def _read_file(filename):
try:
with open(os.path.join(resources, 'cache.pickle'), 'rb') as infile:
result, v = pickle.load(infile)
assert v == __version__
assert not get_versions()['dirty']
return result
except:
icd9cm_expanded_vocab = ICD9CM(use_decimals=True)
icd9pcs_vocab = ICD9PCS(use_decimals=True)
icd9mixed_vocab = ICD9CM(use_decimals=True) | ICD9PCS(
use_decimals=True)
icd10cm_vocab = ICD10CM(use_decimals=True)
icd10pcs_vocab = ICD10PCS(use_decimals=True)
icd10mixed_vocab = ICD10CM(use_decimals=True) | ICD10PCS(
use_decimals=False)
result = {}
with open(filename, 'rt') as infile:
reader = csv.reader(infile)
next(reader)
for row in reader:
if row[3].strip().replace('/', '').upper() != 'NA':
all_codes = icd9mixed_vocab.parse(row[3])
result[(row[1], row[2], icd9cm_expanded_vocab.vocab_domain,
icd9cm_expanded_vocab.vocab_name
)] = icd9cm_expanded_vocab.filter(all_codes)
result[(row[1], row[2], icd9pcs_vocab.vocab_domain,
icd9pcs_vocab.vocab_name
)] = icd9pcs_vocab.filter(all_codes)
# res = (all_codes) - (result[(row[1], row[2], 'icd9cm')] | result[(row[1], row[2], 'icd9pcs')])
# if res:
# print 'icd9:', res
if row[4].strip().replace('/', '').upper() != 'NA':
all_codes = icd10mixed_vocab.parse(row[4])
result[(row[1], row[2], icd10cm_vocab.vocab_domain,
icd10cm_vocab.vocab_name
)] = icd10cm_vocab.filter(all_codes)
result[(row[1], row[2], icd10pcs_vocab.vocab_domain,
icd10pcs_vocab.vocab_name
)] = icd10pcs_vocab.filter(all_codes)
# res = (all_codes) - (result[(row[1], row[2], 'icd10cm')] | result[(row[1], row[2], 'icd10pcs')])
# if res:
# print 'icd10:', res
with open(os.path.join(resources, 'cache.pickle'), 'wb') as outfile:
pickle.dump((result, __version__), outfile)
return result
def test_icd9_cm():
vocab = ICD9CM(match_terminal_only=True)
assert_equal(vocab.parse("250.33, 250.40, 250.41, 250.42, 250.43,"),
{'250.33', '250.40', '250.41', '250.42', '250.43'})
assert_equal(vocab.parse("250.33, 250.40-250.43,"),
{'250.33', '250.40', '250.41', '250.42', '250.43'})
assert_equal(vocab.parse("250.33, 250.4*,"),
{'250.33', '250.40', '250.41', '250.42', '250.43'})
assert_equal(vocab.parse('42'), {'042'})
def test_code_sets():
icd9cm = ICD9CM()
icd9pcs = ICD9PCS()
icd10cm = ICD10CM()
icd10pcs = ICD10PCS()
for k, v in code_sets.collectlevels().items():
assert k[-1] in {'ICD9CM', 'ICD10CM', 'ICD9PCS', 'ICD10PCS'}
assert k[-2] in {'DX', 'PX'}
if k[-1] == 'icd9cm':
assert_equal(icd9cm.filter(v), v)
assert_equal(k[-2], 'DX')
elif k[-1] == 'icd10cm':
assert_equal(icd10cm.filter(v), v)
assert_equal(k[-2], 'DX')
if k[-1] == 'icd9pcs':
assert_equal(icd9pcs.filter(v), v)
assert_equal(k[-2], 'PX')
elif k[-1] == 'icd10pcs':
assert_equal(icd10pcs.filter(v), v)
assert_equal(k[-2], 'PX')
def code_system_standardizers(icd10cm=dict(),
icd10pcs=dict(),
icd9cm=dict(),
icd9pcs=dict(),
hcpcs=dict(),
loinc=dict(),
ndc=dict(),
ubrev=dict()):
result = dict(
ICD10PCS=ICD10PCS(**icd10cm).standardize,
ICD10CM=ICD10CM(**icd10pcs).standardize,
ICD9CM=ICD9CM(**icd9cm).standardize,
ICD9PCS=ICD9PCS(**icd9pcs).standardize,
HCPCS=HCPCS(**hcpcs).standardize,
LOINC=LOINC(**loinc).standardize,
NDC=NDC(**ndc).standardize,
UBREV=UBREV(**ubrev).standardize,
