feat: build (internal) package infrastructure, and export these two functions

[jdhoffa]

workflow.pacta.dashboard/main.R

Lines 8 to 203 in e2f02a4

	prep_techmix_sector <-
	function(equity_results_portfolio,
	bonds_results_portfolio,
	indices_equity_results_portfolio,
	indices_bonds_results_portfolio,
	peers_equity_results_portfolio,
	peers_bonds_results_portfolio,
	investor_name,
	portfolio_name,
	start_year,
	year_span,
	peer_group,
	green_techs,
	all_tech_levels
	) {
	portfolio <-
	list(`Listed Equity` = equity_results_portfolio,
	`Corporate Bonds` = bonds_results_portfolio) %>%
	bind_rows(.id = "asset_class") %>%
	filter(.data$investor_name == .env$investor_name,
	.data$portfolio_name == .env$portfolio_name) %>%
	filter(!is.na(.data$ald_sector))
	asset_classes <-
	portfolio %>%
	pull("asset_class") %>%
	unique()
	equity_sectors <-
	portfolio %>%
	filter(.data$asset_class == "Listed Equity") %>%
	filter(.data$allocation == "portfolio_weight") %>%
	pull("ald_sector") %>%
	unique()
	bonds_sectors <-
	portfolio %>%
	filter(.data$asset_class == "Corporate Bonds") %>%
	pull("ald_sector") %>%
	unique()
	indices <-
	list(`Listed Equity` = indices_equity_results_portfolio,
	`Corporate Bonds` = indices_bonds_results_portfolio) %>%
	bind_rows(.id = "asset_class") %>%
	filter(.data$asset_class %in% .env$asset_classes) %>%
	filter(.data$asset_class == "Listed Equity" & .data$ald_sector %in% .env$equity_sectors |
	.data$asset_class == "Corporate Bonds" & .data$ald_sector %in% .env$bonds_sectors)
	peers <-
	list(`Listed Equity` = peers_equity_results_portfolio,
	`Corporate Bonds` = peers_bonds_results_portfolio) %>%
	bind_rows(.id = "asset_class") %>%
	filter(.data$asset_class %in% .env$asset_classes) %>%
	filter(.data$asset_class == "Listed Equity" & .data$ald_sector %in% .env$equity_sectors |
	.data$asset_class == "Corporate Bonds" & .data$ald_sector %in% .env$bonds_sectors) %>%
	filter(.data$investor_name == .env$peer_group)
	techexposure_data <-
	bind_rows(portfolio, peers, indices) %>%
	filter(.data$allocation == "portfolio_weight") %>%
	filter(.data$scenario_geography == "Global") %>%
	filter(.data$year %in% c(.env$start_year, .env$start_year + .env$year_span))
	if (nrow(techexposure_data) > 0) {
	techexposure_data <-
	techexposure_data %>%
	mutate(green = .data$technology %in% .env$green_techs) %>%
	group_by(.data$asset_class, .data$equity_market, .data$portfolio_name,
	.data$ald_sector, .data$scenario, .data$year) %>%
	mutate(plan_alloc_wt_sec_prod = sum(.data$plan_alloc_wt_tech_prod),
	scen_alloc_wt_sec_prod = sum(.data$scen_alloc_wt_tech_prod)) %>%
	mutate(production_plan = if_else(.data$plan_alloc_wt_tech_prod > 0, .data$plan_alloc_wt_tech_prod / .data$plan_alloc_wt_sec_prod, 0),
	scenario_plan = if_else(.data$scen_alloc_wt_tech_prod > 0, .data$scen_alloc_wt_tech_prod / .data$scen_alloc_wt_sec_prod, 0)) %>%
	group_by(.data$asset_class, .data$equity_market, .data$portfolio_name,
	.data$ald_sector, .data$scenario, .data$year, .data$green) %>%
	mutate(green_sum_prod = sum(.data$production_plan),
	green_sum_scenario = sum(.data$scenario_plan)) %>%
	ungroup() %>%
	select("asset_class", "investor_name", "portfolio_name", "scenario_source",
	"scenario", "allocation", "equity_market", "year", "ald_sector",
	"technology", "production_plan", "scenario_plan", "green",
	"green_sum_prod", "green_sum_scenario") %>%
	pivot_longer(
	cols = -c("asset_class", "investor_name", "portfolio_name", "scenario_source",
	"scenario", "allocation", "equity_market", "year", "ald_sector",
	"technology", "green","green_sum_prod", "green_sum_scenario"),
	names_to = "val_type", values_to = "value") %>%
	mutate(
	green_sum = if_else(.data$val_type == "production_plan", .data$green_sum_prod, .data$green_sum_scenario)
	) %>%
	select(-c("green_sum_prod", "green_sum_scenario")) %>%
	ungroup() %>%
	mutate(this_portfolio = .data$portfolio_name == .env$portfolio_name,
	val_type = if_else(.data$this_portfolio == TRUE, paste0(.data$val_type, "_portfolio"), paste0(.data$val_type, "_benchmark"))) %>%
	mutate(equity_market = case_when(
	.data$equity_market == "GlobalMarket" ~ "Global Market",
	.data$equity_market == "DevelopedMarket" ~ "Developed Market",
	.data$equity_market == "EmergingMarket" ~ "Emerging Market",
	TRUE ~ .data$equity_market)
	) %>%
	# no need for showing scenario mix for the benchmark
	filter(
	.data$val_type != "scenario_plan_benchmark"
	) %>%
	mutate(val_type = case_when(
	.data$val_type == "production_plan_portfolio" ~ "Portfolio",
	.data$val_type == "scenario_plan_portfolio" ~ "Scenario",
	.data$val_type == "production_plan_benchmark" ~ "Benchmark",
	TRUE ~ .data$val_type)
	) %>%
	arrange(
	.data$asset_class,
	factor(.data$equity_market, levels = c("Global Market", "Developed Market", "Emerging Market")),
	desc(.data$this_portfolio),
	factor(.data$val_type, levels = c("Portfolio", "Scenario", "Benchmark")),
	.data$portfolio_name,
	factor(.data$technology, levels = .env$all_tech_levels)
	) %>%
	select("asset_class", "equity_market", "portfolio_name", "scenario", "scenario_source",
	"this_portfolio", "val_type", "ald_sector", "technology", "value",
	"green", "green_sum", "year") %>%
	filter(
	!(.data$year == .env$start_year & .data$val_type == "Scenario")
	)
	}
	techexposure_data
	}
	prep_exposure_stats <- function(audit_file, investor_name, portfolio_name, pacta_sectors) {
	pacta_asset_classes <- c("Bonds", "Equity")
	audit_table <- pacta.portfolio.report:::prep_audit_table(
	audit_file,
	investor_name = investor_name,
	portfolio_name = portfolio_name,
	currency_exchange_value = currency_exchange_value
	)
	exposure_stats <- audit_file %>%
	filter(
	.data$investor_name == .env$investor_name &
	.data$portfolio_name == .env$portfolio_name) %>%
	filter(.data$asset_type %in% pacta_asset_classes) %>%
	filter(.data$valid_input == TRUE) %>%
	mutate(across(c("bics_sector", "financial_sector"), as.character)) %>%
	mutate(
	sector =
	if_else(!.data$financial_sector %in% .env$pacta_sectors,
	"Other",
	.data$financial_sector
	)
	) %>%
	summarise(
	value = sum(.data$value_usd, na.rm = TRUE) / .env$currency_exchange_value,
	.by = c("asset_type", "sector")
	) %>%
	mutate(
	perc_asset_val_sector = .data$value / sum(.data$value, na.rm = TRUE),
	.by = c("asset_type")
	) %>%
	inner_join(audit_table, by = join_by(asset_type == asset_type_analysis)) %>%
	select("asset_type", "percentage_value_invested", "sector", "perc_asset_val_sector")
	asset_classes_in_portfolio <- intersect(pacta_asset_classes, unique(exposure_stats$asset_type))
	all_stats_with_zero_sector_exposure <- expand.grid(
	asset_type = asset_classes_in_portfolio,
	sector = pacta_sectors,
	val_sector = 0
	) %>% inner_join(
	distinct(select(exposure_stats, c("asset_type", "percentage_value_invested"))),
	by = join_by(asset_type)
	)
	exposure_stats_all <- all_stats_with_zero_sector_exposure %>%
	left_join(exposure_stats, by = join_by(asset_type, sector, percentage_value_invested)) %>%
	mutate(
	perc_asset_val_sector = if_else(
	is.na(.data$perc_asset_val_sector),
	.data$val_sector,
	.data$perc_asset_val_sector
	)
	) %>%
	mutate(
	asset_type = case_when(
	.data$asset_type == "Bonds" ~ "Corporate Bonds",
	.data$asset_type == "Equity" ~ "Listed Equity"
	)
	) %>%
	select("asset_type", "percentage_value_invested", "sector", "perc_asset_val_sector")
	exposure_stats_all
	}

