In [1]:
import cPickle as pickle,pandas as pd,nltk,codecs,json
In [2]:
cond_temp = [row.strip().split('|')[2] for row in codecs.open('../data/condition_browse.txt','r','utf-8').readlines()]
corrections = {"Sarcoma, Ewing's": 'Sarcoma, Ewing',
'Beta-Thalassemia': 'beta-Thalassemia',
'Von Willebrand Disease, Type 3': 'von Willebrand Disease, Type 3',
'Von Willebrand Disease, Type 2': 'von Willebrand Disease, Type 2',
'Von Willebrand Disease, Type 1': 'von Willebrand Disease, Type 1',
'Felty''s Syndrome': 'Felty Syndrome',
'Von Hippel-Lindau Disease': 'von Hippel-Lindau Disease',
'Retrognathism': 'Retrognathia',
'Regurgitation, Gastric': 'Laryngopharyngeal Reflux',
'Persistent Hyperinsulinemia Hypoglycemia of Infancy': 'Congenital Hyperinsulinism',
'Von Willebrand Diseases': 'von Willebrand Diseases',
'Pontine Glioma': 'Brain Stem Neoplasms',
'Mental Retardation': 'Intellectual Disability',
'Overdose': 'Drug Overdose',
'Beta-Mannosidosis': 'beta-Mannosidosis',
'Alpha 1-Antitrypsin Deficiency': 'alpha 1-Antitrypsin Deficiency',
'Intervertebral Disk Displacement': 'Intervertebral Disc Displacement',
'Alpha-Thalassemia': 'alpha-Thalassemia',
'Mycobacterium Infections, Atypical': 'Mycobacterium Infections, Nontuberculous',
'Legg-Perthes Disease': 'Legg-Calve-Perthes Disease',
'Intervertebral Disk Degeneration': 'Intervertebral Disc Degeneration',
'Alpha-Mannosidosis': 'alpha-Mannosidosis',
'Gestational Trophoblastic Disease': 'Gestational Trophoblastic Neoplasms'
}
conditions = nltk.FreqDist([c if c not in corrections else corrections[c] for c in cond_temp])
In [3]:
syn = pickle.load(open('../data/mesh_synonyms.pickle','rb'))
In [4]:
cond_dict = {}
def add_to_dict(word, subtotal):
if word not in cond_dict: cond_dict[word] = 0
cond_dict[word] += subtotal
for c, cnt in conditions.items():
add_to_dict(c, cnt)
if c in syn:
for syn_c in syn[c]:
add_to_dict(syn_c, cnt)
In [5]:
final = [c.encode('ascii') for c, cnt in sorted(cond_dict.items(), key=lambda x: (x[1],-1*len(x[0])), reverse=True)]
In [6]:
final
Out[6]:
['Carcinoma',
'DM',
'Diabetes',
'Diabetes Mellitus',
'Diabetes in Children and Teens',
'Tumor',
'Cancer',
'Neoplasms',
'Malignancy',
'Cancer in Children',
'Cancer--Living with Cancer',
'Breast Neoplasms',
'Diabetes Type 2',
'Type 2 Diabetes',
'Diabetes Mellitus, Type 2',
'HIV',
'AIDS',
'HIV/AIDS',
'HIV Infections',
'HIV/AIDS in Women',
'Living with HIV/AIDS',
'Human immunodeficiency virus',
'Acquired immunodeficiency syndrome',
'Acquired Immunodeficiency Syndrome',
'Leukemia',
'Childhood Leukemia',
'Hypertension',
'Lung Cancer',
'Lung Neoplasms',
'Bronchogenic carcinoma',
'Lymphoma',
'Non-Hodgkin lymphoma',
'Obesity',
'Depression',
'Dysthymic disorder',
'Unipolar depression',
'Clinical depression',
'Major depressive disorder',
'Depressive Disorder',
'CAD',
'Coronary Artery Disease',
'Coronary atherosclerosis',
'Coronary arteriosclerosis',
