def run(self, record):
if(record.ruid not in self.masterDict):
self.masterDict[record.ruid] = {}
#TODO: the medications part of this regex needs to get until the end of listed medications, not just 1000 characters after
prescriptionsRegex = r'Medications\sKnown\sto\sbe\sPrescribed(.*?)--|Medications:.{0,1000}'
prescriptionsSearch = re.findall(prescriptionsRegex, record.content, re.IGNORECASE)
match = ""
for group in prescriptionsSearch:
match = match + " " + group
if(match != ""):
for drugName in self.drugNames:
drugNameSearch = re.search(drugName, match, re.IGNORECASE)
if(drugNameSearch):
#find the final records with the corresponding ruid
for fr in self.finalRecords:
if fr.ruid == record.ruid:
#add drug and entry date to final record drug dict
if drugName not in fr.drugs:
dr = Drug()
dr.name = drugName
dr.startDate = record.entry_date
dr.endDate = record.entry_date
fr.drugs[drugName] = dr
else:
#get start date for that drug and see if this date is before it
if record.entry_date < fr.drugs[drugName].startDate:
fr.drugs[drugName].startDate = record.entry_date
#get end date for that drug and see if this date is after it
elif record.entry_date > fr.drugs[drugName].endDate:
fr.drugs[drugName].endDate = record.entry_date
def run(self, record):
if (record.ruid not in self.masterDict):
self.masterDict[record.ruid] = {}
#TODO: the medications part of this regex needs to get until the end of listed medications, not just 1000 characters after
prescriptionsRegex = r'Medications\sKnown\sto\sbe\sPrescribed(.*?)--|Medications:.{0,1000}'
prescriptionsSearch = re.findall(prescriptionsRegex, record.content,
re.IGNORECASE)
match = ""
for group in prescriptionsSearch:
match = match + " " + group
if (match != ""):
for drugName in self.drugNames:
drugNameSearch = re.search(drugName, match, re.IGNORECASE)
if (drugNameSearch):
#find the final records with the corresponding ruid
for fr in self.finalRecords:
if fr.ruid == record.ruid:
#add drug and entry date to final record drug dict
if drugName not in fr.drugs:
dr = Drug()
dr.name = drugName
dr.startDate = record.entry_date
dr.endDate = record.entry_date
fr.drugs[drugName] = dr
else:
#get start date for that drug and see if this date is before it
if record.entry_date < fr.drugs[
drugName].startDate:
fr.drugs[
drugName].startDate = record.entry_date
#get end date for that drug and see if this date is after it
elif record.entry_date > fr.drugs[
drugName].endDate:
fr.drugs[
drugName].endDate = record.entry_date
def main():
first_drug = Drug('essentiale')
second_drug = Drug('Karvalol', 'Use before eating', 30, 'transparent', 4,
10)
third_drug = Drug('Anaferon', 'Use after eating', 15,
'some active subtance', 8, 20)
drugs = [first_drug, second_drug, third_drug]
for drug in drugs:
print(drug)
def search_by_str():
search = raw_input("Enter your search term: ")
drugs = start()
drugs = return_match(drugs, search)
if len(drugs) > 0:
drugs = add_FDA_info(drugs)
count = 1
for drug in drugs:
print str(count) + "."
Drug.printer(drugs[drug])
print
count += 1
else:
print "No results."
def worker(self):
'''worker function to store a list of drug, and sort them'''
fileInput = reader(open(sys.argv[1], "r"), delimiter=',')
fileOutput = writer(open(sys.argv[2], "w+"), delimiter=',')
rowItem = self.readInput(fileInput) # skip the header
rowItem = self.readInput(fileInput)
drugList = []
drugDict = {}
drugPrescriberDict = {}
while rowItem != None:
entry_id, last, first, drug_name, drug_cost = rowItem
drug_cost = int(float(drug_cost))
if drug_name not in drugDict:
drug = Drug(drug_name, 1, drug_cost)
drugList.append(drug)
drugDict[drug_name] = len(drugList) - 1
drugPrescriberDict[(drug_name, last,
first)] = len(drugList) - 1
else:
if (drug_name, last, first) not in drugPrescriberDict:
drugindex = drugDict[drug_name]
drug = drugList[drugindex]
drug.num_pres += 1
drug.total_cost += drug_cost
drugPrescriberDict[(drug_name, last, first)] = drugindex
else:
drugindex = drugDict[drug_name]
drug = drugList[drugindex]
drug.total_cost += drug_cost
rowItem = self.readInput(fileInput)
drugList.sort(reverse=True)
self.saveOutput(drugList, fileOutput)
def search_by_num():
print "Enter minimum and maximum current price and percent increase in two years."
