steelconnection 0.9.9

Simplify access to the Riverbed SteelConnect REST API.

# SteelConnection
##### version 0.9.9
SteelConnection provides a wrapper object to simplify access to the Riverbed SteelConnect REST API.

* Create an object once and it remembers the URL and authentication.
* All subsequent API requests are made via a single TCP connection.
* Call the appropriate method and supply the resource path and any upload data.
* Accepts and returns native Python data - no need to convert to or from JSON.

##### BETA software:
_actively working to simplify - some behavior may change before 1.0.0 release._

## Supports:
Python 2.7, 3.4, 3.5, 3.6, 3.7

## Requires:
Requests

## HOWTO:
#### To install:
```bash
# Python 3
python3 -m pip install steelconnection

# Python 2
python2 -m pip install steelconnection
```

#### To upgrade your copy to the latest version:
```bash
# Python 3
python3 -m pip install --upgrade steelconnection

# Python 2
python2 -m pip install --upgrade steelconnection
```

#### Quickstart:
See the examples folder for sample scripts.\
https://github.com/grelleum/SteelConnection/tree/master/examples

#### Getting Started:
* Make sure the REST API is enabled on your SteelConnect realm before trying to access the REST API.
* Use pip to install steelconnection as shown above.
* Import steelconnection and create a new object by providing the Fully qualified DNS name of your realm. This would typically be `REALM_NAME.riverbed.cc`, where `REALM_NAME` is specific to your realm.
```python
import steelconnection
sc = steelconnection.SConAPI('REALM.riverbed.cc')
```

#### Realms and Organizations:
There is a one to one relationship between a Realm and a SteelConnect Manager. The SteelConnect Manager acts as the controller for a the realm. A newly created realm would not have any organizations, otherwise a realm will have one or more organizations. Each oganization within a realm acts an autonomous network system. In practice, most REST API operations are performed within a specific organization.

You normally access the SteelConnect Manager (SCM) using a web browser.\
The URL you use includes the realm and organization that you are managing and takes the form:
`https://realm.riverbed.cc/admin/Organization`.\
The Organization is case-sensistive and is also known as the organization short name, as opposed to the longname, which is more descriptive and can include spaces, and other characters.

#### Authentication:
SteelConnect REST API uses Basic Auth, meaning a username and password are required for authentication for every request made. The steelconnection object can store the username and pssword for you, or you can use a .netrc file as detailed below. Choose one of the following methods:

##### Interactive login (Optional):
If you do not specify a username and password, and a .netrc file is not configured, steelconnection will interactively prompt you for your username and password. Steelconnection will validate the login by making a 'status' call against the REST API.
```python
>>> import steelconnection
>>> sc = steelconnection.SConAPI('REALM.riverbed.cc')
Enter username: admin
Enter password:
>>>
```

##### Using a .netrc file (Optional):
A .netrc file can be used to store credentials on Mac, Unix, and Linux machines. This file would be stored in the root of your home directory. When using a .netrc file, steelconnection will never have your password, rather the underlying requests library will be responsible for accessing the .netrc file. Use the commands below to setup a .netrc file, replacing REALM, USERNAME, and PASSWORD with your actual values.
```bash
echo "machine REALM.riverbed.cc login USERNAME password PASSWORD" >> ~/.netrc
chmod 600 ~/.netrc
```

##### Specifying username and password (Optional):
If you prefer to use some other method to obtain the username and password, you can supply those as the time of object creation using the username and password keywaord argumets.\
For example, if you want to store your credentials in your system environment variables you could do something similar to the following:
```python
import os
import steelconnection

username = os.environ.get('SCONUSER')
password = os.environ.get('SCONPASSWD')

sc = steelconnection.SConAPI('REALM.riverbed.cc', username=username, password=password)
```

#### Understanding the API:
The Riverbed SteelConnect REST API allows HTTPS access to the SteelConnect Manager (SCM) via the use of GET, POST, PUT, and DELETE commands. SteelConneciton (this module) acts to simplify coding by providing an object that remembers your realm, version, and authentication and builds the HTTPS requests based on that information. A `requests.session` object is used to allow a single TCP connection to be re-used for all subsequent API requests.

