"""
DockerList - command ``/usr/bin/docker (images|ps)``
====================================================
Parse the output of command "docker_list_images" and "docker_list_containers",
which have very similar formats.
The header line is parsed and used as the names for the remaining columns.
All fields in both header and data are assumed to be separated by at least
three spaces. This allows single spaces in values and headers, so headers
such as 'IMAGE ID' are captured as is.
If the header line and at least one data line are not found, no data is
stored.
Each row is stored as a dictionary, keyed on the header fields. The data is
available in two formats:
* The old format is a list of row dictionaries.
* The new format stores each dictionary in a dictionary keyed on the value of
a given field, given by the subclass.
"""
from insights.core import CommandParser
from insights.core.exceptions import SkipComponent
from insights.core.plugins import parser
from insights.parsers import parse_fixed_table
from insights.specs import Specs
[docs]
class DockerList(CommandParser):
"""
A general class for parsing tabular docker list information. Parsing
rules are:
* The first line is the header line.
* The other lines are data lines.
* All fields line up vertically.
* Fields are separated from each other by at least three spaces.
* Some fields can contain nothing, and this is shown as spaces, so we
need to catch that and turn it into None.
Why not just use hard-coded fields and columns? So that we can adapt to
different output lists.
Raises:
NotImplementedError: If `key_field` or `attr_name` is not defined
SkipComponent: If no data to parse
"""
key_field = None
heading_ignore = []
attr_name = ''
substitutions = []
[docs]
def parse_content(self, content):
"""
Parse the lines given into a list of dictionaries for each row. This
is stored in the ``rows`` attribute.
If the ``key_field`` property is set, use this to key a ``data``
dictionary attribute.
"""
if not (self.key_field and self.attr_name):
raise NotImplementedError("'key_field' or 'attr_name' is not defined")
# There is another application named docker that's a kde system tray, that
# will output help when the spec is run due to incorrect arguments. So check
# the content for any lines starting with Usage: so it can be skipped.
if any(l for l in content if l.startswith("Usage: ")):
raise SkipComponent('No data only help output.')
self.rows = parse_fixed_table(content,
heading_ignore=self.heading_ignore,
header_substitute=self.substitutions)
if not self.rows:
raise SkipComponent('No data.')
data = {}
for row in self.rows:
k = row.get(self.key_field)
for sub in self.substitutions:
row[sub[0]] = row.pop(sub[1]) if sub[1] in row else None
if k is not None and k != '<none>':
data[k] = row
setattr(self, self.attr_name, data)
[docs]
@parser(Specs.docker_list_images)
class DockerListImages(DockerList):
"""
Handle the list of docker images using the DockerList parser class.
Sample output of command ``docker images --all --no-trunc --digests``::
REPOSITORY TAG DIGEST IMAGE ID CREATED VIRTUAL SIZE
rhel7_imagemagick latest <none> 882ab98aae5394aebe91fe6d8a4297fa0387c3cfd421b2d892bddf218ac373b2 4 days ago 785.4 MB
rhel6_nss-softokn latest <none> dd87dad2c7841a19263ae2dc96d32c501ee84a92f56aed75bb67f57efe4e48b5 5 days ago 449.7 MB
Attributes:
rows (list): List of row dictionaries.
images (dict): Dictionary keyed on the value of the "REPOSITORY" fileld
Examples:
>>> images.rows[0]['REPOSITORY']
'rhel6_vsftpd'
>>> images.rows[1]['VIRTUAL SIZE']
'785.4 MB'
>>> images.images['rhel6_vsftpd']['CREATED']
'37 minutes ago'
"""
key_field = 'REPOSITORY'
heading_ignore = [key_field]
attr_name = 'images'
substitutions = [("IMAGE ID", "IMAGE_ID"), ("VIRTUAL SIZE", "VIRTUAL_SIZE")]
@property
def data(self):
return self.images
[docs]
@parser(Specs.docker_list_containers)
class DockerListContainers(DockerList):
"""
Handle the list of docker images using the DockerList parser class.
Sample output of command ``docker ps --all --no-trunc --size``::
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES SIZE
95516ea08b565e37e2a4bca3333af40a240c368131b77276da8dec629b7fe102 bd8638c869ea40a9269d87e9af6741574562af9ee013e03ac2745fb5f59e2478 "/bin/sh -c 'yum install -y vsftpd-2.2.2-6.el6'" 51 minutes ago Exited (137) 50 minutes ago tender_rosalind 4.751 MB (virtual 200.4 MB)
03e2861336a76e29155836113ff6560cb70780c32f95062642993b2b3d0fc216 rhel7_httpd "/usr/sbin/httpd -DFOREGROUND" 45 seconds ago Up 37 seconds 0.0.0.0:8080->80/tcp angry_saha 796 B (virtual 669.2 MB)
Attributes:
rows (list): List of row dictionaries.
containers(dict): Dictionary keyed on the value of the "NAMES" field
running_containers(list): List of NAMEs of the running containers.
Examples:
>>> containers.rows[0]['STATUS']
'Up 37 seconds'
>>> containers.containers['tender_rosalind']['STATUS']
'Exited (137) 18 hours ago'
"""
key_field = 'NAMES'
heading_ignore = ['CONTAINER']
attr_name = 'containers'
substitutions = [("CONTAINER ID", "CONTAINER_ID")]
def __init__(self, *args, **kwargs):
super(DockerListContainers, self).__init__(*args, **kwargs)
self.running_containers = []
for name, info in self.containers.items():
if info.get('STATUS', '').startswith('Up '):
self.running_containers.append(name)
@property
def data(self):
return self.containers