def augmentCollection(request, collection_id, seed_level):
if request.method == 'GET':
if collection_id in names:
return HttpResponse(json.dumps({warning: 'Collection already exists! Try with another collection id.'}), status=status.HTTP_304_NOT_MODIFIED)
seeds = Evidence.objects.filter(Q(created_by=collection_id)&~Q(abstract='')&Q(augmentation=seed_level))
counter = 0
start = 306
for e in seeds:
counter += 1
if counter < start:
continue
print 'processing entry #' + str(counter) + ' out of ' + str(seeds.count())
unicodeTitle = e.title.encode('utf-8')
related_evidence, citation_map, pmid = PubMedQuerier.get_related_evidence(unicodeTitle)
print 'found ' + str(len(related_evidence)) + ' related evidence for ' + unicodeTitle
refs, citedin = getRefsAndCitedin(pmid, citation_map)
for re in related_evidence:
if not re.title:
continue
re_object = Evidence.objects.create_evidence(re.title, re.abstract, serializePaperMetadata(re.pmid,re.authors_str,re.journal,re.year,''), collection_id, int(seed_level)+1)
if re.pmid in refs:
Citation.objects.get_or_create(paper_id=e.id, citation_id=re_object.id, collection_id=collection_id)
if re.pmid in citedin:
Citation.objects.get_or_create(paper_id=re_object.id, citation_id=e.id, collection_id=collection_id)
return HttpResponse(json.dumps({}), status=status.HTTP_200_OK)
def augmentCollection(request, collection_id, seed_level):
if request.method == 'GET':
if collection_id in names:
return HttpResponse(json.dumps({warning: 'Collection already exists! Try with another collection id.'}), status=status.HTTP_304_NOT_MODIFIED)
seeds = Evidence.objects.filter(Q(created_by=collection_id)&~Q(abstract='')&Q(augmentation=seed_level))
counter = 0
for e in seeds:
counter += 1
print 'processing entry #' + str(counter) + ' out of ' + str(seeds.count())
# print e.title
# counter += 1
# if counter < 183:
# continue
# if counter >
related_evidence = PubMedQuerier.get_related_evidence(e.title)
try:
print 'found ' + str(len(related_evidence)) + ' related evidence for ' + e.title
except UnicodeEncodeError:
pass
for re in related_evidence:
Evidence.objects.create_evidence(re.title, re.abstract, json.dumps({
'PMID': re.pmid,
'AUTHOR': re.authors_str,
'JOURNAL': re.journal,
'DATE': re.year,
'AFFILIATION': ''
}), collection_id, seed_level+1)
return HttpResponse(json.dumps({}), status=status.HTTP_200_OK)
def completeCitationInfo(request, collection_id):
if request.method == 'GET':
if collection_id == '18':
completeCitationInfoAMiner(collection_id)
else:
evidence = Evidence.objects.filter(created_by=collection_id)
start = 546
counter = 1
for e in evidence:
if counter < start:
counter += 1
continue
print '>> Processing entry ' + str(counter) + ' out of ' + str(evidence.count())
unicodeTitle = e.title.encode('utf-8')
related_evidence, citation_map, pmid = PubMedQuerier.get_related_evidence(unicodeTitle)
refs, citedin = getRefsAndCitedin(pmid, citation_map)
print 'found ' + str(len(refs)) + ' ref, ' + str(len(citedin)) + ' citedin'
for re in related_evidence:
re_objects = Evidence.objects.filter(title=re.title,created_by=collection_id)
# here paper_id cites citation_id
for re_object in re_objects:
if re.pmid in refs:
try:
Citation.objects.get_or_create(paper_id=e.id, citation_id=re_object.id, collection_id=collection_id)
except MultipleObjectsReturned:
print e.id
print re_object.id
if re.pmid in citedin:
try:
Citation.objects.get_or_create(paper_id=re_object.id, citation_id=e.id, collection_id=collection_id)
except MultipleObjectsReturned:
print re_object.id
print e.id
counter += 1
return HttpResponse(json.dumps({}), status=status.HTTP_200_OK)
def process_title_query(request): if request.method == 'POST': query = request.POST['query'] response_data = pq.extract_terms_for_titles(query, min_repeat=1) print 'Extracted terms:' print response_data return HttpResponse(json.dumps(response_data), content_type='application/json')
def search_pubs(request):
if request.method == 'POST':
query = request.POST.get('query')
response_data = pq.search_pubs(query)
print response_data