'Prostate Cancer',
'Prostatic Neoplasms',
'Myocardial Ischemia',
'Carcinoma, Non-Small-Cell Lung',
'Coronary Disease',
'Traumatic injuries',
'Wounds and Injuries',
'Asthma',
'Hepatitis',
'Viral hepatitis',
'HAV',
'Hepatitis A',
'Colon cancer',
'Rectal cancer',
'Colorectal Cancer',
'Colorectal Neoplasms',
'Cardiovascular Diseases',
'Lung Diseases',
'Pulmonary Rehabilitation',
'CHF',
'Heart Failure',
'Cardiac failure',
'Congestive heart failure',
'Left-sided heart failure',
'Right-sided heart failure',
'Arthritis',
'Heart Diseases',
'Cardiac diseases',
'Cardiac Rehabilitation',
'Heart Disease in Women',
'Heart Diseases--Prevention',
'Renal disease',
'Kidney Diseases',
'CVA',
'Stroke',
'Brain attack',
'Stroke Rehabilitation',
'COPD',
'Pulmonary Disease, Chronic Obstructive',
'Schizophrenia',
'Leukemia, Myeloid',
'Lymphoma, Non-Hodgkin',
'OA',
'Osteoarthritis',
'Osteoarthrosis',
'Degenerative joint disease',
'Multiple Myeloma',
'Plasma-cell myeloma',
'ESRD',
'Renal failure',
'Kidney Failure',
'Neoplasms, Plasma Cell',
'Kidney Failure, Chronic',
'End-stage renal disease',
'Respiration Disorders',
'Lung Diseases, Obstructive',
'CLL',
'ALL',
'Leukemia, Lymphoid',
'Acute Lymphocytic Leukemia',
'Acute lymphoblastic leukemia',
'Chronic Lymphocytic Leukemia',
'RA',
'Rheumatoid Arthritis',
'Arthritis, Rheumatoid',
'Fibrosis',
'Cerebral Infarction',
'Neoplasm Metastasis',
'Sclerosis',
'Ovarian Cancer',
'Ovarian Neoplasms',
'HCV',
'Hepatitis C',
'Melanoma',
'Flu',
'Grippe',
'Influenza',
'Influenza, Human',
'AML',
'ANLL',
'Acute Myeloid Leukemia',
'Leukemia, Myeloid, Acute',
'Acute myelogenous leukemia',
'Glioma',
'Diabetes Type 1',
'Juvenile diabetes',
'Diabetes Mellitus, Type 1',
'Insulin-dependent diabetes',
'Anemia',
'Iron poor blood',
'Head and Neck Cancer',
'Head and Neck Neoplasms',
'MDS',
'Myelodysplastic Syndromes',
'Depressive Disorder, Major',
'Preleukemia',
'Hematoma',
'Adenocarcinoma',
'Ulcer',
'Pancreatic Cancer',
'Pancreatic Neoplasms',
'Infarction',
'MI',
'Heart Attack',
'Myocardial Infarction',
'Bleeding',
'Hemorrhage',
'Renal Insufficiency',
'Ischemia',
'Shaking palsy',
'Parkinson Disease',
'Paralysis agitans',
"Parkinson's Disease",
'Alzheimer Disease',
"Alzheimer's Caregivers",
'Inflammation',
'MS',
'Multiple Sclerosis',
'Smoking',
'Teen smoking',
'Pipe smoking',
'Cigar smoking',
'Tobacco smoking',
'Cigarette smoking',
'Smoking and Youth',
'CKD',
'Chronic Kidney Disease',
'Renal Insufficiency, Chronic',
'Precursor Cell Lymphoblastic Leukemia-Lymphoma',
'Hepatitis, Chronic',
'AF',
'Atrial Fibrillation',
'Arteriosclerosis',
'Anxiety',
'Anxiety Disorders',
'Carcinoma, Renal Cell',
'Carcinoma, Squamous Cell',
'Colitis',
'Glaucoma',
'Sleep Apnea',
'Sleep Apnea Syndromes',
'Sleep-disordered breathing',
'Leukemia, Lymphocytic, Chronic, B-Cell',
'Sarcoma',
'Soft Tissue Sarcoma',
'Stomach Cancer',
'Gastric cancer',
'Stomach Neoplasms',
