app.py

from email import message
from os import stat
from flask import Flask,render_template,request,session,jsonify
import csv
import pandas as pd
import numpy as np
import pickle
from os.path import exists
from dataExtractor import sigLocExtract, surveyMCA, surveyPerformance, scatterPca_1_2, eigGap, validIDs
import os

app = Flask(__name__)

surveyFilePath = 'input/surveyData/imputed/locf/features_'
sigLocFilePath = 'input/SignificantLocations/features_'

fileFeatures = {
    surveyFilePath:['MCA Trend', 'MCA Correlation', 'MCA Eigen-Gap', 'Date-Clustering'],
    sigLocFilePath:['Hometime', 'Significant Location'],
}
patientIds = ['XXXXXXXXXX']

meta = {"meta1" : {'panelId':'#panel1Viz','f':None,'f_opt':[],'pid':"XXXXXXXXXX"},
"meta2" : {'panelId':'#panel2Viz','f':None,'f_opt':[],'pid':"XXXXXXXXXX"}}

state = {1:0, 2:0}


#homepage
@app.route('/',methods=['GET','POST'])
def index():
    pfeaturesDict = {}
    for ind in range(1,3):
        i = str(ind)
        p1ID = request.form.get('p'+i+'ID')
        pFeatures = []
        if(p1ID != None and p1ID != meta['meta'+i]['pid']):
            meta['meta'+i]['pid'] = p1ID
            meta['meta'+i]['f'] = None
            state[1] = 1
            state[2] = 0

        for fileloc in list(fileFeatures.keys()):
            if(exists(fileloc+meta['meta'+i]['pid']+'.csv')):
                pFeatures.extend(fileFeatures[fileloc])
        f1 = request.form.get('f'+i)
        if(f1 != None and f1 != meta['meta'+i]['f']):
            meta['meta'+i]['f'] = f1
            state[1] = 1
            state[2] = 0
        if(meta['meta'+i]['f'] == None):
            meta['meta'+i]['f_opt'] = []
        elif(meta['meta'+i]['f'] == 'MCA Trend' or meta['meta'+i]['f'] == 'MCA Correlation'):
            optList = request.form.getlist('svp'+i+'_opt')
            prev_meta_state = meta['meta'+i]['f_opt']
            if (meta['meta'+i]['f_opt'] == [] and optList == []) or meta['meta'+i]['f_opt'][0] not in ['mood','anxiety','social','sleep','psychosis']:
                meta['meta'+i]['f_opt'] = ['mood']
            elif optList != []:
                meta['meta'+i]['f_opt'] = optList
            impList = request.form.getlist('svp'+i+'_imp')
            if (len(meta['meta'+i]['f_opt']) <= 1):
                if(len(impList) == 0):
                    meta['meta'+i]['f_opt'].append('locf') #default option
                else:
                    meta['meta'+i]['f_opt'].extend(impList) #add the selected option
            else:
                meta['meta'+i]['f_opt'] = meta['meta'+i]['f_opt'][:2] #trim the list to 2 if it was bigger
                if(len(impList) == 0 and meta['meta'+i]['f_opt'][-1] not in ['locf','mice']):
                    meta['meta'+i]['f_opt'][-1] = 'locf'
                elif(len(impList) > 0):
                    meta['meta'+i]['f_opt'][-1] = impList[0] 
                #The else case lets the previous value of imputation method be as it is
            if(prev_meta_state != meta['meta'+i]['f_opt']): #detecting change in options
                state[1] = 1
                state[2] = 0
        elif(meta['meta'+i]['f'] == 'MCA Eigen-Gap'):
            optList = request.form.getlist('svp'+i+'_opt')
            if (meta['meta'+i]['f_opt'] == []) and optList == [] or meta['meta'+i]['f_opt'][0] not in ['firstEigenGap', 'secondEigenGap']:
                meta['meta'+i]['f_opt'] = ['firstEigenGap']
            elif optList != []:
                meta['meta'+i]['f_opt'] = optList
                state[1] = 1
                state[2] = 0
            impList = request.form.getlist('svp'+i+'_imp')
            if (len(meta['meta'+i]['f_opt']) <= 1):
                if(len(impList) == 0):
                    meta['meta'+i]['f_opt'].append('locf') #default option
                else:
                    meta['meta'+i]['f_opt'].extend(impList) #add the selected option
            else:
                meta['meta'+i]['f_opt'] = meta['meta'+i]['f_opt'][:2] #trim the list to 2 if it was bigger
                if(len(impList) == 0 and meta['meta'+i]['f_opt'][-1] not in ['locf','mice']):
                    meta['meta'+i]['f_opt'][-1] = 'locf'
                elif(len(impList) > 0):
                    meta['meta'+i]['f_opt'][-1] = impList[0] 
        elif(meta['meta'+i]['f'] == 'Date-Clustering'):
            optList = request.form.getlist('svPer'+i)
            if (meta['meta'+i]['f_opt'] == [] and optList == []) or meta['meta'+i]['f_opt'][0] not in ['Aggregate','Complete']:
                meta['meta'+i]['f_opt'] = ['Aggregate', "Natural clustering"]
            if(optList and (optList[0] == 'Aggregate' or optList[0] == 'Complete')):
                meta['meta'+i]['f_opt'] = [optList[0]]
                state[1] = 1
                state[2] = 0
            if(optList and optList[1]):
                meta['meta'+i]['f_opt'].append(optList[1])
        elif(meta['meta'+i]['f'] == 'Hometime' or meta['meta'+i]['f'] == 'Significant Location'):
            my = request.form.get('p'+i+'MY')
            df = pd.read_csv(sigLocFilePath+meta['meta'+i]['pid']+'.csv')
            temp = pd.to_datetime(df['start'],dayfirst=True)
            validMonths = {}
            month = temp.min().month
            year = temp.min().year
            maxmonth = temp.max().month
            maxyear = temp.max().year
            minyear = temp.min().year
            while(year <= maxyear):
                yearlist = []
                if(year != maxyear):
                    while(month <= 12): 
                        yearlist.append(month)
                        month += 1
                    validMonths[year] = yearlist
                    month = 1
                else:
                    while(month <= maxmonth):
                        yearlist.append(month)
                        month += 1
                    validMonths[year] = yearlist
                year += 1
            timeRange = (minyear, min(validMonths[minyear]), maxyear, max(validMonths[maxyear]))        
            if(my != None):
                state[2] = 0
                state[1] = 1  
                try:
                    M, Y = map(int,my.split('-'))
                    if(Y not in validMonths or M not in validMonths[Y]):
                        M = maxmonth
                        Y = maxyear
                except:
                    M, Y = (timeRange[3], timeRange[2])    
            else:
                try:
                    M, Y, _ = meta['meta'+i]['f_opt']
                except:
                    M, Y = (timeRange[3], timeRange[2])
            meta['meta'+i]['f_opt'] = [M,Y,timeRange]
        pfeaturesDict[ind] = pFeatures

