Tuesday, April 28, 2015

ArcREST Basics - ArcGIS Online and Query()

In this weeks ArcREST basics series, I am going to go into the how to connect to your organization on ArcGIS Online (AGOL) using Python, and perform some basic queries using the manageorg sub-package in ArcREST.

First we start by building off the previous lesson and create a security handler object for AGOL.  If you missed a deeper discussion on how to handle security, please see this post.

One of the most common tasks to perform when working with your site, is to query it to see what content you have.  There are two ways to query a site:

  1. With Credentials - this means a token will be appended on the end of the query search, and users can find both public and non-public items if your users has the proper permissions to do so.
  2. Without Credentials - this means only find public items shared with everyone.
In the example below, a query is being performed with credentials being given.  This means case #1 will be applied.

from arcrest.security import AGOLTokenSecurityHandler
from arcrest.manageorg import Administration
if __name__ == "__main__":
    username = "username"
    password = "password"
    proxy_port = None
    proxy_url = None    
    securityHandler = AGOLTokenSecurityHandler(username, password,
                                               proxy_url=proxy_url,
                                               proxy_port=proxy_port)
    siteObject = Administration(securityHandler=securityHandler,
                                proxy_url=proxy_url,
                                proxy_port=proxy_port)
    results = siteObject.query(q="< some query string >")


This returns a Python dictionary object as shown below:

{
   "query" : "type:feature class",
   "total" : 12345,
   "start" : 1,
   "num" : 10,
   "nextStart" : 11,
   "results" : [ ...list of items... ]
}

Let us examine the results object.  It's a dictionary with many valuable key/value pairs.  Since this query was performed using the default values, this means that the query will begin with the first item and go to the 10th item (num).  The key nextStart tells you where the start value needs to be if you want to page each result.  So the next time you pass the query in order to get all the results you need to set the start = 11.  Manually this seems quite daunting task, but luckily for us, we have looks.  There are many way to create loops to perform this task, but in this example, a while loop will be used to walk the results and put them in a single list object.


def findContentByDate(admin, q):
    """
       finds items based on a query (q)

       Inputs:
    
          admin - manageorg.Administration object
          q - query string
    """
    start = 1
    count  = 0
    nextStart = 0
    num = 100
    results = []
    while nextStart != -1:
        query = admin.query(q=q,
                        sortOrder="desc",
                        sortField="modified",
                        t=None,
                        start=start + (count * num),
                        num=100)
        results = results + query['results']
        nextStart = query['nextStart']
        count += 1
    return results
Now we have all the items based on a query in a single Python list. This is very helpful in cataloging a site/org or even good for finding publicly shared items.

AGOL has it's own query language much like the google or other search engines.  This documentation can be found here (http://doc.arcgis.com/en/arcgis-online/reference/search.htm).  It is a must read for anyone who wants to query AGOL or Portal effectively.

Though the samples shown are created to work with Portal, then same example can be used to query a Portal site.

If you want to learn more, please visit the ArcREST Github Page.  Help us make it better and post comments, suggestions and improvements in the Issues section.  Or better yet, fork the repo and submit a pull request.

Happy Coding and Enjoy!

Wednesday, April 22, 2015

ArcREST Basics - Authentication

When starting working with any new package, there is always a learning curve, and I am trying to make ArcREST easy and fun to use.   So here is my first of many ArcREST posts on how to use this package to meet your Esri REST API use cases.

This post will focus on the most basic thing, authentication.  Authentication is the most basic idea to understand when working with ArcGIS Online, ArcGIS Server, or ArcGIS Portal.  It identifies who you are, and what you are allowed to do.  Not all operations require authentication, but it's good to understand the token based security model and how it applies to you ArcGIS stack.

A quick blurb from the ArcGIS help describes what token security does:
ArcGIS Server provides a proprietary token-based authentication mechanism where users can authenticate themselves by providing a token instead of a user name and password. An ArcGIS token is a string of encrypted information that contains the user's name, the token expiration time, and some proprietary information. To obtain a token, a user provides a valid user name and password. ArcGIS Server verifies the supplied credentials and issues a token. The user presents this token whenever accessing a secured resource.
Before ArcREST, I you probably had to write this function out a lot:


import urllib
import urllib2
import httplib
import time
import json
import contextlib

def submit_request(request):
    """ Returns the response from an HTTP request in json format."""
    with contextlib.closing(urllib2.urlopen(request)) as response:
        job_info = json.load(response)
        return job_info

def get_token(portal_url, username, password):
    """ Returns an authentication token for use in ArcGIS Online."""

