Add fault composition chart

2026-08-10 22:12:30 +02:00
parent 76a8fb4c4a
commit 836197a6ee
2 changed files with 103 additions and 120 deletions
@@ -0,0 +1,103 @@
+library(ggplot2)
+library(dplyr)
+library(readr)
+library(stringr)
+library(tidyr)
+
+# Usage: Rscript marker_composition.r exp_abspath1 ... resultsdata_file
+
+args <- commandArgs(trailingOnly = TRUE)
+if (length(args) < 2) {
+  stop("Need at least 2 experiments")
+}
+
+csv_suffix <- if (grepl("\\.csv$", args[length(args)])) {
+  args[length(args)]
+} else {
+  "resultsdata.csv"
+}
+exp_args <- if (grepl("\\.csv$", args[length(args)])) {
+  args[-length(args)]
+} else {
+  args
+}
+
+extract_info <- function(path) {
+  dir_name <- basename(path)
+  match <- str_match(
+    dir_name,
+    "^\\d{2}-\\d{2}_\\d{2}-\\d{2}-\\d{2}_(.+?)_(c|aot|interp)_"
+  )
+  if (is.na(match[1, 1])) {
+    warning(paste("Could not parse:", dir_name))
+    return(NULL)
+  }
+  list(base_name = match[1, 2], variant = match[1, 3], path = path)
+}
+
+# Load data
+all_data <- data.frame()
+weight_data <- data.frame()
+
+for (arg in exp_args) {
+  info <- extract_info(arg)
+  if (is.null(info)) {
+    next
+  }
+
+  csv_file <- file.path(info$path, csv_suffix)
+  if (!file.exists(csv_file)) {
+    warning(paste("Missing:", csv_file))
+    next
+  }
+  df <- read_csv(csv_file, col_types = cols())
+  df$base_name <- info$base_name
+  df$variant <- info$variant
+  all_data <- bind_rows(all_data, df)
+}
+
+if (nrow(all_data) == 0) {
+  stop("No data loaded")
+}
+
+# Skip OK_MARKERs, sum GROUP1 + TRAP.
+all_data <- all_data |>
+  filter(resulttype != "OK_MARKER") |>
+  mutate(
+    resulttype = ifelse(resulttype == "GROUP1_MARKER", "TRAP", resulttype)
+  ) |>
+  group_by(base_name, variant, benchmark, resulttype) |>
+  summarise(faults = sum(faults), .groups = "drop")
+
+out_suffix <- gsub("^resultsdata|\\.csv$", "", csv_suffix)
+
+# Calculate percentages
+composition <- all_data |>
+  group_by(base_name, variant, benchmark) |>
+  mutate(frac = faults / sum(faults)) |>
+  ungroup()
+
+plot <- ggplot(
+  composition |> filter(variant %in% c("aot", "interp")),
+  aes(x = variant, y = frac, fill = resulttype)
+) +
+  geom_col() +
+  facet_grid(benchmark ~ base_name) +
+  labs(
+    title = "Marker Composition (AOT vs Interp)",
+    x = NULL,
+    y = "Percentage of Faults",
+    fill = "Fault Type"
+  ) +
+  theme_minimal() +
+  theme(
+    plot.title = element_text(size = 13, face = "bold"),
+    axis.text.x = element_text(angle = 45, hjust = 1)
+  )
+
+filename <- paste0(
+  "injections/fault_composition",
+  out_suffix,
+  ".svg"
+)
+ggsave(filename, plot = plot, width = 13, height = 8)
@@ -1,120 +0,0 @@
-library(ggplot2)
-library(dplyr)
-library(readr)
-library(viridisLite)
-
-# Usage: Rscript combined_instr_fault_correlation_heatmap.r exp_abspath1 ... [faults_file]
-
-args <- commandArgs(trailingOnly = TRUE)
-if (length(args) < 1) {