'Osteoporosis',
'Bipolar Disorder',
'Manic-depressive illness',
'Carcinoma, Hepatocellular',
'Insulin Resistance',
'Psychoses',
'Psychotic Disorders',
'Chronic Disease',
'Coping with Chronic Illness',
'Meningioma',
'Apnea',
'Hepatitis C, Chronic',
'Fractures, Bone',
'HBV',
'Hepatitis B',
'Pain, Postoperative',
'Liver Cancer',
'Liver Neoplasms',
'Hepatocellular carcinoma',
'Osteoarthritis, Knee',
'Metabolic Syndrome',
'Metabolic Syndrome X',
'Insulin resistance syndrome',
'Bronchopneumonia',
'Pneumonia',
'Overweight',
'Atherosclerosis',
'Lumbago',
'Backache',
'Back Pain',
'PTSD',
'Post-Traumatic Stress Disorder',
'Stress Disorders, Post-Traumatic',
'Brain Cancer',
'Brain Neoplasms',
'Childhood Brain Tumors',
'Psoriasis',
'Glioblastoma',
'Lymphoma, B-Cell',
'Hypercholesterolemia',
'Birth Defects',
'Congenital Abnormalities',
'Sterility',
'Infertility',
'Malaria',
'Drug Abuse',
'Drunk driving',
'Substance abuse',
'Drugged driving',
'Impaired Driving',
'Distracted driving',
'Drugs and Young People',
'Prescription Drug Abuse',
'Substance-Related Disorders',
'Rhinitis',
'Mental illness',
'Mental Disorders',
'Epilepsy',
'Stress Disorders, Traumatic',
'ADHD',
'Attention Deficit Hyperactivity Disorder',
'Attention Deficit Disorder with Hyperactivity',
'Sleep Apnea, Obstructive',
'Brain Injuries',
'Immunologic Deficiency Syndromes',
'Cervical Cancer',
'Uterine Cervical Neoplasms',
'Colonic Neoplasms',
'Low Back Pain',
'CML',
'Chronic Myeloid Leukemia',
'Chronic myelogenous leukemia',
'Chronic granulocytic leukemia',
'Leukemia, Myelogenous, Chronic, BCR-ABL Positive',
'Neoplasms, Second Primary',
'Hodgkin Disease',
'Hodgkin lymphoma',
'CF',
'Liver Diseases',
'Cystic Fibrosis',
'Hepatic disease',
'Esophageal Cancer',
'Esophageal Neoplasms',
'AMD',
'Rectal Neoplasms',
'Macular Degeneration',
'Pulmonary Hypertension',
'Hypertension, Pulmonary',
'Pulmonary arterial hypertension',
'Age-related macular degeneration',
'Dermatitis',
'Cognition Disorders',
'Sepsis',
'Central Nervous System Neoplasms',
'Hyperkinesis',
'Crohn Disease',
"Crohn's Disease",
'Regional ileitis',
'Regional enteritis',
'Tobacco Use Disorder',
'Nervous System Neoplasms',
'TB',
'Tuberculosis',
'Cirrhosis',
'Liver Cirrhosis',
'Hepatic fibrosis',
'Acute Coronary Syndrome',
'Weight Loss',
'Weight Control',
'PAD',
'Peripheral Arterial Disease',
'Renal cancer',
'Kidney Cancer',
'Hypernephroma',
'Kidney Neoplasms',
'Thrombosis',
'Alcoholism',
'Alcohol dependence',
'Alcoholism and Alcohol Abuse',
'GERD',
'Gastroesophageal Reflux',
'Gastroesophageal reflux disease',
'Senility',
'Dementia',
'DVT',
'Venous Thrombosis',
'Deep Vein Thrombosis',
'Emergencies',
'Winter Weather Emergencies',
'Edema',
'Dropsy',
'Respiratory Distress Syndrome, Adult',
'Migraine',
'Migraine Disorders',
'Hematologic Neoplasms',
'Bladder Cancer',
'Urinary Bladder Neoplasms',
'Lymphoma, Large B-Cell, Diffuse',
'Cataract',
