def load_document_to_es(settings):
"""
load document to elastic search
:param settings:
:return:
"""
doc_folder = settings.get_attr(['epr_index', 'doc_folder'])
d2p_tsv = settings.get_attr(['epr_index', 'doc2patient_tsv'])
es = SemEHRES.get_instance_by_setting(
settings.get_attr(['epr_index', 'es_host']),
settings.get_attr(['epr_index', 'es_index_name']),
settings.get_attr(['epr_index', 'doc_type']), '', '')
tsv_lines = utils.read_text_file(d2p_tsv)
d2p = {}
for l in tsv_lines:
arr = l.split('\t')
if len(arr) > 1:
d2p[arr[0]] = arr[1]
for f in [f for f in listdir(doc_folder) if isfile(join(doc_folder, f))]:
if f in d2p:
p = d2p[f]
t = utils.read_text_file_as_string(join(doc_folder, f))
es.index_new_doc(
index=settings.get_attr(['epr_index', 'es_index_name']),
doc_type=settings.get_attr(['epr_index', 'doc_type']),
data={
settings.get_attr(['epr_index', 'text_field']): t,
settings.get_attr(['epr_index', 'patient_id_field']): p,
"id": f
},
doc_id=f)
def patient_level_indexing(settings, pids):
es = SemEHRES.get_instance_by_setting(
settings.get_attr(['patient_index', 'es_host']),
settings.get_attr(['patient_index', 'patient_index']),
settings.get_attr(['patient_index', 'patient_doct_type']),
settings.get_attr(['patient_index', 'es_concept_type']),
settings.get_attr(['patient_index', 'es_patient_type']))
doc_level_index = settings.get_attr(['patient_index', 'doc_level_index'])
doc_ann_type = settings.get_attr(['patient_index', 'doc_ann_type'])
doc_index = settings.get_attr(['patient_index', 'doc_index'])
doc_pid_field_name = settings.get_attr(
['patient_index', 'doc_pid_field_name'])
doc_text_field_name = settings.get_attr(
['patient_index', 'doc_text_field_name'])
patient_index = settings.get_attr(['patient_index', 'patient_index'])
patient_doct_type = settings.get_attr(
['patient_index', 'patient_doct_type'])
doc_type = settings.get_attr(['patient_index', 'doc_type'])
ann_field_name = settings.get_attr(['patient_index', 'ann_field_name'])
num_procs = 10 if settings.get_attr(['patient_index', 'num_procs']) is None else \
settings.get_attr(['patient_index', 'num_procs'])
ignore_exist = True if settings.get_attr(['patient_index', 'ignore_exist']) is None else \
settings.get_attr(['patient_index', 'ignore_exist'])
utils.multi_process_tasking(lst=pids,
num_procs=num_procs,
process_func=do_patient_indexing,
args=[
es, doc_level_index, doc_ann_type,
doc_index, doc_type, doc_pid_field_name,
doc_text_field_name, patient_index,
patient_doct_type, ann_field_name,
ignore_exist
])
def index_ctx_concept(ann, concept_index, ctx_doc_type, es_inst):
data = {
"doc": {
"cui": ann['cui'],
"negation": ann['negation'],
"experiencer": ann['experiencer'],
"temporality": ann['temporality'],
"prefLabel": ann['pref'],
"STY": ann['sty']
},
"doc_as_upsert": True
}
ctx_id = SemEHRES.get_ctx_concept_id(ann)
es_inst.update_doc(index=concept_index, doc_type=ctx_doc_type, data=data, doc_id=ctx_id)
def es_get_cohort_docs(settings):
pids = utils.read_text_file(settings.get_attr(['cohort_docs', 'es_cohort_file']))
es = SemEHRES.get_instance_by_setting(settings.get_attr(['cohort_docs', 'es_host']),
settings.get_attr(['cohort_docs', 'es_index']),
settings.get_attr(['cohort_docs', 'es_doc_type']),
settings.get_attr(['cohort_docs', 'es_concept_type']),
settings.get_attr(['cohort_docs', 'es_patient_type']))