-  stop("Need at least 1 experiment")
-}
-
-csv_suffix <- if (grepl("\\.csv$", args[length(args)])) {
-  args[length(args)]
-} else {
-  "faults.csv"
-}
-exp_args <- if (grepl("\\.csv$", args[length(args)])) {
-  args[-length(args)]
-} else {
-  args
-}
-
-all_data <- data.frame()
-all_mnem <- data.frame()
-
-for (arg in exp_args) {
-  csv_file <- file.path(arg, csv_suffix)
-  if (!file.exists(csv_file)) {
-    warning(paste("Missing:", csv_file))
-    next
-  }
-
-  df <- read_csv(csv_file, col_types = cols())
-  df$experiment <- basename(arg)
-  all_data <- bind_rows(all_data, df)
-
-  # TODO: This is ignoring any filters currently
-  mnem_file <- file.path(arg, "mnemonics.csv")
-  if (!file.exists(mnem_file)) {
-    warning(paste("Missing:", mnem_file))
-    next
-  }
-  mdf <- read_csv(mnem_file, col_types = cols())
-  all_mnem <- bind_rows(all_mnem, mdf)
-}
-
-if (nrow(all_data) == 0) {
-  stop("No faults.csv data loaded")
-}
-
-if (nrow(all_mnem) == 0) {
-  stop("No mnemonics.csv data loaded")
-}
-
-# no OK_MARKER, sum GROUP1 + TRAP.
-all_data <- all_data |>
-  filter(!is.na(mnemonic), mnemonic != "NULL") |>
-  filter(resulttype != "OK_MARKER") |>
-  mutate(
-    resulttype = ifelse(resulttype == "GROUP1_MARKER", "TRAP", resulttype)
-  )
-
-if (nrow(all_data) == 0) {
-  stop("No failure-marker data to plot")
-}
-
-# Sum faults per (mnemonic, marker) pair for all experiments
-heat <- all_data |>
-  group_by(mnemonic, resulttype) |>
-  summarise(faults = sum(faults, na.rm = TRUE), .groups = "drop")
-
-# Sum mnemonic counts for all experiments
-mnem_counts <- all_mnem |>
-  filter(!is.na(mnemonic), mnemonic != "NULL") |>
-  group_by(mnemonic) |>
-  summarise(count = sum(count, na.rm = TRUE), .groups = "drop")
-
-# Normalize by mnemonic count
-heat <- heat |>
-  left_join(mnem_counts, by = "mnemonic") |>
-  filter(!is.na(count), count > 0) |>
-  mutate(fault_rate = faults / count)
-
-if (nrow(heat) == 0) {
-  stop("Heat join failed")
-}
-
-# Order by fault rate
-mnem_order <- heat |>
-  group_by(mnemonic) |>
-  summarise(total = sum(fault_rate), .groups = "drop") |>
-  arrange(desc(total)) |>
-  pull(mnemonic)
-
-heat <- heat |>
-  mutate(mnemonic = factor(mnemonic, levels = mnem_order))
-
-plot <- ggplot(
-  heat,
-  aes(x = mnemonic, y = resulttype, fill = fault_rate)
-) +
-  geom_tile(colour = "white") +
-  scale_fill_viridis_c(name = "Fault rate", trans = "log10") +
-  labs(
-    title = "Instruction / Fault Rate Heatmap (Normalized)",
-    x = "Instruction",
-    y = "Fault Type"
-  ) +
-  theme_minimal() +
-  theme(
-    axis.text.x = element_text(angle = 90, hjust = 1),
-    panel.grid = element_blank(),
-    plot.title = element_text(size = 14, face = "bold")
-  )
-
-suffix <- gsub("^faults|\\.csv$", "", csv_suffix)
-outfile <- paste0("injections/instr_fault_rate_heatmap", suffix, ".svg")
-ggsave(outfile, plot = plot, width = 12, height = 6)
-print(paste("Saved", outfile))