'Fatigue',
'Tiredness',
'Weariness',
'Spinal Cord Injuries',
'Peritoneal Neoplasms',
'Constriction, Pathologic',
'Critical Illness',
'Lymphoma, Large-Cell, Immunoblastic',
'Glaucoma, Open-Angle',
'Neurotoxicity Syndromes',
'Allergy',
'Hypersensitivity',
'Aneurysm',
'Aneurysms',
'Pancolitis',
'Distal colitis',
'Ulcerative Colitis',
'Colitis, Ulcerative',
'Respiratory Failure',
'Ulcerative proctitis',
'Respiratory Insufficiency',
'Peripheral Vascular Diseases',
'Virus Diseases',
'Viral Infections',
'Hernia',
'Enterocele',
'Dyslipidemias',
'Lymphoma, Follicular',
'Dysentery',
'Retinal Diseases',
'Retinal Disorders',
'Neuralgia',
'Vascular Diseases',
'Emesis',
'Vomiting',
'Nausea and Vomiting',
'ASD',
'Autistic Disorder',
'Autism Spectrum Disorder',
'Pervasive developmental disorder (PDD)',
'Macular Edema',
'Graft vs Host Disease',
'Fallopian Tube Neoplasms',
'Neoplasms, Glandular and Epithelial',
'Myopia',
'Myocardiopathy',
'Cardiomyopathy',
'Cardiomyopathies',
'Venous Thromboembolism',
'Dilated cardiomyopathy',
'Restrictive cardiomyopathy',
'Hypertrophic cardiomyopathy',
'The runs',
'Diarrhea',
'The trots',
'Lymphoma, Mantle-Cell',
'Insomnia',
'Sleep Initiation and Maintenance Disorders',
'Constipation',
'Small Cell Lung Carcinoma',
'Endometrial Neoplasms',
'Myeloproliferative Disorders',
'Plasmacytoma',
'Ocular Hypertension',
'Myofascial Pain Syndromes',
'FMS',
'Fibromyalgia',
'Recurrence',
'Shock',
'Thromboembolism',
'Lymphoma, T-Cell',
'Neuroblastoma',
'Urinary Incontinence',
'Hyperlipidemias',
'Hepatitis B, Chronic',
'Dermatitis, Atopic',
'Hyperplasia',
'Irritable Bowel Syndrome',
'C',
'Snow',
'Coke',
'Blow',
'Coca',
'Crack',
'Flake',
'Cocaine',
'Cocaine-Related Disorders',
'Respiratory Distress Syndrome, Newborn',
'Headache',
'Lupus Erythematosus, Systemic',
'Nausea',
'Sickle Cell Anemia',
'Anemia, Sickle Cell',
'Hemoglobin SS disease',
'Mycoses',
'Fungal Infections',
'Intestinal Diseases',
'Opioid-Related Disorders',
'Glucose Intolerance',
'Respiratory Aspiration',
'Fever',
'Pyrexia',
'Toxemia',
'Polyneuropathies',
'Overactive Bladder',
'Urinary Bladder, Overactive',
'Respiratory Tract Infections',
'Seizures',
'Arrhythmia',
'Irregular heartbeat',
'Arrhythmias, Cardiac',
'PCOS',
'Stein-Leventhal Syndrome',
'Polycystic Ovary Syndrome',
'Delirium',
'Malnutrition',
'Acute Kidney Injury',
'Thyroid Diseases',
'Fatty Liver',
'Dry Eye Syndromes',
'Keratoconjunctivitis Sicca',
'Hemophilia',
'Hemophilia A',
'Christmas disease',
'Factor IX deficiency',
'Factor VIII deficiency',
'ED',
'Impotence',
'Erectile Dysfunction',
'Multiple Sclerosis, Relapsing-Remitting',
'Embolism',
'Hay Fever',
'Pollen allergy',
'Rhinitis, Allergic, Seasonal',
'Acute Lung Injury',
'Zits',
'Acne',
'Atrophy',
'Pimples',
'Acne Vulgaris',
'Diabetic Retinopathy',
'Diabetic Eye Problems',
'Sleep Disorders',
'Parasomnias',
'Quadriplegia',
'Prostatic Hyperplasia',