**With** SteelConnection, a request to get a list of all organizations in the realm would look like this:
```python
orgs = sc.get('orgs')
```
**Without** SteelConnection, the same request would look like this:
```python
response = requests.get(
'https://REALM.riverbed.cc/api/scm.config/1.0/orgs',
auth=(username, password)
)
orgs = response.json()['items']
```

##### Available Methods:
SteelConneciton provides the `.get`, `.getstatus`, `.post`, `.put`, and `.delete` methods to simplify access to the API.\
These methods will build the request to include api version, auth, etc, so you onlu need to specify the resource you are interrested in.

* get: Used for retrieving information about a resource. Expect data to be returned.
* getstatus: Used for retrieving current status about a resource. Expect data to be returned.
* post: Create or deploy a new resource. Requires additional data in the payload and returns the newly created object.
* put: Use to edit or update some existing resource. Requires additional data in the payload.
* delete: Delete an existing resource.

##### Two APIs:
Riverbed divides the REST API into two APIs:
* Config: used to make configurations changes and get information about SteelConnect resources.\
https://support.riverbed.com/apis/scm_beta/scm-2.10.0/scm.config/index.html
* Reporting: used to get current status information about a resource.\
https://support.riverbed.com/apis/scm_beta/scm-2.10.0/scm.reporting/index.html

By nature, the Reporting API only requires the HTTP GET method, where-as the more commonly used Confg API requires GET, POST, PUT and DELETE. SteelConnections combines the two APIs by implementing `.get`, `.post`, `.put`, and `.delete` methods to access to Config API and the `.getstatus` method to access the Reporting API.

For example: Calling `.get('/port/' + port)` would retireve configuration settings on a port, where-as `.getstatus('/port/' + port)` would retreive the actual link state, speed, duplex, etc. for that port.

##### Crafting your API calls:
The Riverbed documentation describes the various REST API calls that can be made.\
These take the form: "_HTTP Method_" "_resource path_".

Take the network section for example:\
https://support.riverbed.com/apis/scm_beta/scm-2.10.0/scm.config/index.html#!/network:
* `GET` `/networks` List networks.
* `GET` `/org/:orgid/networks` Get network for an org.
* `POST` `/org/:orgid/networks` Create network within an org.
* `DELETE` `/networks/:netid` Delete network.
* `GET` `/networks/:netid` Get network.
* `PUT` `/networks/:netid` Update a network.

Within the resource path, you may see a name preceded by a colon `:`. These are considered variables and must be replaced with an actual value. The `/networks/:netid` would require the `:netid` be replaced with the actual network ID for the network you are requesting.

SteelConnection methods mimic the HTTP Methods and accept the short form resource paths.\
To update a network, the documentation lists `PUT` `/networks/:netid`. With the SteelConnection object, you would call the put method as `sc.put('/network/' + net_id)`. Note that the leading `/` in the resource is optional as the SteelConnection object will insert it if it is missing.

##### Model Schema (Data Payload):
Post (create) and Put (update) requests require additional data in the form of a payload. This gets sent to the server in the form of JSON data, however the SteelConnection object will accept either JSON data or a native Python dictionary (`isinstance(data, dict)`). The Riverbed documentation will specify the format of the data as a "Model Schema". Not everything listed in the model schema is required. Generally, you can determine the minimum required data by checking the equivalent function in SteelConnect Manager web GUI.

#### Retrieving Data:
The SteelConnection methods leverage the popular requests package. Methods calls always return a native Python dictionary, or a list of dictionaries, depending on the API call. The `requests.response` object will be stored as an attribute of the object (`sc.response`) so the latest response is always easily accessible. By providing the full `requests.response` object you are free to check status and see all headers.

