diff options
author | chen42 | 2024-06-21 17:16:02 -0500 |
---|---|---|
committer | chen42 | 2024-06-21 17:16:02 -0500 |
commit | a223741881aadb73e75ed6d5d7d4e20dca1cda90 (patch) | |
tree | 66c32bf4a1f114530afad4e7184403ae1f013fc1 /server.py | |
parent | 0d000d860296614d641bc79e1f3e2cca48820fae (diff) | |
download | genecup-a223741881aadb73e75ed6d5d7d4e20dca1cda90.tar.gz |
correct regex order, update gwas data
Diffstat (limited to 'server.py')
-rwxr-xr-x | server.py | 57 |
1 files changed, 28 insertions, 29 deletions
@@ -1,47 +1,44 @@ #!/bin/env python3 from __future__ import print_function -from flask import Flask, render_template, request, session, Response, redirect, url_for, flash -from flask_sqlalchemy import SQLAlchemy -from flask import jsonify -import json -import shutil -import bcrypt import hashlib -import tempfile +import json +import os import random -import string import re -import pytz -import os +import shutil +import string +import tempfile from os import listdir -import pandas as pd + +import bcrypt +import nltk import numpy as np +import pandas as pd +import pytz +from flask import (Flask, Response, flash, jsonify, redirect, render_template, + request, session, url_for) +from flask_sqlalchemy import SQLAlchemy from numpy import array -import nltk nltk.download('punkt') -from nltk.corpus import stopwords -from nltk.stem.porter import PorterStemmer - +import pickle from collections import Counter from datetime import datetime -from more_functions import * import tensorflow import tensorflow.keras -from tensorflow.keras.models import Model -from tensorflow.keras.preprocessing.text import Tokenizer -from tensorflow.keras.preprocessing.sequence import pad_sequences -from tensorflow.keras.layers import * -from tensorflow.keras.models import Sequential -from tensorflow.keras.layers import Dense -from tensorflow.keras.layers import Flatten -from tensorflow.keras.layers import Embedding -from tensorflow.keras import metrics -from tensorflow.keras import optimizers +from nltk.corpus import stopwords +from nltk.stem.porter import PorterStemmer from tensorflow.keras import backend as K -import pickle +from tensorflow.keras import metrics, optimizers +from tensorflow.keras.layers import * +from tensorflow.keras.layers import Dense, Embedding, Flatten +from tensorflow.keras.models import Model, Sequential +from tensorflow.keras.preprocessing.sequence import pad_sequences +from tensorflow.keras.preprocessing.text import Tokenizer + +from more_functions import * app=Flask(__name__) #datadir="/export/ratspub/" @@ -907,8 +904,10 @@ def search(): sentences+=sent if ("GWAS" in search_type): gwas_sent=[] - datf_sub1 = datf[datf['REPORTED GENE(S)'].str.contains('(?:\s|^)'+gene+'(?:\s|$)', flags=re.IGNORECASE) - | (datf['MAPPED_GENE'].str.contains('(?:\s|^)'+gene+'(?:\s|$)', flags=re.IGNORECASE))] + print (datf) + datf_sub1 = datf[datf["MAPPED_GENE"].str.contains('(?:\s|^)'+gene+'(?:\s|$)', flags=re.IGNORECASE) + | (datf["REPORTED GENE(S)"].str.contains('(?:\s|^)'+gene+'(?:\s|$)', flags=re.IGNORECASE))] + print (datf_sub1) for nd2 in dict_onto['GWAS'].keys(): for nd1 in dict_onto['GWAS'][nd2]: for nd in nd1.split('|'): |