Example: cat hunters

... or how to integrate groups of tasks with similar options.

This section will present a more complex use case where we have three commands: bigdog, catkiller and eagle which purpose is to find cats.

We will start by integrating bigdog to secator before realizing that most options can be mutualized between the three tools, and a common output type Cat can be created for all three.

Bigdog

Let's suppose we have a fictional utility called bigdog which purpose is to hunt cats on the internet. We want to add bigdog to secator.

Here are some of bigdog's options:

bigdog can be run on a single site using -site:

$ bigdog -site loadsofcats.com
   / \__
  (    @\___   =============
  /         O  BIGDOG v1.0.0
 /   (_____/   =============
/_____/
garfield [boss, 14]
tony [admin, 18]

bigdog can be run on a list of sites using -list:

$ bigdog -list sites.txt -json
   / \__
  (    @\___   =============
  /         O  BIGDOG v1.0.0
 /   (_____/   =============
/_____/
garfield [boss, 14]
romuald [minion, 5]
tony [admin, 18]

bigdog can output JSON lines using -json:

$ bigdog -site loadsofcats.com -json
   / \__
  (    @\___   =============
  /         O  BIGDOG v1.0.0
 /   (_____/   =============
/_____/
{"name": "garfield", "age": 14, "host": "loadsofcat.com", "position": "boss"}
{"name": "tony", "age": 18, "host": "loadsofcats.com", "position": "admin"}

A basic definition of bigdog using basic secator concepts will be:

secator/tasks/bigdog.py

from secator.runners import Command
from secator.decorators import task


@task
class bigdog(Command):
    cmd = 'bigdog'
    json_flag = '-json'
    input_flag = '-site'
    file_flag = '-list'

You can now run bigdog from the CLI or the library:

secator x bigdog --help
secator x bigdog loadsofcats.com

>>> from secator.tasks import bigdog

# Get all results as a list, blocks until command has finished running
>>> bigdog('loadsofcats.com').run()
[
    {"name": "garfield", "age": 14, "host": "loadofcats.com", "position": "boss"},
    {"name": "tony", "age": 18, "host": "loadsofcats.com", "position": "admin"}
]

# Get result items in real-time as they arrive to stdout
>>> for cat in bigdog('loadsofcats.com'):
>>> ... print(cat['name'] + '(' + cat['age'] + ')')
garfield (14)
tony (18)

Okay, this is a good start.

Now what if the bigdog command has some more options that you would like to integrate ?

-timeout allows to specify a request timeout.
-rate allows to specify the max requests per minute.

You can add the opts parameter to your Command object to define the cmd options:

from secator.runners import Command
from secator.decorators import task


@task
class bigdog(Command):
    cmd = 'bigdog'
    json_flag = '-json'
    input_flag = '-site'
    file_flag = '-list'
    opt_prefix = '-'
    opts = {
        'timeout': {'type': int, 'help': 'Timeout (in seconds)'},
        'rate': {'type': int, 'help': 'Max requests per minute'}
    }

You can now use bigdog with this set of options:

secator x bigdog --help
secator x bigdog loadsofcats.com -json
secator x bigdog loadsofcats.com -timeout 1 -rate 100 -o table,csv,txt,gdrive

from secator.tasks import bigdog
bigdog('loadsofcats.com', rate=100, timeout=3).run()  # adding rate and timeout options

Cat hunters category

One advantage of having class-based definitions is that we can group similar tools together in categories.

Let's assume we have 2 other tools that can hunt cats: catkiller and eagle...

... but each of those tools might be written by a different person, and so the interface and output is different for each of them:

$ catkiller --host loadsofcats.com --max-wait 1000 --max-rate 10 --json
Starting catkiller session ...
{"_info": {"name": "tony", "years": 18}, "site": "loadsofcats.com", "job": "admin"}
{"_info": {"name": "garfield", "years": 14}, "site": "loadsofcats.com", "job": "boss"}

# or to pass multiple hosts, it needs to be called like:
$ cat hosts.txt | catkiller --max-wait 1000 --max-rate 10 --json

Inputs:

--host is equivalent to bigdog's -site.
--max-wait is equivalent to bigdog's -timeout, but in milliseconds instead of seconds.
--max-rate is equivalent to bigdog's -rate.
--json is equivalent to bigdog's -json option, but uses a different option character "--".
cat hosts.txt | catkiller is the equivalent tobigdog's -list.

Output:

_info has the data for name and age, but age is now years.
site is the equivalent of bigdog's host.
job is the equivalent of bigdog's position.

$ eagle -u loadsofcats.com -timeexpires 1 -jsonl
                  _      
                 | |     
  ___  __ _  __ _| | ___ 
 / _ \/ _` |/ _` | |/ _ \
|  __/ (_| | (_| | |  __/  v2.2.0
 \___|\__,_|\__, |_|\___|
             __/ |       
            |___/       
{"alias": "tony", "occupation": "admin", "human_age": 105}

# or to pass multiple hosts, it needs to be called like:
$ eagle -l hosts.txt -timeexpires 1 -jsonl
                  _      
                 | |     
  ___  __ _  __ _| | ___ 
 / _ \/ _` |/ _` | |/ _ \
|  __/ (_| | (_| | |  __/  v2.2.0
 \___|\__,_|\__, |_|\___|
             __/ |       
            |___/    
{"alias": "tony", "occupation": "admin", "human_age": 105, "host": "loadsofcats.com"}

Inputs:

-u is equivalent to bigdog's -site.
-l is equivalent to bigdog's -list.
-timeexpires is equivalent to bigdog's -timeout.
eagle does not support setting the maximum requests per seconds (bigdog's -rate).
-jsonl is the flag to output JSON lines, instead of bigdog's -json.