    selected = {'p1ID':meta['meta1']['pid'], 'p2ID':meta['meta2']['pid'],
                'f1':meta['meta1']['f'], 'f2':meta['meta2']['f']}

    options = {'f1':pfeaturesDict[1], 'f2':pfeaturesDict[2], 
               'f1_opt':meta['meta1']['f_opt'], 'f2_opt':meta['meta2']['f_opt']}
    return render_template('base.html',selected=selected,options=options,patientIds=validIDs(fileFeatures))

@app.route('/render')
def render():
    curState = []
    curState.append(state[1])
    curState.append(state[2])
    curState.append(meta['meta1'])
    curState.append(meta['meta2'])
    return jsonify(curState)

@app.route('/sigLocdata')
def sigLocdata():
    data1 = []
    data2 = []
    if(meta['meta1']['f'] == 'Hometime' or meta['meta1']['f'] == 'Significant Location'):
        data1 = sigLocExtract(sigLocFilePath, meta['meta1'])
    if(meta['meta2']['f'] == 'Hometime' or meta['meta2']['f'] == 'Significant Location'):
        data2 = sigLocExtract(sigLocFilePath, meta['meta2'])
    return jsonify([data1,data2])

@app.route('/surveydata')
def surveyViz():
    data = []
    for i in range(1,3):
        curData = []
        i = str(i)
        filePath = surveyFilePath #By default LOCF
        if('mice' in meta['meta'+i]['f_opt']):
            filePath = 'input/surveyData/imputed/mice/features_'
        if(meta['meta'+i]['f'] == 'MCA Trend'):
            curData = surveyMCA(filePath, meta['meta'+str(i)])
        elif(meta['meta'+i]['f'] == 'Date-Clustering'):
            curData = surveyPerformance(filePath, meta['meta'+str(i)])
        elif(meta['meta'+i]['f'] == 'MCA Correlation'):
            curData = scatterPca_1_2(filePath, meta['meta'+str(i)])
        elif(meta['meta'+i]['f'] == 'MCA Eigen-Gap'):
            curData = eigGap(filePath, meta['meta'+str(i)])
        data.append(curData)
    return jsonify(data)

@app.route("/matrix_seriation/<meta_number>", methods=['GET'])
def matrix_seriation(meta_number):
    return render_template("matrix_seriation.html", url="/matrix_data/" + meta_number)


@app.route("/matrix_data/<meta_number>")
def matrix_data(meta_number):
    meta_mapping = {
        "vis1": meta['meta1'],
        "vis2": meta['meta2'],
    }
    data = surveyPerformance(surveyFilePath, meta_mapping[meta_number])
    json_data = {
        "nodes": [],
        "links": []
    }
    dates = data[1]['dates']
    values = data[2]['data']
    count = 0

    for i in range(len(dates)): 
        for j in range(len(dates)): 
            json_data["links"].append({
                "source": i,
                "target": j,
                "value": values[count]
            })
            count += 1
        json_data["nodes"].append({
            "name": dates[i],
            "group": 1
        })

    return jsonify(json_data)