import requests
ls

r=requests.get("http://api.crossref.org/works/10.1021/jm0203783")
data=r.json()

data['message']['DOI']
# u'10.1021/jm0203783'
data['message']['title'][0]
# Comparative Evaluation of 11 Scoring Functions for Molecular Docking
data['message']['container-title']
# [u'J. Med. Chem.', u'Journal of Medicinal Chemistry']
data['message']['issued']['date-parts'][0][0]
# 2003, [[2003,6]]
data['message']['volume']
# 46
data['message']['issue']
# 12
data['message']['page']
# u'2287-2303'
data['message']['ISSN']
# [u'0022-2623', u'1520-4804']
data['message']['URL']
# u'http://dx.doi.org/10.1021/jm0203783'
data['message']['author']
# [{u'affiliation': [], u'family': u'Wang', u'given': u'Renxiao'},
#  {u'affiliation': [], u'family': u'Lu', u'given': u'Yipin'},
#  {u'affiliation': [], u'family': u'Wang', u'given': u'Shaomeng'}]

r.status_code
# 200
# 无效DOI返回响应码404

r=requests.get("http://api.crossref.org/works?query=Comparative%20Evaluation%20of%2011%20Scoring%20Functions%20for%20Molecular%20Docking&rows=5")
data=r.json()
data['message']
#->facet: 空白字典, 不明
#->items: 一个结果列表,每一项是个字典, 每个字典就是上面的data['message']获得的内容. 其中score项是相关性打分
#->total-results: 搜出的结果总数
#->items-per-page: 5, 由关键词rows=5限定
#->query: 搜索内容, 有'search-terms' 就是搜索query部分; 'start-index' 不明.

class CrossRefAPI(object):
	'''Simple wrapper for CrossRef API'''

	def filterstr(self,filterdict):
		'''Generate string for filter'''
		filters=[]
		for k,v in filterdict.items():
			filters.append(k+':'+str(v))
		return ','.join(filters)

	def getqueryurl(self,query,rows=10,offset=0,papertype='journal-article',sample=0,filterdict={},cursor=''):
		'''Get query url, support rows/offset/cursor/sample
		Default get journal-article. Support more filters in dict format'''
		if (not query):
			return ""
		query=str(query)
		# Basic options
		if (sample>0):
			optstrbase='sample='+str(sample)
		elif (cursor):
			optstrbase="rows="+str(rows)+'&cursor='+cursor
		elif offset<=0:
			optstrbase="rows="+str(rows)
		else:
			optstrbase="rows="+str(rows)+"&offset="+str(offset)

		optstr=optstrbase
		# Deal with filter
		if (papertype):
			optstr+="&filter=type:"+papertype
			if (filterdict):
				optstr+=','+self.filterstr(filterdict)
		elif (filterdict):
			optstr+='&filter='+self.filterstr(filterdict)

		if (len(query) == 9 and query[4] == '-'):
			return "http://api.crossref.org/journals/"+query+"/works?"+optstr
		elif ('10.' in query):
			return "http://api.crossref.org/prefixes/"+query+"/works?"+optstr
		elif ( query.isdigit()):
			return "http://api.crossref.org/members/"+query+"/works?"+optstr
		else:
			return "http://api.crossref.org/works?query="+query+"&"+optstr

	def getquerytotol(self,query,papertype=None,filterdict={}):
		'''Get total result number for a query'''
		if (not query):
			return 0
		j1= self.getqueryurl(query,rows=1,papertype=papertype,filterdict=filterdict)
		if (j1):
			r=urllib2.urlopen(j1,timeout=timeout_setting)
			j=json.loads(r.read())
			r.close()
			return int(j['message'].get('total-results',0))
		else:
			return 0

	def getalldoi2file(self,query,fname,papertype=None,filterdict={},getissn=False):
		'''Get All doi for a query to a file.
		getissn can also analyse all issn for this query'''
		if (not query or not fname):
			return
		issns=set()
		total=self.getquerytotol(query,papertype=papertype,filterdict=filterdict)
		f=open(fname,'w')
		print "Total articles:",total
		f.write("# Total articles: "+str(total)+'\n')
		step=1000
		maxround=total/step+1
		nextcursor="*"
		for i in range(maxround):
			needurl=self.getqueryurl(query,rows=step,cursor=nextcursor,papertype=papertype,filterdict=filterdict)
			r=urllib2.urlopen(needurl,timeout=timeout_setting_download)
			js=json.loads(r.read())
			newcursor=js['message'].get('next-cursor',nextcursor)
			if nextcursor == newcursor:
				break
			nextcursor=newcursor
			for j in js['message']['items']:
				doi=j.get("DOI","")
				if (getissn):
					issn=j.get('ISSN')
					issns.update(issn)
				f.write(doi+'\n')
		if (getissn):
			for issn in issns:
				f.write('# '+issn+'\n')
		f.close()

	def is_doaj(self,query):
		'''Test a query is in DOAJ, often use doi'''
		return self.getquerytotol(query,papertype=None,filterdict={'directory':'DOAJ'})>0

class RandomGet(object):
	'''Wrapper for random get for query list'''
	def __init__(self,query,randnum=100):
		if (not isinstance(query,list)):
			raise ValueError("Should give a list to initialize!")
			sys.exit(1)
		self.query=query
		self.crapi=CrossRefAPI()
		self.qdois={}
		for q in self.query:
			self.qdois[q]=self.sampledois(q,randnum=randnum)

	def sampledois(self,query,randnum=100):
		if not query:
			return []
		needurl=self.crapi.getqueryurl(query, sample=randnum)
		if (needurl):
			r=urllib2.urlopen(needurl,timeout=timeout_setting)
			js=json.loads(r.read())
			r.close()
			return [ item.get("DOI","") for item in js['message'].get('items',[{}]) ]
		return []