc.cr.usgs.gov/appeears/
For comprehensive documentation of the full functionality of the AρρEEARS API, please see the AρρEEARS API Documentation
Throughout the tutorial, specific sections of the API documentation can be accessed by clicking on the bracketed [] links in the section headings.
Files Used in this Tutorial:¶
Source Code used to Generate this Tutorial:¶
1. Getting Started¶
In [1]:
# Import packages
import requests as r
import getpass, pprint, time, os, cgi, json
import geopandas as gpd
If you are missing any of the packages above, download them in order to use the full functionality of this tutorial.¶
In [2]:
# Set input directory, change working directory
inDir = 'D:/appeears-api-getting-started/' # IMPORTANT: Update to reflect directory on your OS
os.chdir(inDir) # Change to working directory
api = 'https://lpdaacsvc.cr.usgs.gov/appeears/api/' # Set the AρρEEARS API to a variable
If you plan to execute this tutorial on your own OS, `inDir` above needs to be changed.
1b. Login [Login]¶
To submit a request, you must first login to the AρρEEARS API. The AρρEEARS API requires the same NASA Earthdata Login as the AρρEEARS user interface. Use the getpass package to enter your NASA Earthdata login Username and Password. When prompted after executing the code block below, enter your username followed by your password.¶
In [3]:
user = getpass.getpass(prompt = 'Enter NASA Earthdata Login Username: ') # Input NASA Earthdata Login Username
password = getpass.getpass(prompt = 'Enter NASA Earthdata Login Password: ') # Input NASA Earthdata Login Password
Use the requests package to post your username and password. A successful login will provide you with a token to be used later in this tutorial to submit a request. For more information or if you are experiencing difficulties, please see the API Documentation.¶
In [4]:
token_response = r.post('{}login'.format(api), auth=(user, password)).json() # Insert API URL, call login service, provide credentials & return json
del user, password # Remove user and password information
token_response # Print response
Out[4]:
Above, you should see a Bearer token. Notice that this token will expire approximately 48 hours after being acquired.¶
2. Query Available Products [Product API]¶
2a. Search and Explore Available Products [List Products]¶
The product API provides details about all of the products and layers available in AρρEEARS. Below, call the product API to list all of the products available in AρρEEARS.¶
In [5]:
product_response = r.get('{}product'.format(api)).json() # request all products in the product service
print('AρρEEARS currently supports {} products.'.format(len(product_response))) # Print no. products available in AppEEARS
Next, create a dictionary indexed by product name, making it easier to query a specific product.¶
In [6]:
products = {p['ProductAndVersion']: p for p in product_response} # Create a dictionary indexed by product name & version
products['MCD15A3H.006'] # Print information for MCD15A3H.006 LAI/FPAR Product
Out[6]:
The product service provides many useful details, including if a product is currently available in AρρEEARS, a description, and information on the spatial and temporal resolution.¶
Below, make a list of all product+version names, and search for products containing Leaf Area Index in their description.¶
In [7]:
prodNames = {p['ProductAndVersion'] for p in product_response} # Make list of all products (including version)
for p in prodNames: # Make for loop to search list of products 'Description' for a keyword
if 'Leaf Area Index' in products[p]['Description']:
pprint.pprint(products[p]) # Print info for each product containing LAI in its description
Using the info above, start a list of desired products by using the highest temporal resolution LAI product, MCD15A3H.006.¶
In [8]:
prods = ['MCD15A3H.006'] # Start a list for products to be requested, beginning with MCD15A3H.006
prods.append('MOD11A2.006') # Append the MOD11A2.006 8 day LST product to the list of products desired
prods.append('SRTMGL1_NC.003') # Append the SRTMGL1_NC.003 product to the list of products desired
prods # Print list
Out[8]:
2b. Search and Explore Available Layers [List Layers]¶
The product API allows you to call all of the layers available for a given product. Each product is referenced by its ProductAndVersion property. For a list of the layer names only, print the keys from the dictionary below.¶
In [9]:
lst_response = r.get('{}product/{}'.format(api, prods[1])).json() # Request layers for the 2nd product (index 1) in the list: MOD11A2.006
list(lst_response.keys())
