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

# This notebook creates a jackknifed PCoA plot based on multiple rarefactions.

# In[ ]:


from emperor import Emperor, nbinstall

nbinstall()

from skbio.stats.ordination import pcoa
from skbio.diversity import beta_diversity

from biom import load_table

import pandas as pd
import numpy as np

def load_mf(fn, index='#SampleID'):
    _df = pd.read_csv(fn, sep='\t', dtype=str, keep_default_na=False, na_values=[])
    _df.set_index(index, inplace=True)
    return _df


# We are going to load data from [Fierer et al. 2010](http://www.pnas.org/content/107/14/6477.full) (the data was retrieved from study [232](https://qiita.ucsd.edu/study/description/232) in [Qiita](https://qiita.ucsd.edu), remember you need to be logged in to access the study).

# In[ ]:


mf = load_mf('keyboard/mapping-file.txt')
bt = load_table('keyboard/otu-table.biom')


# Create 5 rarefactions and compute the Jaccard distance matrix for each resulting table.

# In[ ]:


ordinations = []
for r in range(5):
    rarefied = bt.subsample(1000)
    data = np.array([rarefied.data(i) for i in rarefied.ids()], dtype='int64')
    
    res = pcoa(beta_diversity('jaccard', data, rarefied.ids()))
    
    ordinations.append(res)


# Jackknifed plots need a *master* set of coordinates where there rest of the matrices will be centered around.

# In[ ]:


master = ordinations[0]
jackknifed = ordinations[1:]


# If you want to share your notebook via GitHub use `remote=True` and make sure you share your notebook using nbviewer. Change the jackknifing method to standard deviation.

# In[ ]:


viz = Emperor(master, mf, jackknifed=jackknifed, remote=False)
viz.jackknifing_method = 'sdev'
viz


# Change the jackknifing method to inter quantile ranges:

# In[ ]:


viz.jackknifing_method = 'IQR'
viz