    # Set the username and password parameters before
    #  getting the token. 
    #
    params = {"username": username,
              "password": password,
              "referer": "http://www.arcgis.com",
              "f": "json"}

    token_url = "{}/generateToken".format(portal_url)
    request = urllib2.Request(token_url, urllib.urlencode(params))
    token_response = submit_request(request)
    if "token" in token_response:
        print("Getting token...")
        token = token_response.get("token")
        return token
    else:
        # Request for token must be made through HTTPS.
        #
        if "error" in token_response:
            error_mess = token_response.get("error", {}).get("message")
            if "This request needs to be made over https." in error_mess:
                token_url = token_url.replace("http://", "https://")
                token = get_token(token_url, username, password)
                return token
            else:
                raise Exception("Portal error: {} ".format(error_mess))

It is lots of code just to gain access.  As a user/developer you have not even begun to work on your actual workflow of what you need.

So there has to be an easier way? 

Well there is!  ArcREST supports three main forms of token based security in the form of classes.

  • PortalTokenSecurityHandler - this allows access to ArcGIS Portal Site
  • AGOLTokenSecurityHandler - this generates a token for ArcGIS Online
  • AGSTokenSecurityHandler - this generates a token for ArcGIS Server Sites
Why should I use this? 
  1. allows for cleaner code
  2. pass username/password once in your code
  3. built in proxy support 
  4. handles token expiration automatically
#3 is a biggie, at least in my eyes.  For long running tasks, there is a chance, depending on your expiration time, that your token might fail you because it is too old.  The SecurityHandler classes automatically handle the expiration for you, and regenerate new tokens as needed.

Examples of creating Securityhandler objects:

1. Creating an ArcGIS Server Security Handler
import arcrest
if __name__ == "__main__":
    token_url = "http://mysite.com:6080/arcgis/admin/generateToken"
    username = "username"
    password = "password"
    sh = arcrest.AGSTokenSecurityHandler(username=username,
                                         password=password,
                                         token_url=token_url)

2. Creating an ArcGIS Online Security Handler
import arcrest
if __name__ == "__main__":
    username = "< username >"
    pw = "< password >"
    proxy_url = None
    proxy_port = None
    sh = arcrest.AGOLTokenSecurityHandler(username, password=pw)

3. Creating an ArcGIS for Portal Security Handler
import arcrest
if __name__ == "__main__":
    username = "< username >"
    pw = "< password >"
    tokenUrl = "https://mysite.com/portal/sharing/rest/generateToken"
    org_url = "https://mysite.com/portal/sharing/rest"
    sh = arcrest.PortalTokenSecurityHandler(username=username, 
                                            password=pw, 
                                            org_url=org_url, 
                                            token_url=tokenUrl)

In each example, it shows how to generate the token handlers.  You then can pass the security handlers objects onto other functions like arcrest.manageorg.Administration(), which will allow administrators to perform operations to manage a portal or AGOL site.

Hope this helps!


Please download ArcREST here (http://www.github.com/Esri/ArcREST)
If you use ArcREST, help us out and post comments in the issues sections.  We are always trying to make it better.  If you think you can make it better, fork it and submit a pull request!


Monday, April 13, 2015

ArcREST Update - update your forks!

Today I pushed into the master git repository an update that allows users who use Requests to have ArcREST default to that.  If requests if not present, never fear, the library falls back to the standard urllib modules in the basic python install.

Enjoy!

Wednesday, April 8, 2015

ArcREST - Bigger and Better

I've been working hard on ArcREST (https://github.com/esri/arcrest) these days, and the package is catching on I think.  Over the last couple of months, ArcREST has been shown at both Esri's Federal Developer Summit and at the Esri Developer Summit in California.  I truly think the effort is beginning to pay off, and users are reaping the hard work put into this package.

Simply put, this is a toolbox to manage, manipulate, and control your online GIS presence.  Portal, ArcGIS Server, and ArcGIS Online can now be controlled through Python!  The modules present in the Python package are like hammers, nails, and screw drivers.  They provide the framework to do greater things.  By themselves, they do not seem like much, but together you can build your GIS house through the common scripting language of the science community.

I am constantly looking for feedback on this package, so sign up on github, and test it out with a work flow you need in these environments.  If you find a bug, have a question, or need help, please post it in the issue area of github.  

Thank you everyone!