'Enlarged Prostate (BPH)',
'benign prostatic hyperplasia',
'Hemiplegia',
'TAA',
'AAA',
'Aortic Aneurysm',
'Pinkeye',
'Conjunctivitis',
'Thyroid Cancer',
'Thyroid Neoplasms',
'Carcinoma in Situ',
'Neurologic Manifestations',
'Tachycardia',
'Inflammatory Bowel Diseases',
'Diabetic neuropathy',
'Diabetic Neuropathies',
'Diabetic Nerve Problems',
'Vitamin D Deficiency',
'Nerve Compression Syndromes',
'Bulimia',
'Cholangiocarcinoma',
'Congenital Heart Defects',
'Heart Defects, Congenital',
'Angina',
'Angina Pectoris',
'Demyelinating Diseases',
'Paraplegia',
'Motor Neuron Disease',
'Premature Birth',
'SCA',
'Heart Arrest',
'Cardiac Arrest',
'Obesity, Morbid',
'Sudden cardiac death',
'Child Development Disorders, Pervasive',
'Hyperglycemia',
'High blood sugar',
'High blood glucose',
'Bacterial Infections',
'Peripheral Nervous System Diseases',
'Hypertrophy',
'Encephalitis',
'CP',
'Cerebral Palsy',
'Burns',
'Astrocytoma',
'Leukemia, Myelomonocytic, Chronic',
'ALS',
"Lou Gehrig's disease",
'Amyotrophic Lateral Sclerosis',
'Thrombocytopenia',
'Metabolic Diseases',
'Metabolic Disorders',
'Diabetic Foot',
'Prediabetes',
'Prediabetic State',
'Skin Cancer',
'Skin Neoplasms',
'Pulmonary Embolism',
'Blood clots in the lung',
'Leukemia, Myelomonocytic, Acute',
'Lymphoma, B-Cell, Marginal Zone',
'UTI',
'Shock, Septic',
'Urinary Tract Infections',
'Acute Disease',
'Stress',
'Psychological stress',
'Stress, Psychological',
'Palsy',
'Pertussis',
'Paralysis',
'Whooping Cough',
'Myelodysplastic-Myeloproliferative Diseases',
'Gallstones',
'Waldenstrom Macroglobulinemia',
'Alcohol',
'Drinking',
'Pancreatitis',
'Teenage drinking',
'Alcohol Drinking',
'Underage Drinking',
'Nasopharyngeal Neoplasms',
'Presbycusis',
'Anemia, Iron-Deficiency',
'Neoplasms, Germ Cell and Embryonal',
'Gastrointestinal Neoplasms',
'LBP',
'Hypotension',
'Low Blood Pressure',
'Neurologic Diseases',
'Nervous System Diseases',
'Malaria, Falciparum',
'Burkitt Lymphoma',
'Lymphoma, T-Cell, Cutaneous',
'Anoxia',
'Body Weight',
'Hearing Loss',
'Mood Disorders',
'Deafness',
'Hearing loss',
'Mesothelioma',
'Psychomotor Agitation',
'Malignant mesothelioma',
'Hearing Disorders and Deafness',
'Meningitis',
'Birth Weight',
'Spinal meningitis',
'Tetanus',
'Prolapse',
'Diphtheria',
'Scars',
'Cicatrix',
'Keloid scar',
'Endocrine Diseases',
'Endocrine System Diseases',
'Foot Ulcer',
'Pre-Eclampsia',
'Precancerous Conditions',
'OCD',
'Sinusitis',
'Aortic Valve Stenosis',
'Obsessive-Compulsive Disorder',
'Diabetic nephropathy',
'Diabetic Nephropathies',
'Diabetic Kidney Problems',
'Compulsive Personality Disorder',
'Hypoglycemia',
'Low blood sugar',
'Spondylitis',
'Carcinoma, Transitional Cell',
'Aortic Aneurysm, Abdominal',
'IPF',
'Bleeding Disorders',
'Clotting disorders',
'Pulmonary Fibrosis',
'Bone Marrow Diseases',
'Hemostatic Disorders',
'Gastrointestinal Diseases',
'Blood Coagulation Disorders',