return HttpResponse(json.dumps(response_data), content_type='application/json')
def get(self, request, format=None):
terms = json.loads(request.body)['terms']
evidence = PubMedQuerier.find_evidence_for_terms(terms, skip_no_abstract=True)
serialized_json = serializers.serialize('json', evidence)
evidence_json = flattenSerializedJson(serialized_json)
# let's provide topic modeling results in addition to the raw evidence
output = {}
output['evidence'] = json.loads(evidence_json)
evidencePks = [e.pk for e in evidence]
abstracts = [e.abstract for e in evidence]
output['topics'], output['evidenceTopicMap'] = getTopicsForDocuments(evidencePks, abstracts)
return HttpResponse(json.dumps(output), status=status.HTTP_201_CREATED)
def post(self, request, format=None):
data = json.loads(request.body)
neighbors = PubMedQuerier.find_neighbors_for_terms(data['concepts'], num_neighbors=10, user_id=data['requested_by'])
output = {}
output['counts'] = {}
concepts = []
for pairs in neighbors['keywords']:
output['counts'][pairs[0]] = pairs[1]
concepts.append(Concept.objects.create_concept(pairs[0], data['requested_by']))
serialized_json = serializers.serialize('json', concepts)
concepts_json = flattenSerializedJson(serialized_json)
output['concepts'] = json.loads(concepts_json)
return HttpResponse(json.dumps(output), status=status.HTTP_201_CREATED)
def post(self, request, user_id, format=None):
data = json.loads(request.body)
# data = {
# 'abstract': '',
# 'title': 'Generalized theory of relaxation',
# 'metadata': '',
# 'created_by': 1001
# }
# 01/20/2016 new feature: initiate a google scholar api call to get abstract if not provided
findRelatedEvidence = True;
title = data['title'].replace('{', '').replace('}', '')
abstract = data['abstract']
if abstract == '':
temp_title, temp_abstract = PubMedQuerier.get_abstract_by_title(data['title'])
if temp_title is not None:
title = temp_title
abstract = temp_abstract
evidence = Evidence.objects.create_evidence(title, abstract, data['metadata'], data['created_by'], 0)
serialized_json = serializers.serialize('json', [evidence])
evidence_json = flattenSerializedJson(serialized_json)
return HttpResponse(evidence_json, status=status.HTTP_201_CREATED)
def post(self, request, format=None):
params = json.loads(request.body)
terms = params['terms']
user_id = params['user_id']
texts = Text.objects.filter(created_by=user_id)
serialized_json = serializers.serialize('json', texts)
texts_json = flattenSerializedJson(serialized_json)
evidenceCreated = Evidence.objects.filter(created_by=user_id)
evidenceBookmarks = EvidenceBookmark.objects.filter(user_id=user_id)
evidenceBookmarkedIds = [eb.evidence.pk for eb in evidenceBookmarks]
evidenceBookmarked = Evidence.objects.filter(pk__in=evidenceBookmarkedIds)
evidenceRetrieved = PubMedQuerier.find_evidence_for_terms(terms, skip_no_abstract=True)
evidence = chain(evidenceCreated, evidenceBookmarked, evidenceRetrieved)
print '>> serializing evidence...'
serialized_json = serializers.serialize('json', evidence)
print '>> flatten serialized evidence...'
evidence_json = flattenSerializedJson(serialized_json)
# let's provide topic modeling results in addition to the raw evidence
output = {}
print '>> loading evidence into json...'
output['evidence'] = json.loads(evidence_json)
contents = [t.content for t in texts]
textPks = ['t-'+str(t.pk) for t in texts]
abstracts = [e['abstract'] for e in output['evidence']]
evidencePks = ['e-'+str(e['id']) for e in output['evidence']]
if len(evidencePks + textPks) <= 1:
output['topics'] = []
output['evidenceTextTopicMap'] = []
else:
output['topics'], output['evidenceTextTopicMap'] = getTopicsForDocuments(evidencePks + textPks, abstracts + contents)
return HttpResponse(json.dumps(output), status=status.HTTP_201_CREATED)
def process_term_query(request):
if request.method == 'POST':
query = request.POST.getlist('query[]')
response_data = pq.find_neighbors_for_terms(query)
print response_data
return HttpResponse(json.dumps(response_data), content_type='application/json')
