add plot scripts back

utilities/plotMovie/createMov.sh

@@ -0,0 +1,61 @@
+#!/bin/bash
+module load python
+
+GROUP="FIPS"
+ZOOM="NO"
+BAYAREA="NO"
+for i in "$@"; do
+  case $i in
+    -c=*|--censusMap=*)
+      CENSUS="${i#*=}"
+      shift # past argument=value
+      ;;
+    -i=*|--infectionCases=*)
+      INFECTION="${i#*=}"
+      shift # past argument=value
+      ;;
+    -m=*|--plotMode=*) #CUML/DAILY
+      OUTPUT_MODE="${i#*=}"
+      shift # past argument=value
+      ;;
+    -n=*|--normalize=*) #RAW (no normalization), POP (normalize to population)
+      NORM="${i#*=}"
+      shift # past argument=value
+      ;;
+    -g=*|--group=*) #group data to a higher level (FIPS and GROUP are 2 options for now)
+      GROUP="${i#*=}"
+      shift # past argument=value
+      ;;
+    -z=*|--zoomin=*) #zoom in towards a location, format: "startTime, stopTime, target x, target y, zoomFactor"
+      ZOOM="${i#*=}"
+      shift # past argument=value
+      ;;
+    -b=*|--bayAreaDataFormat=*)
+      BAYAREA="${i#*=}"
+      shift # past argument=value
+      ;;
+    -*|--*)
+      echo "Unknown option $i"
+      exit 1
+      ;;
+    *)
+      ;;
+  esac
+done
+
+echo "PATH TO INFECTION CASES: ${INFECTION}"
+echo "PATH TO CENSUS MAP: ${CENSUS}"
+echo "MODE TO PLOT: ${OUTPUT_MODE}"
+echo "NORMALIZE TO: ${NORM}"
+echo "GROUP TO (DEFAULT: FIPS): ${GROUP}"
+echo "ZOOM IN (DEFAULT: NO): ${ZOOM}"
+echo "BAY AREA DATA FORMAT (DEFAULT: NO): ${BAYAREA}"
+
+#plot everyday's cases on a cencus map
+python3 plotCases.py ${INFECTION} ${CENSUS} ${OUTPUT_MODE} ${NORM} ${GROUP} ${ZOOM} ${BAYAREA}
+
+#create movie
+ffmpeg -framerate 1 -i "casefile%05d.png" -c:v libx264 -r 30 -pix_fmt yuv420p movie.mp4
+
+#remove temporary files
+rm *.png

utilities/plotMovie/plotCases.py

@@ -0,0 +1,154 @@
+import matplotlib as mpl
+import matplotlib.pyplot as plt
+from mpl_toolkits.axes_grid1.inset_locator import inset_axes
+from matplotlib import colors
+from matplotlib import cm
+import os
+import fiona
+import geopandas as geopd
+import pysal as ps
+import shapefile
+import pandas as pd
+from shapely.geometry import shape
+import numpy as np
+import os
+import sys
+import re
+
+argList= sys.argv[1:]
+zoom=""
+group=""
+BayAreaDataFormat=""
+if(len(argList)<5): #5 is the minimum mandatory arguments
+    print("Some mandatory argument is missing")
+    exit
+if len(argList)>= 1: infCasesDir= argList[0]+"/"
+if len(argList)>= 2: CensusDir  = argList[1]+"/"
+if len(argList)>= 3: outputMetric  = argList[2]
+if len(argList)>= 4: datanorm  = argList[3]
+if len(argList)>= 5: group  = argList[4]
+if len(argList)>= 6: zoom  = argList[5] #format: "startTime, stopTime, target x, target y, zoomFactor"
+if len(argList)>= 7: BayAreaDataFormat= argList[6]
+
+gdf = geopd.GeoDataFrame(columns=["STATEFP00", "COUNTYFP00", "TRACTCE00", "CTIDFP00", "NAME00", "NAMELSAD00", "MTFCC00", "FUNCSTAT00", "ALAND00", "AWATER00", "INTPTLAT00", "INTPTLON00", "geometry"])
+#for Bay Area dataset, use the following format:
+if(BayAreaDataFormat=="YES"): gdf = geopd.GeoDataFrame(columns=["geoid", "name_", "namelsad", "mtfcc", "funcstat", "aland", "awater", "intptlat", "intptlon", "SHAPE_Leng", "SHAPE_Area"])
+
+infCasesFiles= [os.path.join(infCasesDir, f) for f in os.listdir(infCasesDir) if f.endswith(".csv")]
+
+for file in os.listdir(CensusDir):
+    filename = CensusDir + file + "/" + file
+    myshp = open(filename+".shp", "rb")
+    mydbf = open(filename+".dbf", "rb")
+    myprj = open(filename+".prj", "rb")
+    r = shapefile.Reader(shp=myshp, dbf=mydbf, prj=myprj)
+    attributes, geometry = [], []
+    field_names = [field[0] for field in r.fields[1:]]
+    for row in r.shapeRecords():
+        geometry.append(shape(row.shape.__geo_interface__))
+        attributes.append(dict(zip(field_names, row.record)))
+    df = geopd.GeoDataFrame(data = attributes, geometry = geometry)
+    gdf= gdf._append(df)
+    r.close()
+
+if(BayAreaDataFormat=="NO"): gdf['fips'] = gdf["CTIDFP00"].astype(int)
+else: gdf['fips'] = gdf["geoid"].astype(int)
+gdf["county"]= gdf["fips"]//1000000
+gdf0= gdf.copy()
+gdf1= gdf.copy()
+gdf0['per'] = 0.0
+gdf1['per'] = 0.0
+gdf1.set_index('fips',inplace=True)
+
+
+sorted_infCasesFiles= sorted(infCasesFiles, key=lambda x: int( re.findall('\d+', x)[0]))
+for caseFile in sorted_infCasesFiles:
+    idx = int(re.findall('\d+', caseFile)[0])
+    print("processing file", caseFile)
+    df000 = pd.read_csv(caseFile, header=None, names=["fips", "per"], dtype={"fips":int, "per":float})
+    #df000.set_index("fips",inplace=True)
+    gdf_join = pd.merge(gdf,df000, left_on="fips", right_on="fips", how='inner')
+    if(group=="COUNTY"):
+        if(datanorm=="RAW"):
+            df000["county"]= df000["fips"]//1000000
+            df000.drop("fips", inplace=True, axis=1)
+            df000= df000.groupby("county").sum()
+            gdf_join = pd.merge(gdf,df000, left_on='county', right_on='county', how='inner')
+    #gdf_join.set_index('fips',inplace=True)
+    columns= gdf1.columns.values
+    columns= columns[:-1].tolist() #exclude per
+    if(outputMetric=="DAILY"):
+        gdf1= pd.concat([gdf0, gdf_join]).groupby(columns).sum().reset_index()
+    if(outputMetric=="CUML"):
+        gdf1= pd.concat([gdf0, gdf_join]).groupby(columns).sum().reset_index()
+
+    fig, ax = plt.subplots(figsize=(10,10))
+
+    gdfplot= geopd.GeoDataFrame(gdf1)
+    #use cmap with a norm to convert a number to RGB
+    #two slope norm
+    #set center to 0, adjust the negative vmin to control how fast color changes from 0 to vmax
+    if(datanorm=="RAW"):
+        vmin = -1.0
+        vmax = 25
+        norm = colors.TwoSlopeNorm(vmin=vmin, vcenter=0.000, vmax=vmax)
+        cmap = cm.cool
+        sm = plt.cm.ScalarMappable(cmap=cmap, norm=norm)
+        gdfplot["per"] = gdfplot["per"] + 1 #avoid log(0)
+        gdfplot["per"] = np.log2(gdfplot["per"])
+        plt.colorbar(sm, ax=ax, ticks=np.linspace(0, 25, 6), label="Log2 Cumulative Cases", orientation="vertical")
+    else:
+        if(datanorm=="POP"):
+            vmin = -10.0
+            vmax = 1.0
+            norm = colors.TwoSlopeNorm(vmin=vmin, vcenter=0.000, vmax=vmax)
+            cmap = cm.cool
+            sm = plt.cm.ScalarMappable(cmap=cmap, norm=norm)
+            plt.colorbar(sm, ax=ax, ticks=np.linspace(0, 1, 11), label="Cumulative Cases per Population", orientation="vertical")
+        else:
+            print("INPUT METRIC NOT PROVIDED", norm)
+            exit
+    ax.xaxis.set_ticks_position('top')
+    ax.set(aspect='equal', xticks=[], yticks=[])
+
+    gdfplot.plot(column= 'per', ax = ax, norm=norm, cmap=cmap, lw=3)#, edgecolor='black',)#, legend=False)
+
+    #adjust the position and zoom in
+    if(zoom!="NO"):
+        startTime, stopTime, targetX, targetY, zoomFactor= zoom.split(',')
+        startTime= int(startTime)
+        stopTime= int(stopTime)
+        shiftTime=10 #we need 10 days to smoothly shift to the target position
+        targetX= float(targetX)
+        targetY= float(targetY)
+        zoomFactor= float(zoomFactor)
+        xrate= (0.5-targetX)/10
+        yrate= (0.5-targetY)/10
+        zoomrate= (1.0-1.0/zoomFactor)/(stopTime-shiftTime-startTime)
+
+        ylim = ax.get_ylim()
+        xlim = ax.get_xlim()
+        if(idx>=startTime and idx<(startTime+shiftTime)):
+            #gradually shift to the interested local area
+            ax.set_xlim((xlim[0]-(xlim[1]-xlim[0])*xrate*(idx-startTime), xlim[1]-(xlim[1]-xlim[0])*xrate*(idx-startTime)))
+            ax.set_ylim((ylim[0]-(ylim[1]-ylim[0])*yrate*(idx-startTime), ylim[1]-(ylim[1]-ylim[0])*yrate*(idx-startTime)))
+        if(idx>=(startTime+shiftTime) and idx<stopTime):
+            #now all figures start looking at the interested local area
+            ax.set_xlim((xlim[0]-(xlim[1]-xlim[0])*xrate*shiftTime, xlim[1]-(xlim[1]-xlim[0])*xrate*shiftTime) )
+            ax.set_ylim((ylim[0]-(ylim[1]-ylim[0])*yrate*shiftTime, ylim[1]-(ylim[1]-ylim[0])*yrate*shiftTime) )
+            #we want to zoom in a little more
+            ylim = ax.get_ylim()
+            xlim = ax.get_xlim()
+            ax.set_xlim((xlim[0]+(xlim[1]-xlim[0])*zoomrate*(idx-startTime-shiftTime)/2, xlim[1]-(xlim[1]-xlim[0])*zoomrate*(idx-startTime-shiftTime)/2) )
+            ax.set_ylim((ylim[0]+(ylim[1]-ylim[0])*zoomrate*(idx-startTime-shiftTime)/2, ylim[1]-(ylim[1]-ylim[0])*zoomrate*(idx-startTime-shiftTime)/2) )
+        if(idx>=stopTime):
+            #stay exactly where last figure was. No more shifting, no more zooming
+            ax.set_xlim((xlim[0]-(xlim[1]-xlim[0])*xrate*shiftTime, xlim[1]-(xlim[1]-xlim[0])*xrate*shiftTime) )
+            ax.set_ylim((ylim[0]-(ylim[1]-ylim[0])*yrate*shiftTime, ylim[1]-(ylim[1]-ylim[0])*yrate*shiftTime) )
+            ylim = ax.get_ylim()
+            xlim = ax.get_xlim()
+            ax.set_xlim((xlim[0]+(xlim[1]-xlim[0])*zoomrate*(stopTime-shiftTime-startTime)/2, xlim[1]-(xlim[1]-xlim[0])*zoomrate*(stopTime-shiftTime-startTime)/2 ))
+            ax.set_ylim((ylim[0]+(ylim[1]-ylim[0])*zoomrate*(stopTime-shiftTime-startTime)/2, ylim[1]-(ylim[1]-ylim[0])*zoomrate*(stopTime-shiftTime-startTime)/2 ))
+    output= "casefile"+str(idx+1).zfill(5)+".png"
+    plt.savefig(output)
+    plt.close()