Polish the HTML report produced by benchmarks/osh-parser.sh.

- New csv2html.py that uses a .schema.csv file to determine column types and presentation. - Remove the old csv_to_html.py and its tests. - Label shells and hosts in R code. Link to the versioned IDs. - Rename "platform ID" to "host ID".
oilshell · Oct 31, 2017 · e7e8f19a5753a265670627b9c8c702f953b985b8 · e7e8f19
1 parent 394b260
commit e7e8f19a5753a265670627b9c8c702f953b985b8
diff --git a/benchmarks/id.sh b/benchmarks/id.sh
@@ -35,7 +35,7 @@ set -o errexit
 #     dash/
 #       HASH.txt
 #       version.txt
-#   platform-id/
+#   host-id/
 #     lisa/
 #       HASH.txt
 #       cpuinfo.txt
@@ -47,7 +47,7 @@ set -o errexit
 #     osh-$HASH/   # osh-cpython, osh-ovm?   osh-opy-ovm?  Too many dimensions.
 #                # the other shells don't have this?
 #     zsh-$HASH/
-#   platform-id/
+#   host-id/
 #     lisa-$HASH/

 die() {
@@ -158,11 +158,11 @@ publish-shell-id() {
 # - kernel upgrade
 # - distro upgrade

-# How about ~/git/oilshell/benchmark-data/platform-id/lisa-$HASH
+# How about ~/git/oilshell/benchmark-data/host-id/lisa-$HASH
 # How to calculate the hash though?

-dump-platform-id() {
-  local out_dir=${1:-_tmp/platform-id/$(hostname)}
+dump-host-id() {
+  local out_dir=${1:-_tmp/host-id/$(hostname)}

  mkdir -p $out_dir

@@ -192,7 +192,7 @@ dump-platform-id() {
 # http://wiki.osdev.org/Target_Triplet
 # It's not exactly the same as what we need here, but close.

-_platform-id-hash() {
+_host-id-hash() {
  local src=$1

  # Don't hash CPU or memory
@@ -210,13 +210,13 @@ _platform-id-hash() {
 }

 # Writes a short ID to stdout.
-publish-platform-id() {
-  local src=$1  # e.g. _tmp/platform-id/lisa
-  local dest_base=${2:-../benchmark-data/platform-id}
+publish-host-id() {
+  local src=$1  # e.g. _tmp/host-id/lisa
+  local dest_base=${2:-../benchmark-data/host-id}

  local name=$(basename $src)
  local hash
-  hash=$(_platform-id-hash $src | md5sum)  # not secure, an identifier
+  hash=$(_host-id-hash $src | md5sum)  # not secure, an identifier

  local id="${hash:0:8}"
  local dest="$dest_base/$name-$id"
diff --git a/benchmarks/osh-parser.R b/benchmarks/osh-parser.R
@@ -7,6 +7,7 @@

 library(dplyr)
 library(tidyr)
+library(stringr)

 options(stringsAsFactors = F)

@@ -15,9 +16,14 @@ Log = function(fmt, ...) {
  cat('\n')
 }

+sourceUrl = function(path) {
+  sprintf('https://github.com/oilshell/oil/blob/master/%s', path)
+}
+
 main = function(argv) {
  out_dir = argv[[1]]

+  # Merge all the inputs
  hosts = list()
  for (i in 2:length(argv)) {
    times_path = argv[[i]]
@@ -45,32 +51,74 @@ main = function(argv) {

    hosts[[i-1]] = host_rows
  }
+
  all_times = bind_rows(hosts)
-  print(all_times)
+  print(summary(all_times))
+
+  #
+  # Find distinct shells and hosts, and label them for readability.
+  #

-  all_times %>% distinct(platform_id) -> distinct_hosts
+  all_times %>% distinct(host_name, host_hash) -> distinct_hosts
+  # Just use the name
+  distinct_hosts$host_label = distinct_hosts$host_name
  print(distinct_hosts)
-  all_times %>% distinct(shell_id) -> distinct_shells
+
+
+  all_times %>% distinct(shell_name, shell_hash) -> distinct_shells
  print(distinct_shells)

-  return()
+  distinct_shells$shell_label = NA  # the column we fill in below
+
+  Log('Labeling shells')
+
+  for (i in 1:nrow(distinct_shells)) {
+    row = distinct_shells[i, ]
+    if (row$shell_name == 'osh') {
+      path = sprintf('../benchmark-data/shell-id/osh-%s/osh-version.txt',
+                     row$shell_hash)
+      Log('Reading %s', path)
+      lines = readLines(path)
+      if (length(grep('OVM', lines)) > 0) {
+        label = 'osh-ovm'
+      } else if (length(grep('CPython', lines)) > 0) {
+        label = 'osh-cpython'
+      }
+    } else {  # same name for other shells
+      label = row$shell_name
+    }
+    distinct_shells[i, ]$shell_label = label
+  }               
+  print(distinct_shells)
+
+  # Replace name/hash combinations with labels.
+  all_times %>%
+    left_join(distinct_hosts, by = c('host_name', 'host_hash')) %>%
+    left_join(distinct_shells, by = c('shell_name', 'shell_hash')) %>%
+    select(-c(host_name, host_hash, shell_name, shell_hash)) ->
+    all_times
+
+  print(summary(all_times))
+  print(head(all_times))