print "You can just hit enter if you don't want a minimum or maximum."
raw_min_price = raw_input("MINIMUM CURRENT PRICE: ")
raw_min_percent = raw_input("MINIMUM PERCENT INCREASE: ")
raw_max_price = raw_input("MAXIMUM CURRENT PRICE: ")
raw_max_percent = raw_input("MAXIMUM PERCENT INCREASE: ")
print
if raw_min_percent:
min_percent = float(remove_stuff(raw_min_percent)) / 100
else:
min_percent = -10000000
if raw_min_price:
min_price = float(remove_stuff(raw_min_price))
else:
min_price = 0
if raw_max_percent:
max_percent = float(remove_stuff(raw_max_percent)) / 100
else:
max_percent = 1000000
if raw_max_price:
max_price = float(remove_stuff(raw_max_price))
else:
max_price = 100000000
drugs = start()
drugs = return_highest(drugs, min_percent, min_price, max_percent, max_price)
if len(drugs) > 0:
drugs = add_FDA_info(drugs)
count = 1
print
for drug in drugs:
print str(count) + "."
Drug.printer(drugs[drug])
print
count += 1
ask_for_graph(drugs)
else:
print "No results."
def __init__(self, filepath):
with io.open(filepath, 'r', encoding='utf-8') as f:
self.raw = json.load(f)
self.compounds = [
Compound(x['name'], x['xrefs']) for x in self.raw['compounds']
]
self.remedies = [
Remedy(x['name'], x['xrefs']) for x in self.raw['remedies']
]
self.enzymes = [
Enzyme(x['name'], x['xrefs']) for x in self.raw['enzymes']
]
self.transporter = [
Transporter(x['name'], x['xrefs']) for x in self.raw['transporter']
]
self.drugs = [Drug(x['name'], x['xrefs']) for x in self.raw['drugs']]
publication = self.raw['publication']
doi = publication['doi'] if 'doi' in publication else None
self.publication = Reference(publication['pmid'], doi,
publication['citation'])
def GetDrugData(self):
"""
Gets info on seperate drugs, such as name and associated terms
terms give info on what the drug does
:return:
"""
print 'Getting drug data /(>_<)\\}'
# drop and re-create table drugs
self.remakeTable("drugs")
# get all the important drug ids (dids) from
# the known gene-drug connections table
self.sql.execute('SELECT DISTINCT did FROM drugpairs')
# fetch matching did and use to create uri for query
for result in tqdm(self.sql.fetchall()):
did = result[0]
d = Drug(did)
for item in d.terms:
term = item['term']
# check for duplicates with chemical name
if d.name not in term.lower():
item = (did, d.name, term)
# insert into drugs table
self.insertValues("drugs", item)
self.conn.commit()
def main():
base_url = "https://www.drugs.com"
alpha_url = "/alpha"
alphabet = "abcdefghijklmnopqrstuvwxyz"
#alphabet = "a"
raw_all_drugs = ""
for letter in alphabet:
print(letter)
url = base_url + alpha_url + "/" + letter + "1.html"
# print(url)
page = BeautifulSoup(requests.get(url).content, "lxml")
letter_drugs = page.find_all("div", ["boxListPopular"])
raw_all_drugs += str(letter_drugs[0])
all_drugs = {}
drug_parsed = BeautifulSoup(raw_all_drugs, "lxml")
for drug in drug_parsed.find_all("a", href=True):
all_drugs[drug.string] = Drug(drug.string, base_url + drug["href"])
with open('all_drugs.json', 'w') as f:
for name, drug in all_drugs.items():
json.dump({name: drug.overview_questions}, f)
f.write('\n')
def new_fda(drug_dict):
