Statistical Algorithms Importer: Create Project

From Gcube Wiki
Revision as of 10:47, 4 February 2016 by Giancarlo.panichi (Talk | contribs) (Use Annotations)

Jump to: navigation, search
How to create a project using Statistical Algorithms Importer(SAI) portlet.

Project Folder

Fist step is to create or select a empty folder in workspace using Create Project button in main menu. After a empty project in that folder is created.
Create Project, SAI

Import Resources

Any resources necessary for project can be imported in Project Folder. If resources is on Workspace you can use Add Resource button in main menu for select the file on workspace.
Add Resource, SAI
Otherwise if the resource is on your pc you can use Drag and Drop, and you can move the file from your pc to Project Explorer panel, multi selection is allowed.
Add Resource DnD, SAI

Set Main Code

After you have add R file, you can set it as Main code by click Set Main button in Project Explorer. The file will be load in the Editor. Now you can change the code and save it by Save button on the Editor panel. If you prefer, you can also use Copy and Paste by writing the code directly in the editor and then save it using Save button in Editor menu(A file name will be required).
Set Main Code, SAI

Input

In this area you can set all input information to create software.

Global Variables

In this panel you can add any useful Global Variable.
Global Variable, SAI

Input/Output

For select input and output, you can select the row on the code in the Editor and than ckick Input or Output button on the Menu Editor. A new row is added to Input/Output list. Now if you need to you you can change your information by clicking on the row. At least one input is required for the project.
Input/Output, SAI

Interprer Info

You can add Version and Packages informations in Interpreter Info panel. The version number is mandatory for the project.
Interpreter Info, SAI

Project Info

Name and Description of project are mandatory, in addittion for Name special characters are not allowed. You can include a list of requested VREs for this algorithm.
Project Info, SAI

Save Project

You can save project by click on Save button in main menu. A file called stat_algo.project is add to Project Folder.
Save Project, SAI


Use Annotations

If you import an R code with annotations, the system automatically try to fills the information into Input/Output panel and Project Info panel. Name of algorithm is mandatory in the annotations. This is an example of annotations:
############################################################################################################################
############# Absence Generation Script - Gianpaolo Coro and Chiara Magliozzi, CNR 2015, Last version 06-07-2015 ###########
############################################################################################################################

#52North WPS annotations
# wps.des: id = Absence_generation_from_OBIS, title = Absence_generation_from_OBIS, abstract = A script to estimate absence records from OBIS;

rm(list=ls(all=TRUE))
graphics.off() 

## charging the libraries
library(DBI)
library(RPostgreSQL)
library(raster)
library(maptools)

# time
t0<-Sys.time()

## parameters 
# wps.in: id = list, type = text/plain, title = list of species beginning with the speciesname header,value="species.txt";
list= "species.txt"
specieslist<-read.table(list,header=T,sep=",") # my short dataset 2 species
#attach(specieslist)
# wps.in: id = res, type = double, title = resolution of the analysis,value=1;
res=1;
extent_x=180
extent_y=90
n=extent_y*2/res;
m=extent_x*2/res;
# wps.in: id = occ_percentage, type = double, title = percentage of observations occurrence of a viable survey,value=0.1;
occ_percentage=0.1 #between 0 and 1

#uncomment for time filtering

#No time filter
TimeStart<-"";
TimeEnd<-"";

TimeStart<-gsub("(^ +)|( +$)", "",TimeStart)
TimeEnd<-gsub("(^ +)|( +$)", "", TimeEnd)

## opening the connection with postgres
cat("Opening the connection with the catalog\n")
drv <- dbDriver("PostgreSQL")
con <- dbConnect(drv, dbname="obis", host="obisdb-stage.vliz.be", port="5432", user="obisreader", password="0815r3@d3r")

cat("Analyzing the list of species\n")
counter=0;
overall=length(specieslist$scientificname)

cat("Extraction from the different contributors the total number of obs per resource id...\n")

timefilter<-""
if (nchar(TimeStart)>0 && nchar(TimeEnd)>0)
  timefilter<-paste(" where datecollected>'",TimeStart,"' and datecollected<'",TimeEnd,"'",sep="");

queryCache <- paste("select drs.resource_id, count(distinct position_id) as allcount from obis.drs", timefilter, " group by drs.resource_id",sep="")
cat("Resources extraction query:",queryCache,"\n")

allresfile="allresources.dat"
if (file.exists(allresfile)){
  load(allresfile)
} else{
  allresources1<-dbGetQuery(con,queryCache)
  save(allresources1,file=allresfile)
}

files<-vector()
f<-0
dir.create("data")

for (sp in specieslist$scientificname){
  f<-f+1
  t1<-Sys.time()
  graphics.off()
  grid=matrix(data=0,nrow=n,ncol=m)
  gridInfo=matrix(data="",nrow=n,ncol=m)
  outputfileAbs=paste("data/Absences_",sp,"_",res,"deg.csv",sep="");
  outputimage=paste("data/Absences_",sp,"_",res,"deg.png",sep="");
  
  counter=counter+1;
  cat("analyzing species",sp,"\n")
  cat("***Species status",counter,"of",overall,"\n")
  
  ## first query: select the species
  cat("Extraction the species id from the OBIS database...\n")
  query1<-paste("select id from obis.tnames where tname='",sp,"'", sep="")
  obis_id<- dbGetQuery(con,query1)
  cat("The ID extracted is ", obis_id$id, "for the species", sp, "\n", sep=" ")
  
  if (nrow(obis_id)==0) {
    cat("WARNING: there is no reference code for", sp,"\n")
    next;
  }
  
  ## second query: select the contributors
  cat("Selection of the contributors in the database having recorded the species...\n")
  query2<- paste("select distinct resource_id from obis.drs where valid_id='",obis_id$id,"'", sep="")
  posresource<-dbGetQuery(con,query2)
  
  if (nrow(posresource)==0) {
    cat("WARNING: there are no resources for", sp,"\n")
    next;
  }
  
  
  ## third query: select from the contributors different observations
  merge(allresources1, posresource, by="resource_id")-> res_ids
  
  ## forth query: how many obs are contained in each contributors for the species
  cat("Extraction from the different contributors the number of obs for the species...\n")
  query4 <- paste("select drs.resource_id, count(distinct position_id) as tgtcount from obis.drs where valid_id='",obis_id,"'group by drs.resource_id ",sep="")
  tgtresources1<-dbGetQuery(con,query4)
  merge(tgtresources1, posresource, by="resource_id")-> tgtresourcesSpecies 
  
  ## fifth query: select contributors that has al least 0.1 observation of the species
  #### we have the table all together: contributors, obs in each contributors for at leat one species and obs of the species in each contributors
  cat("Extracting the contributors containing more than 10% of observations for the species\n")
  tmp <- merge(res_ids, tgtresourcesSpecies, by= "resource_id",all.x=T)
  tmp["species_10"] <- NA 
  tmp$tgtcount / tmp$allcount -> tmp$species_10
  
  viable_res_ids <- subset(tmp,species_10 >= occ_percentage, select=c("resource_id","allcount","tgtcount", "species_10")) 
  #cat(viable_res_ids)
  
  if (nrow(viable_res_ids)==0) {
    cat("WARNING: there are no viable points for", sp,"\n")
    next;
  }
  
  numericselres<-paste("'",paste(as.character(as.numeric(t(viable_res_ids["resource_id"]))),collapse="','"),"'",sep="")
  
  ## sixth query: select all the cell at 0.1 degrees resolution in the main contributors
  cat("Select the cells at 0.1 degrees resolution for the main contributors\n")
  query6 <- paste("select position_id, positions.latitude, positions.longitude, count(*) as allcount ", 
                  "from obis.drs ", 
                  "inner join obis.tnames on drs.valid_id=tnames.id ",
                  "inner join obis.positions on position_id=positions.id ",
                  "where resource_id in (", numericselres,") ",
                  "group by position_id, positions.latitude, positions.longitude, resource_id")
  all_cells <- dbGetQuery(con,query6)
  
  ## seventh query:  select all the cell at 0.1 degrees resolution in the main contributors for the selected species
  cat("Select the cells at 0.1 degrees resolution for the species in the main contributors\n")
  query7 <- paste("select position_id, positions.latitude, positions.longitude, count(*) as tgtcount ",
                  "from obis.drs",
                  "inner join obis.tnames on drs.valid_id=tnames.id ", 
                  "inner join obis.positions on position_id=positions.id ", 
                  "where resource_id in (", numericselres,") ",
                  "and drs.valid_id='",obis_id,"'", 
                  "group by position_id, positions.latitude, positions.longitude")
  presence_cells<-dbGetQuery(con,query7)
  
  ## last query: for every cell in the sixth query if there is a correspondent in the seventh query I can put 1 otherwise 0
  data.df<-merge(all_cells, presence_cells, by= "position_id",all.x=T)
  data.df$longitude.y<-NULL 
  data.df$latitude.y<-NULL
  data.df[is.na(data.df)] <- 0 
  
  ######### Table resulting from the analysis
  pres_abs_cells <- subset(data.df,select=c("latitude.x","longitude.x", "tgtcount","position_id"))
  positions<-paste("'",paste(as.character(as.numeric(t(pres_abs_cells["position_id"]))),collapse="','"),"'",sep="")
  
  query8<-paste("select position_id, resfullname,digirname,abstract,temporalscope,date_last_harvested",
                "from ((select distinct position_id,resource_id from obis.drs where position_id IN (", positions,
                ") order by position_id ) as a",
                "inner join (select id,resfullname,digirname,abstract,temporalscope,date_last_harvested from obis.resources where id in (",
                numericselres,")) as b on b.id = a.resource_id) as d")
  
  resnames<-dbGetQuery(con,query8)
  #sorting data df
  pres_abs_cells<-pres_abs_cells[with(pres_abs_cells, order(position_id)), ]
  
  nrows = nrow(pres_abs_cells)
  ######## FIRST Loop inside the rows of the dataset
  cat("Looping on the data\n")
  for(i in 1: nrows) {
    lat<-pres_abs_cells[i,1]
    long<-pres_abs_cells[i,2]
    value<-pres_abs_cells[i,3]
    resource_name<-paste("\"",paste(as.character(t(resnames[i,])),collapse="\",\""),"\"",sep="")#resnames[i,2]
    k=round((lat+90)*n/180)
    g=round((long+180)*m/360)
    if (k==0) k=1;
    if (g==0) g=1;
    if (k>n || g>m)
      next;
    if (value>=1){
      if (grid[k,g]==0){
        grid[k,g]=1
        gridInfo[k,g]=resource_name
      }
      else if (grid[k,g]==-1){
        grid[k,g]=-2
        gridInfo[k,g]=resource_name
      }
    }
    else if (value==0){
      if (grid[k,g]==0){
        grid[k,g]=-1
        #cat("resource abs",resource_name,"\n")
        gridInfo[k,g]=resource_name
      }
      else if (grid[k,g]==1){
        grid[k,g]=-2
        gridInfo[k,g]=resource_name
      }
      
    }
  }
  cat("End looping\n")
  
  cat("Generating image\n")
  absence_cells<-which(grid==-1,arr.ind=TRUE)
  presence_cells_idx<-which(grid==1,arr.ind=TRUE)
  latAbs<-((absence_cells[,1]*180)/n)-90
  longAbs<-((absence_cells[,2]*360)/m)-180
  latPres<-((presence_cells_idx[,1]*180)/n)-90
  longPres<-((presence_cells_idx[,2]*360)/m)-180
  resource_abs<-gridInfo[absence_cells]
  
  absPoints <- cbind(longAbs, latAbs)
  absPointsData <- cbind(longAbs, latAbs,resource_abs)
  
  if (length(absPoints)==0)
  {
    cat("WARNING no viable point found for ",sp," after processing!\n")
    next;
  }
  data(wrld_simpl)
  projection(wrld_simpl) <- CRS("+proj=longlat")
  png(filename=outputimage, width=1200, height=600)
  plot(wrld_simpl, xlim=c(-180, 180), ylim=c(-90, 90), axes=TRUE, col="black")
  box()
  pts <- SpatialPoints(absPoints,proj4string=CRS(proj4string(wrld_simpl)))
  
  ## Find which points do not fall over land
  cat("Retreiving the poing that do not fall on land\n")
  pts<-pts[which(is.na(over(pts, wrld_simpl)$FIPS))]
  points(pts, col="green", pch=1, cex=0.50)
  datapts<-as.data.frame(pts)
  colnames(datapts) <- c("longAbs","latAbs")
  
  abspointstable<-merge(datapts, absPointsData, by.x= c("longAbs","latAbs"), by.y=c("longAbs","latAbs"),all.x=F)
  
  
  header<-"longitude,latitude,resource_id,resource_name,resource_identifier,resource_abstract,resource_temporalscope,resource_last_harvested_date"
  write.table(header,file=outputfileAbs,append=F,row.names=F,quote=F,col.names=F)
  
  write.table(abspointstable,file=outputfileAbs,append=T,row.names=F,quote=F,col.names=F,sep=",")
  files[f]<-outputfileAbs
  cat("Elapsed:  created imaged in ",Sys.time()-t1," sec \n")
  graphics.off()
}

# wps.out: id = zipOutput, type = text/zip, title = zip file containing absence records and images;
zipOutput<-"absences.zip"
zip(zipOutput, files=c("./data"), flags= "-r9X", extras = "",zip = Sys.getenv("R_ZIPCMD", "zip"))

cat("Closing database connection")
cat("Elapsed:  overall process finished in ",Sys.time()-t0," min \n")
dbDisconnect(con)
graphics.off()


File:AbsencesSpeciesList prod annotated.zip