bittorrent-benchmarks/analysis/final.analysis/static-dissemination.Rmd

---
title: "static-dissemination.Rmd"
output: html_document
date: "2025-01-10"
---

```{r}
library(tidyverse)

devtools::load_all()
```

```{r}
group_id <- 'g1736505161'
```


```{r}
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)
  if (is.null(download_time_stats)) {
    NULL
  } else {
    compute_compact_summary(download_time_stats)
  }
}) |> 
  drop_nulls() |>
  bind_rows() |>
  arrange(seeders) |>
  arrange(network_size) |>
  arrange(file_size)
```