# this function supplies drug and vendor names via NDCs from the FDA's official database.
# sadly the FDA and NADAC have a slightly different format for NDCs, so it's not perfect.
# It is very unlikely to assign an incorrect NDC -- rather, it will pass on drugs it cannot find a match for.
name_file = "FDA-20160118.csv"
headers = ["PRODUCTID", "PRODUCTNDC", "PRODUCTTYPENAME", "PROPRIETARYNAME", "PROPRIETARYNAMESUFFIX",
"NONPROPRIETARYNAME", "DOSAGEFORMNAME", "ROUTENAME", "STARTMARKETINGDATE", "ENDMARKETINGDATE",
"MARKETINGCATEGORYNAME", "APPLICATIONNUMBER", "LABELERNAME", "SUBSTANCENAME",
"ACTIVE_NUMERATOR_STRENGTH", "ACTIVE_INGRED_UNIT", "PHARM_CLASSES", "DEASCHEDULE"]
count = 1
results = {}
current_list = [key for key in drug_dict]
with open(name_file, "r") as namecsv:
csvreader = csv.reader(namecsv)
for line in csvreader:
if count == 1:
pass
else:
drug = {}
for i in range(len(headers)):
if headers [i] == "PRODUCTNDC":
drug["NDC"] = line[i].translate(None,"-")
templist = line[i].split("-")
templist[0] = templist[0] + "0"
drug["NDC2"] = "".join(templist)
else:
drug[headers[i]] = line[i]
# print "Looking for NDC " + drug["NDC"] + "..."
findings = [item for item in current_list if ( drug["NDC"] in item ) or ( drug["NDC2"] in item )]
if len(findings) == 1:
this_drug = drug_dict[findings[0]]
Drug.add_vendor(this_drug, drug["LABELERNAME"])
Drug.add_sci_name(this_drug, drug["NONPROPRIETARYNAME"])
Drug.add_desc(this_drug, drug["PHARM_CLASSES"])
current_list.remove(findings[0])
print "%i drugs left to ID..." % len(current_list)
elif len(findings) == 0:
pass
else:
findings_as_drugs = [drug_dict[key] for key in findings] #converts list of keys into list of Drugs
findings = multiple_results(drug, findings_as_drugs) #list of Drugs
for finding in findings: #finding is a Drug
Drug.add_vendor(finding, drug["LABELERNAME"])
Drug.add_sci_name(finding, drug["NONPROPRIETARYNAME"])
Drug.add_desc(finding, drug["PHARM_CLASSES"])
current_list.remove(finding.id)
print "%i drugs left to ID..." % len(current_list)
