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

# In[1]:


from __future__ import print_function

from collections import defaultdict

import requests

import numpy as np
import matplotlib.pyplot as plt
get_ipython().run_line_magic('matplotlib', 'inline')


# In[2]:


def do_request(service, a1=None, a2=None, a3=None, **kwargs):
    server = 'http://api.gbif.org/v1'
    params = ''
    for a in [a1, a2, a3]:
        if a is not None:
            params += '/' + a
    req = requests.get('%s/%s%s' % (server, service, params),
                       params=kwargs,
                       headers={'Content-Type': 'application/json'})
    if not req.ok:
        req.raise_for_status()
    return req.json()


# In[3]:


req = do_request('species', 'search', q='bear')
print(req['count'])
req['results'][0]


# In[4]:


req_short = do_request('species', 'search', q='bear', rank='family')
print(req_short['count'])
bear = req_short['results'][0]
bear


# In[5]:


import time
def get_all_records(rec_field, service, a1=None, a2=None, a3=None, **kwargs):
    records = []
    all_done = False
    offset = 0
    num_iter = 0
    while not all_done and num_iter < 100:  # arbitrary
        req = do_request(service, a1=a1, a2=a2, a3=a3, offset=offset, **kwargs)
        all_done = req['endOfRecords']
        if not all_done:
            time.sleep(1)
        offset += req['limit']
        records.extend(req[rec_field])
        num_iter += 1
    return records


# In[6]:


def get_leaves(nub):
    leaves = []
    recs = get_all_records('results', 'species', str(nub), 'children')
    if len(recs) == 0:
        return None
    for rec in recs:
        rec_leaves = get_leaves(rec['nubKey'])
        if rec_leaves is None:
            leaves.append(rec)
        else:
            leaves.extend(rec_leaves)
    return leaves


# In[7]:


records = get_all_records('results', 'species', str(bear['nubKey']), 'children')
leaves = get_leaves(bear['nubKey'])


# In[8]:


#[leaf['nameType'] for leaf in leaves]
for rec in leaves:
    print(rec['scientificName'], rec['rank'], end=' ')
    vernaculars = do_request('species', str(rec['nubKey']), 'vernacularNames', language='en')['results']
    for vernacular in vernaculars:
        if vernacular['language'] == 'eng':
            print(vernacular['vernacularName'], end='')
            break
    print()


# In[9]:


basis_of_record = defaultdict(int)
country = defaultdict(int)
zero_occurrences = 0
count_extinct = 0
for rec in leaves:
    #print(rec['scientificName'], rec['rank'], rec['taxonID'])
    occurrences = get_all_records('results', 'occurrence', 'search', taxonKey=rec['nubKey'])
    for occurrence in occurrences:
        basis_of_record[occurrence['basisOfRecord']] += 1
        country[occurrence.get('country', 'NA')] += 1
        #there is also publisingCountry
    if len(occurrences) > 0:
        zero_occurrences += 1
    profiles = do_request('species', str(rec['nubKey']), 'speciesProfiles')['results']
    for profile in profiles:
        if profile.get('extinct', False):
            count_extinct += 1
            break


# In[10]:


countries, obs_countries = zip(*sorted(country.items(), key=lambda x: x[1]))
basis_name, basis_cnt = zip(*sorted(basis_of_record.items(), key=lambda x: x[1]))
fig = plt.figure(figsize=(16, 9))
ax = fig.add_subplot(1, 2, 1)
ax.barh(np.arange(10) - 0.5, obs_countries[-10:])
ax.set_title('Top 10 countries per occurences')
ax.set_yticks(range(10))
ax.set_ylim(0.5, 9.5)
ax.set_yticklabels(countries[-10:])
#refer metadata problems

ax = fig.add_subplot(2, 2, 2)
ax.set_title('Basis of record')
ax.bar(np.arange(len(basis_name)), basis_cnt, color='g')
basis_name = [x.replace('OBSERVATION', 'OBS').replace('_SPECIMEN', '') for x in basis_name]
ax.set_xticks(0.5 + np.arange(len(basis_name)))
ax.set_xticklabels(basis_name, size='x-small')

ax = fig.add_subplot(2, 2, 4)
other = len(leaves) - zero_occurrences - count_extinct
pie_values = [zero_occurrences, count_extinct, other]
labels = ['No occurence (%d)' % zero_occurrences,
          'Extinct (%d)' % count_extinct, 'Other (%d)' % other]
ax.pie(pie_values, labels=labels,
       colors=['cyan', 'magenta', 'yellow'])
ax.set_title('Status for each species')


# In[11]:


#distribution of observations of horribilis per year
#probably not put...


# In[12]:


horribilis = get_all_records('results', 'occurrence', 'search', taxonKey=6163845)


# In[13]:


years = defaultdict(int)
for x in horribilis:
    years[x.get('year', None)] += 1
xs, ys = zip(*[(y, cnt) for y, cnt in years.items() if y is not None])
fig = plt.figure()
ax = fig.add_subplot(111)
ax.plot(xs, ys, '.')


# In[ ]: