#!/usr/bin/python

# program to interogate 1-wire bus and report all known temperatures
# with some commentary

# by ian
# This code has no warranty, including any implied fitness for purpose
# or implied merchantability.  It's free, and worth every penny.

# standard library modules used in code
import sys
import time
import argparse

# statistics from scipy
from scipy import stats

# use OWFS module
import ow

# set up parser with command summary
parser = argparse.ArgumentParser(
        description='Rolling 1-wire temperatures report')
# set up arguments with associated help and defaults 
parser.add_argument('-i', 
        dest='sample_interval', 
        help='interval in seconds between samples', 
        default='30')
parser.add_argument('-t', 
        dest='trend_time', 
        help='time in seconds over which trend is calculated', 
        default='300')

# process the arguments
args=parser.parse_args()

# turn the arguments into numbers
sample_interval=float(args.sample_interval)
trend_time=int(args.trend_time)

# tell the user what is happening
print "1 wire bus reports"
print "  sample interval: "+str(sample_interval) + " seconds"
print "  report trend over "+ str(trend_time) + " seconds"

# connect to localhost port 4304 where owserver should be running
ow.init('4304');

# now determine what sensors we will consider
print "locate sensors and take initial readings:"
# every sensor on the bus added to the list
list = ow.Sensor('/').sensorList()

# create lists for teh data we will accumulate
sensorlist=[]
sensordata=[]
readtimes=[]

# get time in seconds now
tt=int(time.time())

# number of temperature sensors found so far
n=0

for s in list:
    sys.stdout.write("  consider " + str(s) + ": ")
    if 'temperature' in s.entryList():
        # get teh temperature from that sensor
        T = s.temperature
        # increment count of sensors
        n = n + 1  
        # tell the user teh good news
        sys.stdout.write("T: " + T)
        # record that as a sensor to interrogate
        sensorlist.append(s)
        # set found sensors to use uncached results
        # so will run new conversion every read cycle
        s.useCache( False )
        # record teh time of reading and value read
        sensordata.append([float(T)])
        readtimes.append([tt])
    else:
        # give the user the bad news
        sys.stdout.write("no temperature value")
    print

print "bus search done - " + str(n) + " temperature sensors found"
print

# now loop forever reading the identified sensors
while 1:
    # determine how long to wait until next interval
    # note this will skip some if specified interval is too short
    #  - it finds the next after now, not next after last 
    sleeptime = sample_interval - (time.time() % sample_interval)
    time.sleep(sleeptime) 

    # get time now and record it
    tt = int(time.time())  # we only record to integer seconds
    
    # show the time
    print (str(tt))
    
    # work through list of sensors
    for s in sensorlist:
        n = sensorlist.index(s)  # the array register of this sensor
        try:  # just in case it has been unplugged
            T=float(s.temperature)
        except ow.exUnknownSensor:  # it has been unplugged
            print '  sensor ' + str(s) + ' gone away - just ignore'
            continue  # so we'll jump to teh next in the list

        # record value and time of this reading
        sensordata[n].append(T)
        readtimes[n].append(tt)

        # but we only want to keep data that's not too old
        # so drop any values older than  trend_time value
        while ((tt-readtimes[n][0]) > trend_time):
            readtimes[n].pop(0)
            sensordata[n].pop(0)

        # print sensor name and current value
        print '  {!s}: {:-6.2f}'.format(s,T)

        # if we have at least three samples, print aggregate info
        if len(readtimes[n]) >= 3:
            slope, intercept, r_value, p_value, std_err = \
                stats.linregress(readtimes[n],sensordata[n])
            print '     max {:-6.2f}, min {:-6.2f}, avg {:-6.2f}'.format( \
                max(sensordata[n]),min(sensordata[n]), \
                    sum(sensordata[n])/len(sensordata[n]))
            print '     trend {:-8.4f} per second'.format(slope)
            print '     from {:d} samples over {:d} seconds'.format( \
                len(sensordata[n]),max(readtimes[n])-min(readtimes[n]))
                
    print