'Idiopathic pulmonary fibrosis',
'Hypogonadism',
'MRSA',
'Staph',
'Joint Diseases',
'Joint Disorders',
'Staphylococcal Infections',
'Methicillin-resistant Staphylococcus aureus',
'Sexual Dysfunctions, Psychological',
'Neutropenia',
'Hyperparathyroidism',
'Hip Fractures',
'Scleroderma, Diffuse',
'Scleroderma, Systemic',
'Blood Disorders',
'Hematologic Diseases',
'Hypothermia',
'Cold-related illness',
'Rheumatoid spondylitis',
'Ankylosing Spondylitis',
'Spondylitis, Ankylosing',
'Neovascularization, Pathologic',
'Leiomyoma',
'Aplastic Anemia',
'Anemia, Aplastic',
'Herpes Zoster',
'Diabetes, Gestational',
'Osteoporosis, Postmenopausal',
'Myofibroma',
'Behavior, Addictive',
'Primary Myelofibrosis',
'Carcinoma, Basal Cell',
'Rupture',
'After Surgery',
'Postoperative care',
'Recovery from surgery',
'Postoperative Complications',
'Emphysema',
'Endo',
'Bone Cancer',
'Endometriosis',
'Bone Neoplasms',
'Osteosarcoma',
'Pulmonary Emphysema',
'MD',
'Muscular Dystrophy',
'Muscular Dystrophies',
'Oropharyngeal Neoplasms',
'Conjunctivitis, Allergic',
'Uveitis',
'Lung Injury',
'Marijuana Abuse',
'Dwarfism, Pituitary',
'Neuroendocrine Tumors',
'Intervertebral Disc Degeneration',
'Mucositis',
'Thalassemia',
"Cooley's anemia",
'Mediterranean anemia',
'Sarcoma, Endometrial Stromal',
'Chorioretinitis',
'Lymphoproliferative Disorders',
'Urinary Incontinence, Stress',
'Muscle Spasticity',
'Lymphoma, Large-Cell, Anaplastic',
'Purpura',
'Cavities',
'Tooth Decay',
'Hot Flashes',
'Dental Caries',
'Gastrointestinal Stromal Tumors',
'Gliosarcoma',
'Keratosis',
'Bone Diseases, Metabolic',
'Mycosis Fungoides',
'Hernia, Inguinal',
'Necrosis',
'Binge eating',
'Anorexia nervosa',
'Eating Disorders',
'Purpura, Thrombocytopenic',
'Choroidal Neovascularization',
'Death',
'Terminal care',
'Death and dying',
'End of Life Issues',
'Keratosis, Actinic',
'Phobias',
'Phobic Disorders',
'Osteoarthritis, Hip',
'CMV',
'Stomach flu',
'Esophagitis',
'Periodontitis',
'Gastroenteritis',
'Digestive Diseases',
'Norovirus Infections',
'Norwalk virus infections',
'Digestive System Diseases',
'Gastrointestinal diseases',
'Cytomegalovirus Infections',
'Otitis',
'Arterial Occlusive Diseases',
'Postoperative Nausea and Vomiting',
'Bronchiolitis',
'Neck Pain',
'Weight Gain',
'Epilepsies, Partial',
'Meningococcal Infections',
'Fistula',
'Fistulas',
'Moniliasis',
'Candidiasis',
'Yeast Infections',
'Subarachnoid Hemorrhage',
'Adenoma',
'Dysphagia',
'Eye Diseases',
'Laryngeal Neoplasms',
'Swallowing Disorders',
'Deglutition Disorders',
'Lymphoma, T-Cell, Peripheral',
'PPS',
'Infection',
'Poliomyelitis',
'HIV Seropositivity',
'Infantile paralysis',
'Death, Sudden, Cardiac',
'Polio and Post-Polio Syndrome',
'Bronchitis',
'Skin Diseases',
'Skin Conditions',
'Acute Bronchitis',
'Cutaneous disorders',
'Hypertriglyceridemia',
'Retinal Vein Occlusion',
'Dermatologic disorders',
'Chest Pain',
'Wet Macular Degeneration',