Out[9]:
Use the dictionary key 'LST_Day_1km' to see the information for that layer in the response.¶
In [10]:
lst_response['LST_Day_1km'] # Print layer response
Out[10]:
AρρEEARS also allows subsetting data spectrally (by band). Create a tupled list with product name and specific layers desired.¶
In [11]:
layers = [(prods[1],'LST_Day_1km'),(prods[1],'LST_Night_1km')] # Create tupled list linking desired product with desired layers
Next, request the layers for the MCD15A3H.006 product.¶
In [12]:
lai_response = r.get('{}product/{}'.format(api, prods[0])).json() # Request layers for the 1st product (index 0) in the list: MCD15A3H.006
list(lai_response.keys()) # Print the LAI layer names
Out[12]:
In [13]:
lai_response['Lai_500m']['Description'] # Make sure the correct layer is requested
Out[13]:
In [14]:
layers.append((prods[0],'Lai_500m')) # Append to tupled list linking desired product with desired layers
Thirdly, request the layers for the SRTMGL1_NC.003 product.¶
In [15]:
dem_response = r.get('{}product/{}'.format(api, prods[2])).json() # Request layers for the 3rd product (index 2) in the list: SRTMGL1_NC.003
list(dem_response.keys()) # Print the SRTM DEM layer names
Out[15]:
Finally, append SRTMGL1_DEM to the tupled list of desired products/layers.¶
In [16]:
layers.append((prods[2], 'SRTMGL1_DEM')) # Append to tupled list linking desired product with desired layers
Below, take the tupled list (layers) and create a list of dictionaries to store each layer+product combination. This will make it easier to insert into the json file used to submit a request in Section 3.¶
In [17]:
prodLayer = []
for l in layers:
prodLayer.append({
"layer": l[1],
"product": l[0]
})
prodLayer
Out[17]:
3. Submit an Area Request [Tasks]¶
The Submit task API call provides a way to submit a new request to be processed. It can accept data via JSON, query string, or a combination of both. In the example below, compile a json and submit a request. Tasks in AρρEEARS correspond to each request associated with your user account. Therefore, each of the calls to this service requires an authentication token (see Section 1c.), which is stored in a header below.¶
In [18]:
token = token_response['token'] # Save login token to a variable
head = {'Authorization': 'Bearer {}'.format(token)} # Create a header to store token information, needed to submit a request
3a. Import a Shapefile¶
In this section, begin by importing a shapefile using the geopandas package. The shapefile is publically available for download from the NPS website.¶
In [19]:
nps = gpd.read_file('{}nps_boundary.shp'.format(inDir + os.sep)) # Read in shapefile as dataframe using geopandas
print(nps.head()) # Print first few lines of dataframe
Below, query the geopandas dataframe for the national park that you are interested in using for your region of interest, here Grand Canyon National Park.¶
In [20]:
nps_gc = nps[nps['UNIT_NAME']=='Grand Canyon National Park'].to_json() # Extract Grand Canyon NP and set to variable
nps_gc = json.loads(nps_gc) # Convert to json format
3b. Search and Explore Available Projections [Spatial API]¶
The spatial API provides some helper services used to support submitting area task requests. The call below will retrieve the list of supported projections in AρρEEARS.¶
In [21]:
projections = r.get('{}spatial/proj'.format(api)).json() # Call to spatial API, return projs as json
projections # Print projections and information
Out[21]:
Create a dictionary of projections with projection Name as the keys.¶
In [22]:
projs = {} # Create an empty dictionary
for p in projections: projs[p['Name']] = p # Fill dictionary with `Name` as keys
list(projs.keys()) # Print dictionary keys
Out[22]:
Print the information for the projection used in this tutorial.¶
In [23]:
projs['geographic']
Out[23]:
3c. Compile a JSON [Task Object]¶
In this section, begin by setting up the information needed to compile an acceptable json for submitting an AρρEEARS area request. For detailed information on required json parameters, see the API Documentation.¶
In [24]:
task_name = input('Enter a Task Name: ') # User-defined name of the task: 'NPS Vegetation Area' used in example
In [25]:
task_type = ['point','area'] # Type of task, area or point
proj = projs['geographic']['Name'] # Set output projection
outFormat = ['geotiff', 'netcdf4'] # Set output file format type
startDate = '07-01-2017' # Start of the date range for which to extract data: MM-DD-YYYY
endDate = '07-31-2017' # End of the date range for which to extract data: MM-DD-YYYY
recurring = False # Specify True for a recurring date range