  # Summarize rates by platform/shell
  all_times %>%
-    group_by(shell_id, platform_id) %>%
+    group_by(host_label, shell_label) %>%
    summarize(total_lines = sum(num_lines), total_ms = sum(elapsed_ms)) %>%
    mutate(lines_per_ms = total_lines / total_ms) ->
-    rate_summary
+    shell_summary

-  print(rate_summary)
+  print(shell_summary)

  # Elapsed seconds for each shell by platform and file
  all_times %>%
    select(-c(lines_per_ms)) %>% 
-    spread(key = shell_id, value = elapsed_ms) %>%
-    arrange(platform_id, num_lines) ->
+    spread(key = shell_label, value = elapsed_ms) %>%
+    arrange(host_label, num_lines) %>%
+    mutate(filename = basename(path), filename_HREF = sourceUrl(path)) %>% 
+    select(c(host_label, bash, dash, mksh, zsh, `osh-ovm`, `osh-cpython`,
+             num_lines, filename, filename_HREF)) ->
    elapsed
-    #select(c(bash, dash, mksh, zsh, osh, num_lines, path)) ->

  Log('\n')
  Log('ELAPSED')
@@ -79,25 +127,65 @@ main = function(argv) {
  # Rates by file and shell
  all_times  %>%
    select(-c(elapsed_ms)) %>% 
-    spread(key = shell_id, value = lines_per_ms) %>%
-    arrange(platform_id, num_lines) ->
+    spread(key = shell_label, value = lines_per_ms) %>%
+    arrange(host_label, num_lines) %>%
+    mutate(filename = basename(path), filename_HREF = sourceUrl(path)) %>% 
+    select(c(host_label, bash, dash, mksh, zsh, `osh-ovm`, `osh-cpython`,
+             num_lines, filename, filename_HREF)) ->
    rate
-    #select(c(bash, dash, mksh, zsh, osh, num_lines, path)) ->

  Log('\n')
  Log('RATE')
  print(rate)

-  write.csv(rate_summary,
-            file.path(out_dir, 'rate_summary.csv'), row.names = F)
-  write.csv(elapsed, file.path(out_dir, 'elapsed.csv'), row.names = F)
-  write.csv(rate, file.path(out_dir, 'rate.csv'), row.names = F)
+  # Should be:
+  # host_id_url
+  # And then csv_to_html will be smart enough?  It should take --url flag?
+  host_table = data_frame(
+    host_label = distinct_hosts$host_label,
+    host_id = paste(distinct_hosts$host_name,
+                    distinct_hosts$host_hash, sep='-'),
+    host_id_HREF = sprintf('../../../../benchmark-data/host-id/%s', host_id)
+  )
+  print(host_table)
+
+  shell_table = data_frame(
+    shell_label = distinct_shells$shell_label,
+    shell_id = paste(distinct_shells$shell_name,
+                     distinct_shells$shell_hash, sep='-'),
+    shell_id_HREF = sprintf('../../../../benchmark-data/shell-id/%s', shell_id)
+  )
+  print(shell_table)
+
+  writeCsv(host_table, file.path(out_dir, 'hosts'))
+  writeCsv(shell_table, file.path(out_dir, 'shells'))
+  writeCsv(shell_summary, file.path(out_dir, 'summary'))
+  writeCsv(elapsed, file.path(out_dir, 'elapsed'))
+  writeCsv(rate, file.path(out_dir, 'rate'))

  Log('Wrote %s', out_dir)

  Log('PID %d done', Sys.getpid())
 }

+# Write a CSV file along with a schema.
+writeCsv = function(table, prefix) {
+  data_out_path = paste0(prefix, '.csv')
+  write.csv(table, data_out_path, row.names = F)
+
+  fieldType = function(field_name) { typeof(table[[field_name]]) }
+
+  types_list = lapply(names(table), fieldType)
+  types = as.character(types_list)
+
+  schema = data_frame(
+    column_name = names(table),
+    type = types
+  )
+  schema_out_path = paste0(prefix, '.schema.csv')
+  write.csv(schema, schema_out_path, row.names = F)
+}
+
 if (length(sys.frames()) == 0) {
  # increase ggplot font size globally
  #theme_set(theme_grey(base_size = 20))