'Dystonia',
'Rhinitis, Allergic, Perennial',
'STDs',
'Wound Infection',
'Venereal disease',
'Tachycardia, Ventricular',
'Sexually Transmitted Diseases',
'Sexually transmitted infections',
'Infectious Diseases',
'Somatoform Disorders',
'Communicable Diseases',
'Dystonic Disorders',
'Hepatic Encephalopathy',
'Anorexia',
'Cervical Intraepithelial Neoplasia',
'Alopecia',
'Bacteremia',
'Oligodendroglioma',
'Dyspepsia',
'Indigestion',
'Upset stomach',
'Idiopathic Pulmonary Fibrosis',
'Retinitis',
'Sezary Syndrome',
'Psoriatic Arthritis',
'Arthritis, Psoriatic',
'Pelvic Organ Prolapse',
'Paresis',
'Hyperopia',
'Barrett Esophagus',
'Ventricular Dysfunction',
'Leukemia, Myeloid, Chronic-Phase',
'RLS',
'Restless Legs',
'Osteonecrosis',
'Aseptic necrosis',
'Ischemic necrosis',
'Avascular necrosis',
'Reperfusion Injury',
'Deficiency Diseases',
'Restless Legs Syndrome',
'Lymphoma, Extranodal NK-T-Cell',
'Polyps',
'Dyspnea',
'Angina, Unstable',
'Brain Damage, Chronic',
'Otitis Media',
'Ear Infections',
'Pneumonia, Ventilator-Associated',
'Suicide',
'Purpura, Thrombocytopenic, Idiopathic',
'Tinnitus',
'Aggression',
'Leukemia, T-Cell',
'Disease Susceptibility',
'Arthritis, Juvenile Rheumatoid',
'Sarcoma, Ewing',
'Leukemia-Lymphoma, Adult T-Cell',
'HPV',
'Stomatitis',
'Human papillomavirus',
'Philadelphia Chromosome',
'Papillomavirus Infections',
'Hyperparathyroidism, Secondary',
'Warts',
'Verruca',
'Carotid Artery Disease',
'Carotid Artery Diseases',
'GI bleeding',
'Breast Cancer',
'Panic Disorder',
'Carcinoma, Ductal, Breast',
'Gastrointestinal Bleeding',
'Gastrointestinal Hemorrhage',
'Muscular Dystrophy, Duchenne',
'Cholestasis',
'Cerebral Hemorrhage',
'Disease Progression',
'Neoplasms, Squamous Cell',
'Amyloidosis',
'Alopecia Areata',
'Neuroectodermal Tumors, Primitive',
'Male Breast Cancer',
'Breast Neoplasms, Male',
'Calculi',
'Spinal Diseases',
'Bile Duct Cancer',
'Bile Duct Neoplasms',
'TIA',
'Cystitis',
'Shingles',
'Acromegaly',
'Mini-stroke',
'Blast Crisis',
'Herpes zoster',
'Uterine Cancer',
'Uterine Neoplasms',
'Endometrial cancer',
'Postherpetic neuralgia',
'Neuralgia, Postherpetic',
'Mycobacterium Infections',
'Transient Ischemic Attack',
'Ischemic Attack, Transient',
'Kala-azar',
'Leishmaniasis',
'Brain Ischemia',
'Personality Disorders',
'Dyssomnias',
'Proteinuria',
'Iron Overload',
'Coronary Stenosis',
'Heart Valve Diseases',
'Valvular heart disease',
'Lymphomatoid Granulomatosis',
'Pleural Effusion',
'POF',
'Carotid Stenosis',
'Rhabdomyosarcoma',
'Hemorrhagic Disorders',
'Ovarian insufficiency',
'Premature Ovarian Failure',
'Primary Ovarian Insufficiency',
'KS',
'Bellyache',
'Cholelithiasis',
'Abdominal Pain',
'Sarcoma, Kaposi',
"Kaposi's Sarcoma",
'Muscular Atrophy',
'Cholecystolithiasis',
'Liver Failure',
'Hypopharyngeal Neoplasms',
'Astigmatism',
'Hemophilia B',
'Autoimmune Diseases',
'Tetany',
'Lymphedema',
'Pelvic Pain',
