From 700753cf3dd3d80505fd453248252fe7e611bd27 Mon Sep 17 00:00:00 2001
From: gmega <giuliano.mega@gmail.com>
Date: Mon, 13 Jan 2025 20:16:41 -0300
Subject: [PATCH] chore: move analysis code into analysis; properly discard
 experiments that error out

---
 analysis/final.analysis/R/analysis.R          | 62 +++++++++++++++
 .../final.analysis/static-dissemination.Rmd   | 76 +------------------
 2 files changed, 64 insertions(+), 74 deletions(-)

diff --git a/analysis/final.analysis/R/analysis.R b/analysis/final.analysis/R/analysis.R
index e6da2e2..0a5bb0d 100644
--- a/analysis/final.analysis/R/analysis.R
+++ b/analysis/final.analysis/R/analysis.R
@@ -1,3 +1,9 @@
+PIECE_SIZE <- 262144
+
+piece_count <- function(experiment) {
+  experiment$meta$file_size / PIECE_SIZE
+}
+
 extract_repetitions <- function(deluge_torrent_download) {
   deluge_torrent_download |>
     mutate(
@@ -104,3 +110,59 @@ download_time_stats <- function(download_times) {
     )
 }
 
+compute_download_time_stats <- function(experiment) {
+  meta <- experiment$meta
+  pieces <- experiment |> piece_count()
+  downloads <- experiment$deluge_torrent_download |>
+    extract_repetitions() |>
+    compute_pieces(pieces)
+
+  if (!check_incomplete_downloads(downloads, pieces)) {
+    warning(glue::glue('Discard experiment {experiment$experiment_id} ',
+                       'due to incomplete downloads'))
+    return(NULL)
+  }
+
+  if (!check_mismatching_repetitions(downloads, meta$repetitions)) {
+    warning(glue::glue('Discard experiment {experiment$experiment_id} ',
+                       'due to mismatching repetitions'))
+    return(NULL)
+  }
+
+  download_times <- compute_download_times(
+    meta,
+    experiment$request_event,
+    downloads,
+    group_id
+  )
+
+  if (!check_seeder_count(download_times, meta$seeders)) {
+    warning(glue::glue('Undefined download times do not match seeder count'))
+    return(NULL)
+  }
+
+  network_size <- meta$nodes$network_size
+
+  download_times |>
+    download_time_stats() |>
+    mutate(
+      network_size = network_size,
+      seeders = meta$seeders,
+      leechers = network_size - meta$seeders,
+      file_size = meta$file_size
+    )
+}
+
+
+compute_compact_summary <- function(download_ecdf) {
+  lapply(c(0.05, 0.5, 0.95), function(p)
+    download_ecdf |>
+      filter(completed >= p) |>
+      slice_min(completed)
+  ) |>
+    bind_rows() |>
+    select(completed, network_size, file_size, seeders, leechers, median) |>
+    pivot_wider(id_cols = c('file_size', 'network_size', 'seeders', 'leechers'),
+                names_from = completed, values_from = median)
+}
+
diff --git a/analysis/final.analysis/static-dissemination.Rmd b/analysis/final.analysis/static-dissemination.Rmd
index 469dc5c..b063758 100644
--- a/analysis/final.analysis/static-dissemination.Rmd
+++ b/analysis/final.analysis/static-dissemination.Rmd
@@ -19,80 +19,10 @@ group_id <- 'g1736505161'
 deluge <- read_all_experiments('./data/deluge')
 ```
 
-The torrent piece size is set at torrent creation time by [torrentool](https://github.com/idlesign/torrentool/blob/5f37d6dcc304758bae46d01c63e5be0f0a348bfc/torrentool/torrent.py#L354).
-```{r}
-PIECE_SIZE <- 262144
-``` 
-
-```{r}
-piece_count <- function(experiment) {
-  experiment$meta$file_size / PIECE_SIZE
-}
-```
-
-```{r}
-compute_download_time_stats <- function(experiment) {
-  meta <- experiment$meta
-  pieces <- experiment |> piece_count()
-  downloads <- experiment$deluge_torrent_download |> 
-    extract_repetitions() |>
-    compute_pieces(pieces)
-
-  if (!check_incomplete_downloads(downloads, pieces)) {
-    warning(glue::glue('Discard experiment {experiment$experiment_id} ',
-                       'due to incomplete downloads'))
-    return(NULL)
-  }
-
-  if (!check_mismatching_repetitions(downloads, meta$repetitions)) {
-    warning(glue::glue('Discard experiment {experiment$experiment_id} ',
-                       'due to mismatching repetitions'))
-    return(NULL)
-  }
-
-  download_times <- compute_download_times(
-    meta, 
-    experiment$request_event, 
-    downloads, 
-    group_id
-  )
-
-  if (!check_seeder_count(download_times, meta$seeders)) {
-    warning(glue::glue('Undefined download times do not match seeder count'))
-    return(NULL)
-  }
-  
-  network_size <- meta$nodes$network_size
-  
-  download_times |> 
-    download_time_stats() |>
-    mutate(
-      network_size = network_size,
-      seeders = meta$seeders,
-      leechers = network_size - meta$seeders,
-      file_size = meta$file_size
-    )
-}
-
-
-compute_compact_summary <- function(download_ecdf) {
-  lapply(c(0.05, 0.5, 0.95), function(p)
-    download_ecdf |>
-      filter(completed >= p) |>
-      slice_min(completed)
-    ) |>
-    bind_rows() |>
-    select(completed, network_size, file_size, seeders, leechers, median) |>
-    pivot_wider(id_cols = c('file_size', 'network_size', 'seeders', 'leechers'),
-                names_from = completed, values_from = median)
-}
-
-```
-
 ```{r}
 lapply(deluge, function(experiment) {
   print(glue::glue('Process {experiment$experiment_id}'))
-  download_time_stats <- tryCatch(compute_download_time_stats(experiment), error = function(e) NULL)
+  download_time_stats <- tryCatch(compute_download_time_stats(experiment), error = function(e) { print(e); NULL })
   if (is.null(download_time_stats)) {
     NULL
   } else {
@@ -101,9 +31,7 @@ lapply(deluge, function(experiment) {
 }) |> 
   drop_nulls() |>
   bind_rows() |>
-  arrange(seeders) |>
-  arrange(network_size) |>
-  arrange(file_size)
+  arrange(file_size, network_size, seeders, leechers)
 ```