patiet_id_field = settings.get_attr(['cohort_docs', 'patiet_id_field'])
docs = []
docs2p = {}
for pid in pids:
container = []
cohort_analysis_helper.query_collect_patient_docs({'_id': pid}, es, '*', patiet_id_field, container)
if len(container) > 0:
docs += [{'docid': d} for d in container[0]['docs']]
for d in container[0]['docs']:
docs2p[d] = pid
return docs, docs2p, pids
def es_populate_patient_study_table_post_ruled(study_analyzer, out_file, rule_executor,
sample_size, sample_out_file, ruled_ann_out_file,
es_conn_file, text_preprocessing=False,
retained_patients_filter=None,
filter_obj=None):
"""
populate patient study result with post processing to remove unwanted mentions
:param cohort_name:
:param study_analyzer:
:param out_file:
:param rule_executor:
:param sample_size:
:param sample_out_file:
:return:
"""
es = SemEHRES.get_instance_by_setting_file(es_conn_file)
if filter_obj is not None:
fes = SemEHRES.get_instance_by_setting_file(filter_obj['doc_es_setting'])
patient_id_field = filter_obj['patient_id_field']
filter_obj['es'] = fes
if retained_patients_filter is None:
pids = es.search_by_scroll("*", es.patient_type)
else:
pids = retained_patients_filter
patients = [{'brcid': p} for p in pids]
id2p = {}
for p in patients:
id2p[p['brcid']] = p
print 'total patients is %s' % len(patients)
non_empty_concepts = []
study_concepts = study_analyzer.study_concepts
term_to_docs = {}
ruled_anns = []
for sc in study_concepts:
positive_doc_anns = []
sc_key = '%s(%s)' % (sc.name, len(sc.concept_closure))
print 'working on %s' % sc_key
if sc.name.startswith('ess_'):
non_empty_concepts.append(sc_key)
# elasticsearch concepts
p2docs = chelper.query_doc_by_search(es, fes, sc.concept_closure, patient_id_field,
retained_patients_filter=retained_patients_filter,
filter_obj=filter_obj, doc_filter_function=patient_timewindow_filter)
for pd in p2docs:
id2p[pd['pid']][sc_key] = str(len(pd['docs']))
# continue without to do the rest
continue
doc_anns = []
if len(sc.concept_closure) > 0:
doc_anns = chelper.query_doc_anns(es, sc.concept_closure, study_analyzer.skip_terms,
retained_patients_filter=retained_patients_filter,
filter_obj=filter_obj, doc_filter_function=patient_timewindow_filter
)
if len(doc_anns) > 0:
p_to_dfreq = {}
counted_docs = set()
for d in doc_anns:
doc = doc_anns[d]
p = doc['pid']
if d in counted_docs:
continue
for ann in doc['anns']:
ruled, rule = rule_executor.execute(doc['text'] if not text_preprocessing else
preprocessing_text_befor_rule_execution(doc['text']),
int(ann['s']),
int(ann['e']))
if not ruled:
counted_docs.add(d)
p_to_dfreq[p] = 1 if p not in p_to_dfreq else 1 + p_to_dfreq[p]
positive_doc_anns.append({'id': d,
'content': doc['text'],
'annotations': [{'start': ann['s'],
'end': ann['e'],
'concept': ann['inst']}]})
else:
ruled_anns.append({'p': p, 'd': d, 'ruled': rule})
if len(counted_docs) > 0:
non_empty_concepts.append(sc_key)
for p in p_to_dfreq:
id2p[p][sc_key] = str(p_to_dfreq[p])
# save sample docs
if sample_size >= len(positive_doc_anns):
term_to_docs[sc_key] = positive_doc_anns
else:
sampled = []
for i in xrange(sample_size):
index = random.randrange(len(positive_doc_anns))
sampled.append(positive_doc_anns[index])
positive_doc_anns.pop(index)
term_to_docs[sc_key] = sampled
concept_labels = sorted(non_empty_concepts)
s = '\t'.join(['brcid'] + concept_labels) + '\n'