Output:

alias is the equivalent of bigdog's name.
occupation is the equivalent of bigdog's job.
human_age is the human age conversion of the cat age.

Cat output type

We first define a base Cat dataclass to define the common output schema and a CatHunter category as an input interface.

We take bigdog's output schema as reference to create the Cat output type:

secator/output_types/cat.py

from secator.definitions import OPT_NOT_SUPPORTED
from secator.output_types import OutputType
from secator.decorators import task
from dataclasses import dataclass, field

# common output type `Cat`
@dataclass
class Cat(OutputType):
    name: str
    age: int
    alive: bool = False
    _source: str = field(default='', repr=True)
    _type: str = field(default='ip', repr=True)
    _uuid: str = field(default='', repr=True, compare=False)

    _table_fields = [name, age]
    _sort_by = (name, age)

    def __str__(self) -> str:
        return self.ip

CatHunter category

We take bigdog's options names as reference and add the ones that can be mutualized to the CatHunter category:

secator/categories.py

from secator.output_types import Cat
...

class CatHunter(Command):
    meta_opts = {
        'timeout': {'type': int, 'default': 1, 'help': 'Timeout (in seconds)'},
        'rate': {'type': int, 'default': 1000, 'help': 'Max requests per minute'},
    }
    output_types = [Cat]

Tools implementation

Finally we inherit all commands implementation from CatHunter and write the option mapping for the remaining cat-hunter commands:

secator/tasks/bigdog.py

from secator.categories import CatHunter
from secator.decorators import task


@task
class bigdog(CatHunter):
    cmd = 'bigdog'
    json_flag = '-json'
    input_flag = '-site'
    file_flag = '-list'
    opt_prefix = '-'

secator/tasks/catkiller.py

from secator.categories import CatHunter
from secator.decorators import task
from secator.output_types import Cat


@task
class catkiller(CatHunter):
    cmd = 'catkiller'
    json_flag = '--json'
    input_flag = '--host'

    # stdin-like input using 'cat <FILE> | <COMMAND>'
    file_flag = None

    # catkiller options start with "--" unlike the other tools
    opt_prefix = '--' 

    # Map `catkiller` options to CatHunter.meta_opts
    opt_key_map = {
        'rate': 'max-rate'
        'timeout': 'max-wait'
    }
    opt_value_map = {
        'timeout': lambda x: x / 1000 # converting milliseconds to seconds
    }

    # Map `catkiller` output schema to Cat schema
    output_map = {
	Cat: {
	    'name': lambda x: x['_info']['name'], # note: you can use any function, we use
	    'age': lambda x: x['_info']['age'],   #       lambdas for readability here
	    'host': 'site',   # 1:1 mapping
	    'job': 'job' # 1:1 mapping
	}
    }

secator/tasks/eagle.py

from secator.categories import CatHunter
from secator.decorators import task
from secator.definitions import OPT_NOT_SUPPORTED
from secator.output_types import Cat


@task
class eagle(CatHunter):
    cmd = 'eagle'
    json_flag = '-jsonl'
    input_flag = '-u'
    file_flag = '-l'

    # Map `eagle` input options to CatHunter.meta_opts
    opt_key_map = {
        'rate': 'timeexpires',
        'timeout': OPT_NOT_SUPPORTED # explicitely state that this option not supported by the target tool
    }

    # Map `eagle` output schema to Cat schema:
    output_map = {
	Cat: {
	    'name': 'alias',
	    'age': lambda x: human_to_cat_age(x['human_age']),
	    'job': 'occupation',
	}
    }

    # Add 'host' key dynamically after the item has been converted to the output schema,
    # since `eagle` doesn't return the host systematically.
    @staticmethod
    def on_item(self, item):
        item['host'] = item.get('host') or self.input
        return item


def human_to_cat_age(human_age):
    cat_age = 0
    if human_age <= 22:
        cat_age = human_age // 11
    else:
        cat_age = (human_age - 22) // 5 + 2
    return cat_age

Using these definitions, we can now use all the cat-hunter commands with a common interface (input options & output schema):

secator x bigdog loadsofcats.com -rate 1000 -timeout 1 -json
secator x eagle loadsofcats.com -rate 1000 -timeout 1 -json
secator x catkiller loadsofcats.com -rate 1000 -timeout 1 -json

>>> from secator.tasks import bigdog, catkiller, eagle
>>> meta_opts = {'timeout': 1, 'rate': 1000, 'json': True}
>>> bigdog('loadsofcats.com', **meta_opts).run()
[
    Cat(name="garfield", age=14, host="loadsofcats.com", position="boss", _source="bigdog"),
    Cat(name="tony", age=18, host="loadsofcats.com", position="admin", _source="bigdog")
]
>>> catkiller('catrunner.com', **meta_opts).run()
[
    Cat(name=fred, age=12, host="catrunner.com", position="minion", _source="catkiller"},
    Cat(name=mark, age=20, host="catrunner.com", position="minion", _source="catkiller"}
]
>>> eagle('allthecats.com', **meta_opts).run()
[
    Cat(name="marcus", age=4, host="allthecats.com", position="minion", _source="eagle"},
    Cat(name="rafik", age=7, host="allthecats.com", position="minion", _source="eagle"}
]

PreviousExample: ls NextAdvanced options

Last updated 5 months ago