# BECAUSE THE FDA NDCS ARE ONLY EIGHT DIGITS LONG, THIS SEARCH OFTEN RETURNS 2+ RESULTS.
# TODO: write a function to check other info to figure out which is the correct one.
# or, worst-case scenario, prompt user to choose which looks right, maybe?
count += 1
return drug_dict
def write(drugs):
fieldnames = ["ID", "Name", "Scientific Name", "Unit", "OTC", "Brand/Generic", "Vendor", "Package", "Description"]
dates = get_date_list()
for i in range(len(dates)):
fieldnames.append(dates[i])
with open("FullPrices.csv","a") as prices_csv:
writer = csv.writer(prices_csv, delimiter='|')
writer.writerow(fieldnames)
for entry in drugs:
drug = drugs[entry]
drugrow = [drug.id, drug.name, drug.scientific_name, drug.unit, drug.otc, drug.b_or_g, drug.vendor, drug.package, drug.desc]
for i in range(len(dates)):
try:
drugrow.append(drug.prices[dates[i]])
except KeyError:
drugrow.append(None)
writer.writerow(drugrow)
prices_csv.close()
print "Done"
if __name__ == '__main__':
drugs = create_list()
count = 0
for drug in drugs:
print "%i." % count
Drug.printer(drugs[drug])
print
count += 1
write(drugs)
def main(argv):
i = int(argv[1])
u = Utils()
db = Database('Mimir from Munnin')
np.random.seed(u.RANDOM_STATE)
df_patients = u.load_df('df_patients')
sex_adr = db.get_list('select meddra_pt_id from gender_terms')
drugs = db.get_list('select atc_5_id from atc_5_name')
test = [(x, y) for x, y in u.load_np('prr_test')]
df_test = pd.DataFrame(test, columns=['drug',
'adr']).sort_values(by='drug')[i:]
PID_M = set(df_patients.query('Sex=="M"').get('PID').values)
PID_F = set(df_patients.query('Sex=="F"').get('PID').values)
for drugID, data in df_test.groupby('drug'):
filename = 'Post_PRR/' + str(drugID)
pth = u.DATA_PATH + filename + '.feather'
if os.path.exists(pth):
print(drugID)
continue
prr_counts = []
q = 'select PID from atc_5_patient where atc_5_id = ' + str(drugID)
takes_drug = set(db.get_list(q))
try:
drug = Drug(drugID)
drug.match()
except NameError:
df = pd.DataFrame(
columns=['drug', 'adr', 'sex', 'a_post', 'c_post'])
u.save_df(df, filename)
if drug.match_m is None or drug.match_f is None:
df = pd.DataFrame(
columns=['drug', 'adr', 'sex', 'a_post', 'c_post'])
u.save_df(df, filename)
db = Database('Mimir from Munnin')
continue
for adrID in data.adr.values:
q = 'select PID from pt_patient where meddra_concept_id = ' + str(
adrID)
has_adr = set(db.get_list(q))
for sex in ['M', 'F']:
if sex == 'M': PSM = drug.match_m
else: PSM = drug.match_f
a_post = len([1 for x in PSM if x in (has_adr & takes_drug)])
c_post = len([1 for x in PSM if x in (takes_drug - has_adr)])
res = {
'drug': drugID,
'adr': adrID,
'sex': sex,
'a_post': a_post,
'c_post': c_post
}
prr_counts.append(res)
df = pd.DataFrame(prr_counts)
u.save_df(df, filename)
break
def main(argv):
u = Utils()
iterations = 25
idx = int(argv[1])
drugs = u.load_np('drugs')
drugID = drugs[idx]
status = u.read_status(drugID)
if status == 'no':
try:
u.write_status(drugID, 'working')
drug = Drug(drugID)
for itr in range(1, iterations + 1):
drug.match()
drug.count_adr()
drug.assign_abcd(itr)
drug.do_chi_square()
drug.calc_logROR()
drug.reset_for_next_itr()
x = drug.save_results(iterations)
if x:
u.write_status(drugID, 'yes')
else:
u.write_status(drugID, 'no')
except:
info = str(sys.exc_info()[1])
u.write_status(drugID, 'error ' + info)
def get_drugs(file_path):
with open(file_path, 'r') as drugs_file:
reader = csv.reader(drugs_file, delimiter=",")
header = next(reader)
return [(Drug(line[0], line[1])) for line in reader if len(line) == 2]
index = self.mapItemtoHeapIndex[item]
#print index
item.total_cost += amount
self.sift_up(index)
def __str__(self):
''' print the heap out'''
res = '['
for item in self.hq:
res += item.__str__() + ', '
res += ']'
return res
if __name__ == "__main__":
drug1 = Drug("drug1", 2, 1000)
drug2 = Drug("drug2", 1, 1500)
drug3 = Drug("drug3", 4, 500)
drug4 = Drug("drug4", 3, 1500)
drug5 = Drug("drug5", 1, 1500)
drug6 = Drug("drug06", 1, 1500)
maxheap = MaxHeap()
#print maxheap
maxheap.heappush(drug1)
#print maxheap
maxheap.heappush(drug2)
#print maxheap
maxheap.heappush(drug3)
print maxheap
maxheap.heappush(drug4)
maxheap.heappush(drug5)