for p in patients:
s += '\t'.join([p['brcid']] + [p[k] if k in p else '0' for k in concept_labels]) + '\n'
utils.save_string(s, out_file)
utils.save_json_array(convert_encoding(term_to_docs, 'cp1252', 'utf-8'), sample_out_file)
utils.save_json_array(convert_encoding(ruled_anns, 'cp1252', 'utf-8'), ruled_ann_out_file)
print 'done'
def do_semehr_doc_anns_analysis(settings):
anns_folder = settings.get_attr(['doc_ann_analysis', 'ann_docs_path'])
text_folder = settings.get_attr(['doc_ann_analysis', 'full_text_folder'])
full_text_file_pattern = settings.get_attr(
['doc_ann_analysis', 'full_text_fn_ptn'])
rule_config = settings.get_attr(['doc_ann_analysis', 'rule_config_path'])
output_folder = settings.get_attr(['doc_ann_analysis', 'output_folder'])
study_folder = settings.get_attr(['doc_ann_analysis', 'study_folder'])
combined_anns = settings.get_attr(['doc_ann_analysis', 'combined_anns'])
es_output_index = settings.get_attr(
['doc_ann_analysis', 'es_output_index'])
es_output_doc = settings.get_attr(['doc_ann_analysis', 'es_output_doc'])
output_file_pattern = settings.get_attr(
['doc_ann_analysis', 'output_fn_pattern'])
thread_num = settings.get_attr(['doc_ann_analysis', 'thread_num'])
if thread_num is None:
thread_num = 10
process_mode = settings.get_attr(['doc_ann_analysis', 'process_mode'])
if process_mode is not None and process_mode != 'sql':
if settings.get_attr(['doc_ann_analysis', 'es_host']) is not None:
es = SemEHRES.get_instance_by_setting(
settings.get_attr(['doc_ann_analysis', 'es_host']),
settings.get_attr(['doc_ann_analysis', 'es_index']),
settings.get_attr(['doc_ann_analysis', 'es_doc_type']),
settings.get_attr(['doc_ann_analysis', 'es_concept_type']),
settings.get_attr(['doc_ann_analysis', 'es_patient_type']))
docanalysis.process_doc_anns(
anns_folder=anns_folder,
full_text_folder=text_folder,
rule_config_file=rule_config,
output_folder=output_folder,
study_folder=study_folder,
full_text_fn_ptn=full_text_file_pattern,
fn_pattern=output_file_pattern,
thread_num=thread_num,
es_inst=es,
es_text_field=settings.get_attr(
['doc_ann_analysis', 'full_text_field']),
patient_id_field=settings.get_attr(
['doc_ann_analysis', 'patielt_id_field']),
combined_anns=combined_anns,
es_output_index=es_output_index,
es_output_doc=es_output_doc)
else:
docanalysis.process_doc_anns(
anns_folder=anns_folder,
full_text_folder=text_folder,
rule_config_file=rule_config,
output_folder=output_folder,
study_folder=study_folder,
full_text_fn_ptn=full_text_file_pattern,
fn_pattern=output_file_pattern,
thread_num=thread_num)
else:
ann_list_sql = settings.get_attr(['doc_ann_analysis', 'ann_list_sql'])
primary_keys = settings.get_attr(['doc_ann_analysis', 'primary_keys'])
ann_inst_sql = settings.get_attr(['doc_ann_analysis', 'ann_inst_sql'])
full_text_sql = settings.get_attr(
['doc_ann_analysis', 'full_text_sql'])
update_query_template = settings.get_attr(
['doc_ann_analysis', 'update_query_template'])
update_status_template = settings.get_attr(
['doc_ann_analysis', 'update_status_template'])
dbconn_file = settings.get_attr(['doc_ann_analysis', 'dbconn_file'])
docanalysis.analyse_db_doc_anns(
ann_list_sql,
ann_inst_sql,
primary_keys,
update_query_template,
full_text_sql,
dbconn_file,
thread_num=thread_num,
study_folder=study_folder,
rule_config_file=rule_config,
update_status_template=update_status_template)