)
return result
# from clinvoc.icd9 import ICD9PCS, ICD9CM
import resources
import os
import re
from base import CCS
from clinvoc.icd9 import ICD9CM, ICD9PCS
import pandas as pd
from toolz.dicttoolz import merge
# '''
# Single category parsing
# '''
icd9cm_vocab = ICD9CM(use_decimals=False)
icd9pcs_vocab = ICD9PCS(use_decimals=False)
def _get_icd9_codes(filename, code_type):
assert code_type in ['dx', 'px']
vocab = icd9cm_vocab if code_type == 'dx' else icd9pcs_vocab
file_path = os.path.join(resources.resources, filename)
df = pd.read_csv(file_path)
code_column = df.columns[0]
result = {}
for _, row in df.iterrows():
key = (re.sub('\[[^\]]*\]', '', row[2]).strip(), re.sub('\[[^\]]*\]', '', row[4]).strip(), re.sub('\[[^\]]*\]', '', row[6]).strip(), vocab.vocab_domain, vocab.vocab_name)
# if not key[2]:
# key = key[:2] + key[1:2] + key[3:]
if key not in result:
result[key] = set()
result[key].add(vocab.standardize(row[code_column].strip('\'')))
return result
from clinvoc.icd9 import ICD9CM # This string describes a set of ICD 9 codes codestring = '745.0-745.3, 745.6*, 746, 747.1-747.49, 747.81, 747.89, 35.8, 35.81, 35.82, 35.83, 35.84' # Use clinvoc to parse and standardize the above codes vocab = ICD9CM() codeset = vocab.parse(codestring) print(sorted(codeset))
import os
import pandas as pd
from clinvoc.icd9 import ICD9CM
from clinvoc.code_collections import CodeCollection
from modulecache.invalidators import FileChangeInvalidator
from modulecache.backends import PickleBackend
from .resources import resources
vocab = ICD9CM(use_decimals=True)
def _construct_icd9_categories(def_column="Enhanced ICD-9-CM"): # args?
# df = pd.read_csv(os.path.join(resources, 'NDF_January_2016.xlsx'))
df = pd.read_csv(os.path.join(resources, 'elixhauser_definitions.csv'))
elix_cats = {}
for _, row in df.iterrows():
category = row['Category']
codes = row[def_column]
parsed_codes = vocab.parse(codes)
elix_cats[(category, vocab.vocab_domain, vocab.vocab_name)] = parsed_codes
return elix_cats
cache_filename = os.path.join(os.path.abspath(os.path.dirname(__file__)), 'elix_icd9_cache.pkl')
suppress = globals().keys()
with PickleBackend(cache_filename, suppress) as cache, FileChangeInvalidator(cache, os.path.abspath(__file__)):
code_set = CodeCollection(*_construct_icd9_categories().items(), name='elix_icd9')
from clinvoc.code_collections import CodeCollection
import os
from clinvoc.icd9 import ICD9CM
from clinvoc.code_maps import CodeMap
from clinvoc.icd10 import ICD10CM
# This CSV file contains sets of codes in several categories
# See: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4134331/
filename = os.path.join(os.path.dirname(os.path.abspath(__file__)),
'pccccs2.csv')
# Create a CodeCollection based on the CSV file
collection = CodeCollection.from_csv(filename,
vocabs={'ICD-9': ICD9CM()},
header=True,
ignore=[0, 'ICD-10'])
# Create a CodeMap from the collection
code_map = CodeMap.from_code_collection(collection,
standardizers={
'ICD-9': ICD9CM().standardize,
'ICD-10': ICD10CM().standardize
})
# Map ICD9 code V45.2 to see that it's a neurologic or neuromuscular device
print(code_map[('ICD-9', 'V45.2')])