CODEWITHSUNDEEP
pythonlinuxjson
Dayan
Dayan

Reputation: 8031

Python iterate a subprocess that returns json?

I have a subprocess executing:

lshw -json -C network

If i receive the following back:

    {
    "id" : "network",
    "class" : "network",
    "claimed" : true,
    "handle" : "PCI:0000:00:05.0",
    "description" : "Ethernet interface",
    "product" : "82545EM Gigabit Ethernet Controller (Copper)",
    "vendor" : "Intel Corporation",
    "physid" : "5",
    "businfo" : "pci@0000:00:05.0",
    "logicalname" : "eth0",
    "version" : "00",
    "serial" : "00:8c:42:77:58:49",
    "units" : "bit/s",
    "size" : 1000000000,
    "capacity" : 1000000000,
    "width" : 32,
    "clock" : 66000000,
    "configuration" : {
      "autonegotiation" : "on",
      "broadcast" : "yes",
      "driver" : "e1000",
      "driverversion" : "7.3.21-k8-NAPI",
      "duplex" : "full",
      "firmware" : "N/A",
      "ip" : "10.211.55.10",
      "latency" : "0",
      "link" : "yes",
      "multicast" : "yes",
      "port" : "twisted pair",
      "speed" : "1Gbit/s"
    },
    "capabilities" : {
      "msi" : "Message Signalled Interrupts",
      "bus_master" : "bus mastering",
      "cap_list" : "PCI capabilities listing",
      "ethernet" : true,
      "physical" : "Physical interface",
      "logical" : "Logical interface",
      "tp" : "twisted pair",
      "10bt" : "10Mbit/s",
      "10bt-fd" : "10Mbit/s (full duplex)",
      "100bt" : "100Mbit/s",
      "100bt-fd" : "100Mbit/s (full duplex)",
      "1000bt-fd" : "1Gbit/s (full duplex)",
      "autonegotiation" : "Auto-negotiation"
    }
  },

Can i possibly iterate over this to ensure i capture all network interfaces (in the case that there's more than one) which is not the case with my system.. Also, how can i pick 1 or two from this output, i don't need the entire data.

I had the following in mind:

 def get_nic_data():
        lshw_cmd = "lshw -json -C network"
        proc = subprocess.Popen(lshw_cmd, shell=True, stdout=subprocess.PIPE,
                                                      stdin=subprocess.PIPE)
        return proc.stdout


 def read_data(proc_output):
        import simplejason as json
        json_obj = json

        json_obj.loads(proc_output)

        #Obtain Vendor,Description,Product
        #...
        #...

        json_obj.dumps(obtained_data_here)

        #Not sure if this would work this way.


  read_data(get_nic_data())

Upvotes: 3

Views: 6067

Answers (2)

Martijn Pieters
Martijn Pieters

Reputation: 1123410

Unfortunately, you cannot combine -C class filtering with -json output; even in the latest version JSON output is severely broken. Instead, filter the full JSON output yourself. Note that you should avoid using shell=True when using subprocess, pass in a list instead; no need to pipe stdin either, but do capture (silence) stderr.

Then we can recurse over the 'children' structures, picking out anything that has a matching 'class' key:

def get_nic_data():
    lshw_cmd = ['lshw', '-json']
    proc = subprocess.Popen(lshw_cmd, stdout=subprocess.PIPE,
                                      stderr=subprocess.PIPE)
    return proc.communicate()[0]

def find_class(data, class_):
    for entry in data.get('children', []):
        if entry.get('class') == class_:
            yield entry

        for child in find_class(entry, class_):
            yield child

def read_data(proc_output, class_='network'):
    import json

    for entry in find_class(json.loads(proc_output), class_):
        yield entry['vendor'], entry['description'], entry['product']

then loop over read_data(get_nic_data()):

for vendor, description, product in read_data(get_nic_data()):
    print vendor, description, product

Upvotes: 8

jfs
jfs

Reputation: 414585

If there are multiple network cards lshw doesn't return a valid json text. It could be fixed by adding [ / ] before / after the output and adding , in between objects:

import json
import re
from subprocess import STDOUT, check_output as qx

# get command output
output = qx("lshw -json -C network".split(), stderr=STDOUT)

# get json parts
_, sep, json_parts = output.rpartition(b"\r")
if not sep: # no \r in the output
    json_parts = output

# convert it to valid json list
jbytes = b"[" + re.sub(b"}( *){", b"}, {", json_parts) + b"]"
L = json.loads(jbytes.decode())

# pretty print
import sys
json.dump(L, sys.stdout, indent=4)

A cleaner solution would use lshw -xml that produces output that could be easily converted to a well-formed xml by wrapping it in a root element: '<root>' + output + '</root>'.

Upvotes: 0

Related Questions