Convert sas2ircu output to json for easier use in other tools

sas2ircu_json.py

#!/usr/bin/env python3

"""
Usage:
  ./sas_devices.py - json output
  ./sas_devices.py --summary - human output with locations and serial
  
Warning: sas2ircu invocations have occassionally caused kernel
      panics on previous FreeBSD installations
      Related https://redmine.ixsystems.com/issues/14859

Note: This was only tested on a setup with a single controller
      which had a single enclosure attached. Parsing may not 
      be sufficient for other configurations

Field naming conventions were copied directly from command output
in places where field names have been hardcoded

sas2ircu: Version 20.00.00.00 (2014.09.18) 
os: FreeBSD nas 12.2-RELEASE-p3 FreeBSD 12.2-RELEASE-p3 7851f4a452d(HEAD) TRUENAS  amd64

By: Nick Venenga <nick@venenga.com>
"""

import json
import re
import shutil
import subprocess
import sys

if not shutil.which("sas2ircu"):
    sys.stderr.write(f"sas2ircu: Command not found.\n")
    sys.exit(1)

cmd_output = subprocess.check_output(["sas2ircu", "list"], text=True).splitlines()
lines = iter(cmd_output)
# skip lines til hyphen/space divider
while not re.match(r"^[- ]+$", next(lines)):
    pass

controllers = [
    dict(
        zip(
            [
                "Index",
                "Adapter Type",
                "Vendor ID",
                "Device ID",
                "Pci Address",
                "SubSys Ven ID",
                "SubSys Dev ID",
            ],
            line.split(),
        )
    )
    for line in lines
    if line.strip() and line.strip()[0].isnumeric()
]


def parse_controller_display(index):
    cmd_output = subprocess.check_output(
        ["sas2ircu", index, "display"],
        text=True,
    )
    sections = "|".join(
        re.escape(s)
        for s in [
            "Controller information",
            "IR Volume information",
            "Physical device information",
            "Enclosure information",
        ]
    )
    section_finder = re.compile(
        "(" + sections + ")\n[-]{8,}\n(.*?)[-]{8,}\n", flags=re.DOTALL | re.MULTILINE
    )
    results = {}
    for header, data in section_finder.findall(cmd_output):
        # This section is "special" in that it has "sub-sections" denoted by non-indented fields
        # Each subsection is a device
        if header == "Physical device information":
            results.update(
                {
                    header: [
                        {
                            line.split(":")[0].strip(): line.split(":", 1)[1].strip()
                            for line in device.strip().splitlines()
                        }
                        for device in re.split(r"^Device is .+\n", data, flags=re.M)
                        if re.match("\s+[^\s]+", device)
                    ]
                }
            )
        elif header == "IR Volume information":
            results.update(
                {
                    header: [
                        {
                            line.split(":")[0].strip(): line.split(":", 1)[1].strip()
                            for line in device.strip().splitlines()
                        }
                        for device in re.split(r"^IR volume .+\n", data, flags=re.M)
                        if re.match("\s+[^\s]+", device)
                    ]
                }
            )
        else:
            results.update(
                {
                    header: {
                        line.split(":")[0].strip(): line.split(":", 1)[1].strip()
                        for line in data.splitlines()
                    }
                }
            )
    return results


devices = [parse_controller_display(controller["Index"]) for controller in controllers]

if len(sys.argv) == 2 and sys.argv[1] == "--summary":
    print(
        "\n".join(
            [
                f'Controller: {c["Controller information"]["Device"]} Enclosure: {device["Enclosure #"]} Slot: {device["Slot #"]} Serial: {device["Serial No"]}'
                for c in devices
                for device in c["Physical device information"]
                if "Drive Type" in device
                # if any(typ in device.get("Drive Type") for typ in ("SAS", "SATA"))
            ]
        )
    )
else:
    print(json.dumps(devices, indent=2))