#!/usr/bin/env python
# coding: utf-8

# Notebook to determine the temporal and spatial coverage of IOS data that Rich has stored.
# 

# Data is stored in several directories: 
# 
# * /ocean/rich/home/SoG/ios/che/
# * /ocean/rich/home/SoG/ios/che2
# * /ocean/nsoontie/MEOPAR/IOS/2011/
# * /ocean/nsoontie/MEOPAR/IOS/2013/

# In[1]:


import numpy as np
import matplotlib.pyplot as plt
import matplotlib.cm as cm
import glob
import os
import datetime
import pandas as pd
import netCDF4 as nc

from salishsea_tools.nowcast import figures
from salishsea_tools import viz_tools
get_ipython().run_line_magic('matplotlib', 'inline')


# In[2]:


#Load NEMO grid
f = '/data/nsoontie/MEOPAR/NEMO-forcing/grid/bathy_meter_SalishSea2.nc'
grid = nc.Dataset(f)


# #Loading data

# In[3]:


def read_IOS(filename):
    """Funtion to read IOS data in filename. Only good for reading lat, lon and date """
    filehandle = open(filename,'rt')
    with filehandle as data:
            headerover = False
            for line in data:
                if headerover:
                    break
                    word = line.split()
                elif line[0:14] == "*END OF HEADER":
                    headerover = True
                else:
                    word = line.split()
                    if len(word) > 1:
                        if word[0] == "LATITUDE" and word[1] !='2':
                            lat = float(word[2])+ float(word[3])/60
                        if word[0] == "LONGITUDE" and word[1] !='2':
                            lon = -(float(word[2])+ float(word[3])/60)
                        if word[0] == "START":
                            string = word[4] +' ' + word[5]
                            time = datetime.datetime.strptime(string,'%Y/%m/%d %H:%M:%S.%f')
    return lat,lon,time


# Loop through files in first directory.

# In[4]:


ios_dir = '/ocean/rich/home/SoG/ios/che/'
files = glob.glob(os.path.join(ios_dir,'*'))
files.sort(key=os.path.basename)
#new IOS data (may have repeats)
files2 = glob.glob(os.path.join('/ocean/nsoontie/MEOPAR/IOS/2011/','*'))
files2.sort(key=os.path.basename)
files.extend(files2)
#another new source
files3=glob.glob(os.path.join('/ocean/nsoontie/MEOPAR/IOS/2013/','*'))
files3.sort(key=os.path.basename)
files.extend(files3)


# In[5]:


times = []
lats=[]
lons=[]
years = []
months = []
counter = 0
for filename in files:
    lat,lon,time = read_IOS(filename)
    times.append(time)
    lats.append(lat)
    lons.append(lon)
    years.append(time.year)
    months.append(time.month)
    counter=counter+1
#Check number of files looked at
print counter
print len(files)


# Loop through files in second directory

# In[6]:


ios_dir = '/ocean/rich/home/SoG/ios/che2/'
files = glob.glob(os.path.join(ios_dir,'*'))
files.sort(key=os.path.basename)


# In[7]:


counter=0
for filename in files:
    lat,lon,time= read_IOS(filename)
    times.append(time)
    lats.append(lat)
    lons.append(lon)
    years.append(time.year)
    months.append(time.month)
    counter=counter+1
print counter
print len(files)


# Add position and dates to a data frame.

# In[8]:


data = pd.DataFrame({'year': years, 'date': times, 'lat':lats, 'lon': lons,'month':months})


# #Plotting

# Where and when do we have data?

# In[9]:


fig,ax=plt.subplots(1,1,figsize=(10,10))
# Group the data by year
grouped = data.groupby('year')
# Grab the keys of this group, (years) and sort 
y = grouped.groups.keys()
y.sort()
#create a colour map for plotting each year a different colour
num_groups = grouped.ngroups
cmap = cm.get_cmap('spectral_r',num_groups)
#iterate through the years an plot location. Give each year a different colour
count = 0
for year in y:
    g = grouped.get_group(year)
    g.plot(x='lon',y='lat',kind='scatter',marker='.',ax=ax,label=year,color=cmap(count))
    count=count+1
viz_tools.plot_coastline(ax,grid,coords='map')
ax.set_xlim([-125.5,-122])
ax.set_ylim([48,50.5])

print grouped.size() #shows number of data points in each year


# * Points on land? I think the model is missing some of those inlets (Sechelt)
# * Lots of points in 2011 because of duplicates. Rob's function can take care of that

# During which months?

# In[10]:


fig,ax=plt.subplots(1,1,figsize=(10,10))
# Group the data by month
grouped = data.groupby('month')
# Grab the keys of this group, (years) and sort 
m = grouped.groups.keys()
m.sort()
#create a colour map for plotting each year a different colour
num_groups = grouped.ngroups
cmap = cm.get_cmap('spectral_r',num_groups)
#iterate through the years an plot location. Give each year a different colour
count = 0
for mon in m:
    g = grouped.get_group(mon)
    g.plot(x='lon',y='lat',kind='scatter',marker='.',ax=ax,label=mon,color=cmap(count))
    count=count+1
viz_tools.plot_coastline(ax,grid,coords='map')
ax.set_xlim([-125.5,-122])
ax.set_ylim([48,50.5])

print grouped.size() #show number of data point in each month


# In[ ]: