analysis.csp

# Generated by Extended CovScript Compiler
# DO NOT MODIFY
# Date: Fri Nov  3 00:45:35 2023
import ecs as analysis_ecs
package analysis
import analysis_impl as impl
import codec.json as json
import regex
struct operator
	var name = null
	var column = null
	var impl_func = null
	function as(str)
		name = str
		return this
	end
	function run(data)
		return impl_func(data, column)
	end
end
function __impl_make_operator(name, col, func)
	var op = new operator
	op.name = name
	op.column = col
	op.impl_func = func
	return move(op)
end
function avg(col)
	return __impl_make_operator("avg", col, impl.avg)
end
function average(col)
	return __impl_make_operator("avg", col, impl.avg)
end
function sum(col)
	return __impl_make_operator("sum", col, impl.sum)
end
function summery(col)
	return __impl_make_operator("sum", col, impl.sum)
end
function count(col)
	return __impl_make_operator("count", col, impl.count)
end
function count_distinct(col)
	return __impl_make_operator("count_distinct", col, impl.count_distinct)
end
function __impl_align(text, width)
	var txt = " " + text
	var pad = width - text.size - 1
	foreach it in range(pad) do txt += " "
	return move(txt)
end
function __impl_show(header, data)
	var maxs = new array
	var width = header.size
	foreach it in header do maxs.push_back(it.size)
	foreach line in data
		if line.size != width
			system.out.println("analysis.dataframe.show: broken line founded!")
			return this
		end
		foreach i in range(width)
			var size = to_string(line.at(i)).size
			if size > maxs.at(i)
				maxs.at(i) = size
			end
		end
	end
	system.out.print("#")
	foreach idx in range(width) do system.out.print(__impl_align(header.at(idx), maxs.at(idx) + 2) + "#")
	system.out.println("")
	foreach line in data
		system.out.print("|")
		foreach idx in range(width) do system.out.print(__impl_align(to_string(line.at(idx)), maxs.at(idx) + 2) + "|")
		system.out.println("")
	end
end
class datarecord
	var host = null
	var data = null
	function construct(df, line)
		host = df
		data = line
	end
	function by_name(name)
		return data->at(host->header_map.at(name))
	end
	function by_index(idx)
		return data->at(idx)
	end
	function show()
		__impl_show(host->header, {*data})
		return this
	end
end
class dataframe
	var header_map = null
	var header = null
	var data = null
	function construct(_header)
		header_map = new hash_map
		foreach idx in range(_header.size) do header_map.insert(_header.at(idx), idx)
		header = _header
		data = new array
	end
	function filter(args, cond)
		if args.empty()
			return this
		end
		var cols = new array
		foreach it in args do cols.push_back(header_map.at(it))
		return __impl_make_dataframe(header, impl.filter(data, cols, cond))
	end
	function __groupby_impl(cols, iteration, child_data, idx)
		var result = (idx == 0 ? impl.groupby(child_data, cols.at(idx)) : impl.groupby_group(data, child_data, cols.at(idx)))
		if iteration > 0
			foreach it in result do it.second = __groupby_impl(cols, iteration - 1, it.second, idx + 1)
		end
		return move(result)
	end
	function groupby(...args)
		if args.empty()
			return this
		end
		var cols = new array
		foreach it in args do cols.push_back(header_map.at(it))
		return __impl_make_dataframe_groupby(header, &data, cols, __groupby_impl(cols, cols.size - 1, data, 0))
	end
	function select(...args)
		var cols = new array
		foreach it in args do cols.push_back(header_map.at(it))
		return __impl_make_dataframe(args, impl.select(data, cols))
	end
	function aggregate(...args)
		var result_header = new array, result_array = new array
		foreach op in args
			result_header.push_back(op.name)
			op.column = header_map.at(op.column)
			result_array.push_back(op.run(data))
		end
		return __impl_make_dataframe(result_header, {result_array})
	end
	function attach(...args)
		var result_header = header, result_array = new array
		foreach op in args
			result_header.push_back(op.name)
			op.column = header_map.at(op.column)
			result_array.push_back(op.run(data))
		end
		var new_data = data
		foreach line in new_data
			foreach it in result_array do line.push_back(it)
		end
		return __impl_make_dataframe(result_header, new_data)
	end
	function apply(col, func)
		var col_idx = header_map.at(col)
		foreach line in data do func(line.at(col_idx))
		return this
	end
	function enumerate(func)
		foreach i in range(data.size)
			func(i, analysis_ecs.param_new(datarecord, {&this, &data.at(i)}))
		end
		return this
	end
	function foreach(func)
		foreach it in data
			func(analysis_ecs.param_new(datarecord, {&this, &it}))
		end
		return this
	end
	function pick(func)
		var results = new array
		foreach line in data
			if func(analysis_ecs.param_new(datarecord, {&this, &line}))
				results.push_back(line)
			end
		end
		return __impl_make_dataframe(header, move(results))
	end
	function by_index(idx)
		return analysis_ecs.param_new(datarecord, {&this, &data.at(idx)})
	end
	function append(...lines)
		foreach line in lines
			if line.size == header.size
				data.push_back(move(line))
			end
		end
		return this
	end
	function join_v(df, pkey, fill_val, drop_invalid)
		var pidx = null, pmap = null
		if pkey != null && header_map.exist(pkey) && df.header_map.exist(pkey)
			pidx = header_map.at(pkey)
			pmap = new hash_map
			foreach i in range(data.size) do pmap.insert(data.at(i).at(pidx), i)
		end
		var index = new array
		foreach key in header
			if df.header_map.exist(key)
				index.push_back(df.header_map.at(key))
			else
				index.push_back(null)
			end
		end
		foreach rec in df.data
			var line = new array
			foreach pos in index
				if pos != null
					line.push_back(rec.at(pos))
				else
					line.push_back(fill_val)
				end
			end
			if pmap != null
				if pmap.exist(line.at(pidx))
					link target = data.at(pmap.at(line.at(pidx)))
					foreach i in range(target.size)
						if i != pidx && index.at(i) != null
							swap(target.at(i), line.at(i))
						end
					end
				else
					if !drop_invalid
						data.push_back(move(line))
					end
				end
			else
				data.push_back(move(line))
			end
		end
		return this
	end
	function join(df)
		return join_v(df, null, null, false)
	end
	function find_broken_lines()
		var lines = new array
		foreach i in range(data.size)
			if data.at(i).size != header.size
				lines.push_back(i + 2)
			end
		end
		return move(lines)
	end
	function to_csv(file)
		var csv_data = data
		csv_data.push_front(header)
		impl.write_csv(csv_data, file)
		return this
	end
	function to_json()
		var j_arr = json.value.make_array()
		foreach line in data
			var j_obj = json.value.make_object()
			foreach i in range(header.size) do j_obj.set_member(header.at(i), json.value.make_string(to_string(line.at(i))))
			j_arr.arr_append(j_obj)
		end
		return move(j_arr)
	end
	function show()
		__impl_show(header, data)
		return this
	end
end
function __impl_make_dataframe(header, data)
	var df = new dataframe
	df.header_map = new hash_map
	foreach idx in range(header.size) do df.header_map.insert(header.at(idx), idx)
	df.header = header
	df.data = data
	return move(df)
end
function read_csv(path)
	var data = impl.read_csv(path)
	if data != null
		link header = data.front
		data.pop_front()
		return __impl_make_dataframe(header, move(data))
	end
end
function cartesian_product(dfx, dfy, pkey, fill_val, drop_invalid)
	var pidx = null, pmap = null
	if pkey != null && dfx.header_map.exist(pkey) && dfy.header_map.exist(pkey)
		pidx = dfx.header_map.at(pkey)
		pmap = new hash_map
		foreach i in range(dfx.data.size) do pmap.insert(dfx.data.at(i).at(pidx), i)
	end
	var keys = hash_set.intersect(analysis_ecs.type_constructor.__hash_set(dfx.header), analysis_ecs.type_constructor.__hash_set(dfy.header))
	var header = dfx.header
	var index = new array
	foreach key in dfx.header
		if dfy.header_map.exist(key)
			index.push_back(dfy.header_map.at(key))
		else
			index.push_back(null)
		end
	end
	foreach key in dfy.header
		if !keys.exist(key)
			header.push_back(key)
			index.push_back(dfy.header_map.at(key))
		end
	end
	var data = dfx.data
	foreach rec in data
		foreach i in range(header.size - rec.size)
			rec.push_back(fill_val)
		end
	end
	foreach rec in dfy.data
		var line = new array
		foreach pos in index
			if pos != null
				line.push_back(rec.at(pos))
			else
				line.push_back(fill_val)
			end
		end
		if pmap != null
			if pmap.exist(line.at(pidx))
				link target = data.at(pmap.at(line.at(pidx)))
				foreach i in range(target.size)
					if i != pidx && index.at(i) != null
						swap(target.at(i), line.at(i))
					end
				end
			else
				if !drop_invalid
					data.push_back(line)
				end
			end
		else
			data.push_back(line)
		end
	end
	return __impl_make_dataframe(header, data)
end
function __impl_get_group(group, key, target)
	if typeid group != typeid hash_map
		throw analysis_ecs.throw_exception(runtime.exception("Group not exist!"))
	end
	if group.exist(key)
		target.push_back(group.at(key))
	else
		foreach it in group do __get_group_impl(it.second, key, target)
	end
end
function __impl_unstack_group_impl(group, target)
	if typeid group == typeid hash_map
		foreach it in group do __impl_unstack_group_impl(it.second, target)
	else
		foreach it in group do target.push_back(it)
	end
end
function __impl_unstack_group(group, data)
	var lines = new array
	__impl_unstack_group_impl(group, lines)
	var result = new array
	foreach idx in lines do result.push_back(data.at(idx))
	return move(result)
end
function __impl_aggregate_group(stack, data, group, target, ops)
	if typeid group == typeid hash_map
		target = new hash_map
		foreach it in group
			var child_data = null
			stack.push_back(it.first)
			__impl_aggregate_group(stack, data, it.second, child_data, ops)
			stack.pop_back()
			target.insert(it.first, child_data)
		end
	else
		target = stack
		var lines = new array
		foreach idx in group do lines.push_back(data.at(idx))
		foreach op in ops do target.push_back(op.run(lines))
		target = {target}
	end
end
function __impl_attach_group(data, group, target, ops)
	if typeid group == typeid hash_map
		target = new hash_map
		foreach it in group
			var child_data = null
			__impl_attach_group(data, it.second, child_data, ops)
			target.insert(it.first, child_data)
		end
	else
		target = new array
		var result = new array
		var lines = new array
		foreach idx in group do lines.push_back(data.at(idx))
		foreach op in ops do result.push_back(op.run(lines))
		foreach it in lines do target.push_back({it..., result...})
	end
end
class dataframe_groupby
	var header_map = null
	var header = null
	var groups = null
	var g_cols = null
	var data = null
	function get_group(key)
		var group_set = new array, lines = new array, result = new array
		__get_group_impl(groups, key, group_set)
		foreach it in group_set do __impl_unstack_group_impl(it, lines)
		foreach idx in lines do result.push_back(data->at(idx))
		return __impl_make_dataframe(header, result)
	end
	function unstack()
		return __impl_make_dataframe(header, __impl_unstack_group(groups, *data))
	end
	function aggregate(...args)
		var result_header = new array, result_map = null, result_array = new array
		foreach idx in g_cols do result_header.push_back(header.at(idx))
		foreach op in args
			result_header.push_back(op.name)
			op.column = header_map.at(op.column)
		end
		var stack = new array
		__impl_aggregate_group(stack, *data, groups, result_map, args)
		__impl_unstack_group_impl(result_map, result_array)
		return __impl_make_dataframe(result_header, result_array)
	end
	function attach(...args)
		var result_header = header, result_map = null, result_array = new array
		foreach op in args
			result_header.push_back(op.name)
			op.column = header_map.at(op.column)
		end
		__impl_attach_group(*data, groups, result_map, args)
		__impl_unstack_group_impl(result_map, result_array)
		return __impl_make_dataframe(result_header, result_array)
	end
end
function __impl_make_dataframe_groupby(header, data, cols, groups)
	var df = new dataframe_groupby
	df.header_map = new hash_map
	foreach idx in range(header.size) do df.header_map.insert(header.at(idx), idx)
	df.header = header
	df.groups = groups
	df.g_cols = cols
	df.data = data
	return move(df)
end
function enumerate(data, func)
	foreach i in range(data.size)
		func(i, data.at(i))
	end
end
function match_regex(reg, str)
	var result = reg.match(str)
	if result.empty()
		return null
	end
	var data = new array
	foreach i in range(1, result.size())
		data.push_back(result.str(i))
	end
	return move(data)
end