#yearRange = [2000,2016] # if recurring = True, set yearRange, change start/end date to MM-DD
Below, compile the JSON to be submitted as an area request. Notice that nps_gc is inserted from the shapefile transformed to a json via the geopandas and json packages above in section 3a.¶
In [26]:
task = {
'task_type': task_type[1],
'task_name': task_name,
'params': {
'dates': [
{
'startDate': startDate,
'endDate': endDate
}],
'layers': prodLayer,
'output': {
'format': {
'type': outFormat[0]},
'projection': proj},
'geo': nps_gc,
}
}
3d. Submit a Task Request [Submit Task]¶
Below, post a call to the API task service, using the task json created above.¶
In [27]:
task_response = r.post('{}task'.format(api), json=task, headers=head).json() # Post json to the API task service, return response as json
task_response # Print task response
Out[27]:
3e. Retrieve Task Status [Retrieve Task]¶
This API call will list all of the requests associated with your user account, automatically sorted by date descending with the most recent requests listed first.¶
The AρρEEARS API contains some helpful formatting resources. Below, limit the API response to 2 entries and set pretty to True to format the response as an organized json, making it easier to read. Additional information on AρρEEARS API pagination and formatting can be found in the API documentation.¶
In [28]:
params = {'limit': 2, 'pretty': True} # Limit API response to 2 most recent entries, return as pretty json
In [29]:
tasks_response = r.get('{}task'.format(api), params=params, headers=head).json() # Query task service, setting params and header
tasks_response # Print tasks response
Out[29]:
Next, take the task id returned from the task_response that was generated when submitting your request, and use the AρρEEARS API status service to check the status of your request.¶
In [30]:
task_id = task_response['task_id'] # Set task id from request submission
status_response = r.get('{}status/{}'.format(api, task_id), headers=head).json() # Call status service with specific task ID & user credentials
status_response
Out[30]:
In [31]:
# Ping API until request is complete, then continue to Section 4
starttime = time.time()
while r.get('{}task/{}'.format(api, task_id), headers=head).json()['status'] != 'done':
print(r.get('{}task/{}'.format(api, task_id), headers=head).json()['status'])
time.sleep(20.0 - ((time.time() - starttime) % 20.0))
print(r.get('{}task/{}'.format(api, task_id), headers=head).json()['status'])
4. Download a Request [Bundle API]¶
Before downloading the request output, set up an output directory to store the output files, and examine the files contained in the request output.¶
In [32]:
destDir = os.path.join(inDir, task_name) # Set up output directory using input directory and task name
if not os.path.exists(destDir):os.makedirs(destDir) # Create the output directory
4a. Explore Files in Request Output [List Files]¶
The bundle API provides information about completed tasks. For any completed task, a bundle can be queried to return the files contained as a part of the task request. Below, call the bundle API and return all of the output files.¶
In [33]:
bundle = r.get('{}bundle/{}'.format(api,task_id)).json() # Call API and return bundle contents for the task_id as json
bundle # Print bundle contents
Out[33]:
4b. Download Files in a Request (Automation) [Download File]¶
Now, use the contents of the bundle to select the file name and id and store as a dictionary.¶
In [34]:
files = {} # Create empty dictionary
for f in bundle['files']: files[f['file_id']] = f['file_name'] # Fill dictionary with file_id as keys and file_name as values
files
Out[34]:
Use the files dictionary and a for loop to automate downloading all of the output files into the output directory.¶
In [35]:
for f in files:
dl = r.get('{}bundle/{}/{}'.format(api, task_id, f), stream=True) # Get a stream to the bundle file
filename = os.path.basename(cgi.parse_header(dl.headers['Content-Disposition'])[1]['filename']) # Parse the name from Content-Disposition header
filepath = os.path.join(destDir, filename) # Create output file path
with open(filepath, 'wb') as f: # Write file to dest dir
for data in dl.iter_content(chunk_size=8192): f.write(data)
print('Downloaded files can be found at: {}'.format(destDir))
5. Explore AρρEEARS Quality Service [Quality API]¶
The quality API provides quality details about all of the data products available in AρρEEARS. Below are examples of how to query the quality API for listing quality products, layers, and values. The final example (Section 5c.) demonstrates how AρρEEARS quality services can be leveraged to decode pertinent quality values for your data.¶
