Created
May 21, 2018 19:37
-
-
Save jceloria/36290e0067f326121528fe15a9ed3ba3 to your computer and use it in GitHub Desktop.
Telegraf input plugin for Arris TM1602AP2
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
#!/usr/bin/env python | |
# -*- coding: utf-8 -*- | |
######################################################################################################################## | |
""" | |
Telegraf input plugin for Arris TM1602AP2 | |
Copyright © 2018 by John Celoria. | |
This program is free software; you can redistribute it and/or modify | |
it under the terms of the GNU General Public License as published by | |
the Free Software Foundation; either version 2, or (at your option) | |
any later version. | |
This program is distributed in the hope that it will be useful, | |
but WITHOUT ANY WARRANTY; without even the implied warranty of | |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
GNU General Public License for more details. | |
You should have received a copy of the GNU General Public License | |
along with this program; if not, write to the Free Software Foundation, | |
Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. | |
""" | |
######################################################################################################################## | |
# Library imports | |
from __future__ import print_function | |
from bs4 import BeautifulSoup | |
import argparse | |
import json | |
import re | |
import sys | |
try: | |
# For Python 3.0 and later | |
from urllib.request import urlopen | |
except ImportError: | |
# Fall back to Python 2's urllib2 | |
from urllib2 import urlopen | |
######################################################################################################################## | |
# Set some defaults | |
__author__ = 'John Celoria' | |
__version__ = '0.1' | |
MODEM_URL = 'http://192.168.100.1/cgi-bin/status_cgi' | |
######################################################################################################################## | |
def getUpstreamInformation(html, channels): | |
# Empty list and dictionary | |
AllChannelInfo = [] | |
CurrentChannelInfo = {} | |
# Loop through all detected channels | |
for idx in range(0, channels): | |
# Scrape information from HTML tables | |
ChannelID = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[1] | |
Frequency = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[2] | |
PowerLevel = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[3] | |
SymbolRate = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[5] | |
Modulation = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[6] | |
# Put information from current channel into dictionary | |
CurrentChannelInfo = { | |
"Channel ID": ChannelID.text.encode('ascii', "ignore").strip(), | |
"Frequency": Frequency.text.encode("ascii", "ignore").strip(), | |
"Power Level": PowerLevel.text.encode("ascii", "ignore").strip(), | |
"Symbol Rate": SymbolRate.text.encode("ascii", "ignore").strip(), | |
"Modulation": Modulation.text.encode("ascii", "ignore").strip() | |
} | |
# Add dictionary to list | |
AllChannelInfo.append(CurrentChannelInfo) | |
return AllChannelInfo | |
def getDownstreamInformation(html, channels): | |
# Empty list and dictionary | |
AllChannelInfo = [] | |
CurrentChannelInfo = {} | |
# Loop through all detected channels | |
for idx in range(0, channels): | |
# Scrape information from HTML tables | |
ChannelID = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[1] | |
Frequency = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[2] | |
PowerLevel = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[3] | |
SNR = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[4] | |
Modulation = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[5] | |
CorrectableCodewords = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[7] | |
UncorrectableCodewords = html.findAll("tr")[idx + 1].findAll("td", recursive=False)[8] | |
# Put information from current channel into dictionary | |
CurrentChannelInfo = { | |
"Channel ID": ChannelID.text.encode('ascii', "ignore").strip(), | |
"Frequency": Frequency.text.encode("ascii", "ignore").strip(), | |
"Power Level": PowerLevel.text.encode("ascii", "ignore").strip(), | |
"SNR": SNR.text.encode("ascii", "ignore").strip(), | |
"Modulation": Modulation.text.encode("ascii", "ignore").strip(), | |
"Correctable Codewords": CorrectableCodewords.text.encode("ascii", "ignore").strip(), | |
"Uncorrectable Codewords": UncorrectableCodewords.text.encode("ascii", "ignore").strip() | |
} | |
# Add dictionary to list | |
AllChannelInfo.append(CurrentChannelInfo) | |
return AllChannelInfo | |
def normalize(obj): | |
# Replace spaces w/ underscores, lowercase dictionary keys and strip/convert values to integers | |
return [{k.replace(" ", "_").lower(): float(re.sub("[^0-9\.-]", "", v)) for k, v in d.items()} for d in obj] | |
def main(arguments): | |
parser = argparse.ArgumentParser( | |
description=__doc__, | |
formatter_class=argparse.RawDescriptionHelpFormatter | |
) | |
parser.add_argument("-d", "--downstream", help="Return downstream information", action="store_true") | |
parser.add_argument("-u", "--upstream", help="Return upstream information", action="store_true") | |
parser.add_argument("-r", "--raw", help="Display unmodified raw data", action="store_true") | |
args = parser.parse_args(arguments) | |
if not (args.downstream or args.upstream): | |
parser.error('Please specify one of -d or -u options') | |
# Request status_cgi and initialize BeautifulSoup | |
resp = urlopen(MODEM_URL) | |
body = resp.read() | |
soup = BeautifulSoup(body, 'html.parser') | |
# Parse HTML tables | |
ustream = soup.findAll("table")[3] | |
dstream = soup.findAll("table")[1] | |
# Get the number of channels in use by counting rows | |
uschan = len(ustream.findAll("tr")) - 1 | |
dschan = len(dstream.findAll("tr")) - 1 | |
if args.upstream: | |
upstream_list = getUpstreamInformation(ustream, uschan) | |
if not (args.raw): | |
upstream_list = normalize(upstream_list) | |
# Add telegraf tag | |
[d.update({'stream': 'upstream'}) for d in upstream_list] | |
# JSON-ify our lists | |
print(json.dumps(upstream_list, indent=2, sort_keys=True)) | |
if args.downstream: | |
downstream_list = getDownstreamInformation(dstream, dschan) | |
if not (args.raw): | |
downstream_list = normalize(downstream_list) | |
# Add telegraf tag | |
[d.update({'stream': 'downstream'}) for d in downstream_list] | |
# JSON-ify our lists | |
print(json.dumps(downstream_list, indent=2, sort_keys=True)) | |
if __name__ == '__main__': | |
sys.exit(main(sys.argv[1:])) | |
######################################################################################################################## |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment