# complete documentation for this script can be found at :
# http://ucsb.piscoweb.org/~cburt/intertidal/workflow.html
classes = c(
 	'factor',		#site_code
 	'factor',		#survey_rep
 	'integer',	#year
 	'integer',	#month
 	'integer',	#day
 	'factor',		#section
 	'factor',		#transect
 	'factor',		#location
	'factor',		#classification_code
 	'factor',		#sampler_code
 	'factor',		#zone_code
 	'integer'		#count
)
quad_data	= read.csv(
	file="quad_numeric.csv", 
	na.strings='.', 
	head=TRUE, sep=",", 
	colClasses = classes
)
# this script will only handle the number of rows in nrow for the output.
# as of 2006 this is 30,000 rows
data = matrix(nrow=100000, ncol=7)
#pointer for adding rows to the output matrix
pointer = 0
sites = unique(quad_data$site_code);
class_codes = levels(quad_data$class_code)
num_of_sites = length(sites)
site_count = 0
for(site in sites){
	site_count = site_count + 1
	message = paste('site', site_count, '/', num_of_sites)
	print(message)
	data_for_site = subset(quad_data, site_code == sites[site]);
	years = unique(data_for_site$year)
	num_of_years = length(years)
	year_counter = 0
	for(y in years){
		year_counter = year_counter + 1
		message = paste('year', year_counter, '/', num_of_years)
		print(message)
		data_for_year = subset(data_for_site, year == y)
		#taking the taxon list from the year data. 
		#So assuming that if a species does not have a zero datapoint, it has not been sampled
		cc = unique(data_for_year$class_code)
		for(c in cc){
			data_for_taxon = subset(data_for_year, class_code == class_codes[c])
			counts = data_for_taxon$count * 4
			mean = mean(counts)
			if(!is.na(mean)){
				pointer = pointer + 1
				standard_deviation = sd(counts)
				n = length(counts)
				standard_error = standard_deviation / sqrt(n)
				data[pointer, ] = c(
					data_for_taxon$year[1],		#year 
					sites[site], 							#site
					class_codes[c], 					#class_code
					mean, 										#mean
					standard_deviation, 			#standard_deviation
					standard_error, 					#standard_error
					n													#replicates, number of quads
				)
			}else{
				print('warning: mean is NA')
			}
		}
	}
}
write.table(
	na.omit(data),
	file="./quad_output.csv",
	row.names=FALSE,
	col.names = FALSE,
	sep="\t",
	quote=FALSE
)