First, reset pagination to include offset which allows you to set the number of results to skip before starting to return entries. Next, make a call to list all of the data product layers and the associated quality product and layer information.¶
In [36]:
params = {'limit': 6, 'pretty': True, 'offset': 20} # Limit response to 6 entries, start w/ 20th entry, return pretty json
quality_response = r.get('{}quality'.format(api), params=params).json() # Call quality API using pagination and return json
quality_response # Print response
Out[36]:
5a. List Quality Layers [List Quality Layers]¶
This API call will list all of the quality layer information for a product.¶
In [37]:
product = 'MCD15A3H.006' # Product used in the example
ql_response = r.get('{}quality/{}'.format(api,product)).json() # Call API to retrieve quality layers for selected product
ql_response # Print response
Out[37]:
5b. Show Quality Values [List Quality Values]¶
This API call will list all of the values for a given quality layer.¶
In [38]:
qlayer = ql_response[1]['QualityLayers'][0] # Set quality layer from ql_response for 'Lai_500m'
qv_response = r.get('{}quality/{}/{}'.format(api, product, qlayer)).json() # Call API for list of bit-word quality values
qv_response # Print response
Out[38]:
5c. Decode Quality Values [Decode Quality Values]¶
This API call will decode the bits for a given quality value.¶
In [39]:
val = 1 # Set a specific value
q_response = r.get('{}quality/{}/{}/{}'.format(api, product, qlayer, val)).json() # Call quality API for specific value
q_response # Print response
Out[39]:
In [40]:
# Import packages
import matplotlib.pyplot as plt
import numpy as np
from osgeo import gdal
list(files.values()) # List files downloaded
Out[40]:
In [41]:
dem = gdal.Open(destDir + '/SRTMGL1_NC.003_SRTMGL1_DEM_doy2000042_aid0001.tif' ) # Read file in
demBand = dem.GetRasterBand(1) # Read the band (layer)
demData = demBand.ReadAsArray().astype('float') # Import band as an array with type float
Next, query the metadata for the fill value, and set fill value equal to nan.¶
In [42]:
demFill = demBand.GetNoDataValue() # Returns fill value
demData[demData == demFill] = np.nan # Set fill value to nan
In [43]:
# Set matplotlib plots inline
%matplotlib inline
First, make a basic plot of the DEM data.¶
In [44]:
plt.imshow(demData); # Visualize a basic plot of the DEM data
Next, add some additional parameters to the plot.¶
In [47]:
import warnings
warnings.filterwarnings("ignore")
fig = plt.figure(figsize = (10,7.5)) # Set the figure size (x,y)
plt.axis('off') # Remove the axes' values
ax = fig.add_subplot(111)
# Plot the array, using a colormap and setting a custom linear stretch based on the min/max Elevation values
plt.imshow(demData, vmin = np.nanmin(demData), vmax = np.nanmax(demData), cmap = 'terrain');
Finally, add important map items including a legend and title.¶
In [48]:
plt.style.use("dark_background") # Default to a black background
fig2 = plt.figure(figsize=(10,7.5)) # Set the figure size
plt.axis('off') # Remove the axes' values
ax1 = fig2.add_subplot(111) # Make a subplot
fig2.subplots_adjust(top=3.8) # Adjust spacing
ax1.set_title('SRTM DEM: Grand Canyon NP',fontsize=15,fontweight='bold',color='white') # Add title
# Plot the masked data, using a colormap and setting a custom linear stretch based on the min/max DEM values
im = plt.imshow(demData, vmin = np.nanmin(demData), vmax = np.nanmax(demData), cmap = 'terrain');
cb = plt.colorbar(im, orientation='horizontal', fraction=0.047, pad=0.004, shrink=0.6) # Add a colormap legend
cb.set_label(label='Elevation (m)', color = 'white') # Set Label and color
cb.outline.set_edgecolor('white') # Set edge color
This example can provide a template to use for your own research workflows. Leveraging the AρρEEARS API for searching, extracting, and formatting analysis ready data, and importing it directly into Python means that you can keep your entire research workflow in a single software program, from start to finish.¶
Contact Information
Material written by Cole Krehbiel$^{1}$
Contact: LPDAAC@usgs.gov
Voice: +1-605-594-6116
Organization: Land Processes Distributed Active Archive Center (LP DAAC)
Website: https://lpdaac.usgs.gov/
Date last modified: 04-28-2020
$^{1}$Innovate! Inc., contractor to the U.S. Geological Survey, Earth Resources Observation and Science (EROS) Center, Sioux Falls, South Dakota, 57198-001, USA. Work performed under USGS contract G15PD00467 for LP DAAC$^{2}$.
$^{2}$LP DAAC Work performed under NASA contract NNG14HH33I.