[AlexAxthelm]

@jdhoffa: the functions are prep_techmix_sector() and prep_exposure_stats()?

is there a reverse dependency I need to track?

[jdhoffa]

I suppose there are two main things to be aware of. Number 1 is in-scope for this current issue. Number 2 is something to be aware of, but probably a can we can kick down the road for now.

1. In main.R there are 5 functions that we should package and export:

• prep_techmix_sector
• prep_exposure_stats
• prep_company_bubble
• prep_key_bars_company
• prep_key_bars_portfolio

2. We are utilizing un-exported function from various PACTA packages, e.g. pacta.portfolio.report

workflow.pacta.dashboard/main.R

Lines 610 to 619 in 32fa462

	audit_file %>%
	pacta.portfolio.report:::prep_exposure_pie(
	asset_type = "Equity",
	investor_name = investor_name,
	portfolio_name = portfolio_name,
	pacta_sectors = pacta_sectors,
	currency_exchange_value = currency_exchange_value
	) %>%
	pacta.portfolio.report:::translate_df_contents("data_value_pie_equity", dictionary) %>%
	jsonlite::write_json(path = file.path(output_dir, "data_value_pie_equity.json"))

[jdhoffa]

is there a reverse dependency I need to track?

No reverse dependency in terms of function names/ arguments etc.! However of course the final datasets output are fixed, and have an expected format so that must be preserved.

[cjyetman]

I think this is resolved by #30

[AlexAxthelm]

Not resolved, as no functions were exported in that PR.