For example, the 'get orgs' request should always provide a list of orgs within the realm, so we can directly assign the result as a native Python list.\
`list_of_all_orgs = sc.get('orgs')`

Here are the rules to determine what gets returned by an API request:
* If response.json() is True and the 'items' key exists, then return a python list of response.json()['items'].
* If response.json() is True and the 'items' key _does not_ exist, then return a python dictionary.
* If response.json() is False, return an empty python dictionary.

##### Binary Data:
In the event an API call returns binary data, as is the case with "GET /node/:nodeid/get_image", the return value of the call will include a message on how to access the binary data. You can access it directly through the object's '.response.content' attribute, or by calling the '.savefile(filename)' method, which will save the binary data to a file.

#### Errors and Exceptions:
The **_Zen of Python_** states:
> Errors should never pass silently.\
Unless explicitly silenced.

With this in mind, steelconnection assumes all REST API calls should complete without error. Succeful requests will return with an HTTP 200-level response. Any other response if considered a failed request and will cause steelconnection to raise either a `RuntimeError`, or a custom exceptions that inherits from `RuntimeError`. Exception handling can be used to catch the exception:
```python
try:
sc.put(f'node/{node_id}', data={'location': 'LAB'})
except RuntimeError as e:
handle_exception(e)
```
More specific exceptions that might be generated (all inherit from `RuntimeError`):
* `steelconnection.exceptions.AuthenticationError`: 401 - Check your username and password.
* `steelconnection.exceptions.APINotEnabled`: 502 - Check if REST API is enabled on Realm settings.
* `steelconnection.exceptions.BadRequest`: 400 - Possibly tried creating a resource that already exists.
* `steelconnection.exceptions.InvalidResource`: 404 - Path or resource not found.

If you prefer to handle errors manually and do not want steelconnection to generate exceptions based on HTTP response code, you can instead use the child class `SConAPIwithoutExceptions` to create your object. The `SConAPIwithoutExceptions` class replaces the exception handling method with a method that does nothing. The steelconnection object will evaluate as `True` after a successful request and `False` otherwise. This reflects the status of the obect attribute `.response.ok`.
```python
sc = SConAPIwithoutExceptions('REALM.riverbed.cc')
```

#### Convienience functions:
##### Object-level Convienience functions:
The SteelConnect Manager stores resources in a database with a uniquie identifier (id). Many API calls require that you know the id number of the resource you are interested in.\
SteelConnection provides a collection of `lookup` functions to look up the resources based on known values. These functions return a tuple of the id and the actual resouce.\
Currently these are the available lookup functions:
* `lookup.org(org_short_name)`
* `lookup.node(serial)`
* `lookup.site(site_name, org_id=org_id)`

These functions are accessed directly from the object you created and are specific to the SteelConnect API.

###### Lookup Organization:
Many REST API calls require that you know the org id of your organization. You can provide the organization short name to the function and it will return the org id and the org object.
```python
>>> org_id, org = sc.lookup.org('Spacely')
>>> org_id
'org-Spacely-0a0b1cbadb33f34'
>>>
```
###### Lookup Node:
Similarly, the `lookup.node` method exists to privide the node id and node when you supply the commonly known appliance serial number.
```python
>>> node_id, node = sc.lookup.node('XN00012345ABCDEF')
>>> node_id
'node-56f1968e222ab789'
>>>
```
###### Lookup Site:
The site id can be found in a similar way, but since the same site name, like HQ, could exist in multiple organizations, the org_id is required.
```python
>>> site_id, site = sc.lookup.site('Skypad', orgid='org-Spacely-0a501e7f27b2c03e')
>>> site_id
'site-Skypad-884b9071141e4bc0'
>>>
```

##### Module-level Convienience functions:
These functions are accessed directly from the imported module and can be used independently of the SteelConnect API.

###### Get Input:
`get_input(prompt)` function works with both Python 2 and Python 3 to get user input.

###### Get Username:
`get_username(prompt)` function works with both Python 2 and Python 3 to get username.

###### Get Password:
`get_password(prompt)` function works with both Python 2 and Python 3 to get user input. Uses getpass to provide discretion. Requires user to input password to be typed twice for verification.