'Varicose Ulcer',
'Spinal Stenosis',
'Lymphatic obstruction',
'Immunoblastic Lymphadenopathy',
'Dyskinesias',
'Polycythemia',
'Signs and Symptoms',
'Intestinal Neoplasms',
'HSV',
'Myopathy',
'Sarcopenia',
'Herpes Simplex',
'Anorexia Nervosa',
'Muscle Disorders',
'Muscular Diseases',
'Gallbladder Cancer',
'Gallbladder Neoplasms',
'Helicobacter Infections',
'Tuberculosis, Pulmonary',
'Intermittent Claudication',
'Helicobacter Pylori Infections',
'Cough',
'Oral Cancer',
'Colon polyps',
'Colonic Polyps',
'Mouth Neoplasms',
'HD',
'Abscess',
'Huntington Disease',
'Movement Disorders',
"Huntington's chorea",
"Huntington's Disease",
'Bronchopulmonary Dysplasia',
'Thrombocythemia, Essential',
'Nevus',
'Leg Ulcer',
'Trisomy 21',
'Appendicitis',
'Down Syndrome',
'Muscle Weakness',
'Carcinoid Tumor',
'Medulloblastoma',
'Carcinoid Tumors',
'Neuroectodermal Tumors',
'Cerebrovascular Disorders',
'Ventricular Dysfunction, Left',
'Lipodystrophy',
'Femoral Fractures',
'Duodenal Neoplasms',
'IC',
'PBS',
'Bed sores',
'Varicella',
'Chickenpox',
'Bradycardia',
'Shoulder Pain',
'Pressure Sores',
'Pressure Ulcer',
'Pressure ulcers',
'Decubitus ulcers',
'Laryngeal Diseases',
'Interstitial Cystitis',
'Bladder pain syndrome',
'Cystitis, Interstitial',
'Painful bladder syndrome',
'Mitral Valve Insufficiency',
'Eczema',
'Sarcoidosis',
'Gastroparesis',
'Varicose Veins',
'Thrombocytosis',
'Interstitial Lung Diseases',
'Lung Diseases, Interstitial',
'Hives',
'Urticaria',
'Peptic Ulcer',
'Gastric ulcer',
'Stomach ulcer',
'Duodenal ulcer',
'Polycythemia Vera',
'Postpartum Hemorrhage',
'Biliary Tract Neoplasms',
'Leukemia, Erythroblastic, Acute',
'Carcinoma, Intraductal, Noninfiltrating',
'Pain',
'Dengue',
'Memory',
'Cachexia',
'Dengue fever',
'Kidney Calculi',
'Break-bone fever',
'Memory Disorders',
'Periodontal Diseases',
'Jaundice',
'Gastritis',
'Fecal Incontinence',
'Radiation Injuries',
'Cardiomyopathy, Dilated',
'Scoliosis',
'Fabry Disease',
'Chronic Bronchitis',
'Bronchitis, Chronic',
'Retinitis Pigmentosa',
'Dilatation, Pathologic',
'Leukemia, Monocytic, Acute',
'Leukemia, Megakaryoblastic, Acute',
'Ataxia',
...]
In [7]:
inst = pickle.load(open('../data/facility_match_good.pkl','rb'))
In [8]:
typeahead = []
for t in final:
typeahead.append({'label': t, 'category': 'Condition'})
inst_order = [k for k, c in sorted([(b, inst[b]['trials']) for b in inst.keys()], key=lambda x: x[1], reverse=True)]
for b in inst_order:
if 'types' in inst[b]:
typeahead.append({'label': b, 'category': 'Institution'})
In [9]:
with open('../data/typeahead.json','wb') as fp:
json.dump(typeahead, fp)
In [18]:
locations = []
for b in inst.keys():
if 'geo' in inst[b]:
locations.append([inst[b]['geo']['lat'], inst[b]['geo']['lng'], inst[b]['name']])
In [21]:
with open('../data/locations.json','wb') as fp:
json.dump(locations, fp)
In [ ]:
Content source: jasonost/clinicaltrials
Similar notebooks: