From 4aa26f6d1d0daf70833a79c48be9578f76f1719b Mon Sep 17 00:00:00 2001
From: Ethan Williams <ethan@ethan.ws>
Date: Thu, 24 Oct 2024 14:18:00 -0400
Subject: [PATCH] Just covid mts (#27)

* work on covid-mts

* add covid-mts to CI
---
 .github/workflows/tests.yml           |  2 +-
 covid-mts/cleanup.sh                  | 10 ++--
 covid-mts/deps.sh                     | 14 ++++++
 covid-mts/hashes/outputs.md5sum       |  4 ++
 covid-mts/hashes/outputs_small.md5sum |  4 ++
 covid-mts/hashes/small/1.hash         |  1 -
 covid-mts/hashes/small/2.hash         |  1 -
 covid-mts/hashes/small/3.hash         |  1 -
 covid-mts/hashes/small/4.hash         |  1 -
 covid-mts/input.sh                    | 11 +++++
 covid-mts/input/.gitignore            |  1 +
 covid-mts/inputs.sh                   | 15 ------
 covid-mts/run.sh                      | 57 +++++++---------------
 covid-mts/verify.sh                   | 69 +++++----------------------
 14 files changed, 71 insertions(+), 120 deletions(-)
 create mode 100755 covid-mts/deps.sh
 create mode 100644 covid-mts/hashes/outputs.md5sum
 create mode 100644 covid-mts/hashes/outputs_small.md5sum
 delete mode 100644 covid-mts/hashes/small/1.hash
 delete mode 100644 covid-mts/hashes/small/2.hash
 delete mode 100644 covid-mts/hashes/small/3.hash
 delete mode 100644 covid-mts/hashes/small/4.hash
 create mode 100755 covid-mts/input.sh
 create mode 100644 covid-mts/input/.gitignore
 delete mode 100755 covid-mts/inputs.sh

diff --git a/.github/workflows/tests.yml b/.github/workflows/tests.yml
index 2d3e3dca..dbf5a5d3 100644
--- a/.github/workflows/tests.yml
+++ b/.github/workflows/tests.yml
@@ -11,7 +11,7 @@ jobs:
     runs-on: ubuntu-latest
     strategy:
       matrix:
-        benchmark: [nlp, file-enc, unix50, log-analysis, max-temp, uniq-ips, media-conv, sklearn]
+        benchmark: [nlp, file-enc, unix50, log-analysis, max-temp, uniq-ips, media-conv, sklearn, covid-mts]
 
     steps:
       - name: Checkout code
diff --git a/covid-mts/cleanup.sh b/covid-mts/cleanup.sh
index 2e2e12f3..d8680650 100755
--- a/covid-mts/cleanup.sh
+++ b/covid-mts/cleanup.sh
@@ -1,5 +1,9 @@
 #!/bin/bash
 
-cd "$(realpath $(dirname "$0"))"
-rm -rf ./inputs
-rm -rf ./outputs
+REPO_TOP=$(git rev-parse --show-toplevel)
+eval_dir="${REPO_TOP}/covid-mts"
+outputs_dir="${eval_dir}/outputs"
+input_dir="${eval_dir}/input"
+
+rm -rf "$outputs_dir"
+rm -rf "$input_dir"
diff --git a/covid-mts/deps.sh b/covid-mts/deps.sh
new file mode 100755
index 00000000..4d423348
--- /dev/null
+++ b/covid-mts/deps.sh
@@ -0,0 +1,14 @@
+#!/bin/bash
+
+sudo apt update 
+
+sudo apt install -y \
+    sudo \
+    curl \
+    wget \
+    unzip \
+    python3-pip \
+    vim \
+    libarchive-tools libncurses5-dev libncursesw5-dev zstd liblzma-dev libbz2-dev zip unzip \
+    git
+
diff --git a/covid-mts/hashes/outputs.md5sum b/covid-mts/hashes/outputs.md5sum
new file mode 100644
index 00000000..a5d5e642
--- /dev/null
+++ b/covid-mts/hashes/outputs.md5sum
@@ -0,0 +1,4 @@
+49b36b9a7531dc79e0d9040ceced1461  outputs/1.out
+2596cd73ea1406ae863d853899957def  outputs/2.out
+6a6df2aec7e1ba347f0c7f5b39fbdc6b  outputs/3.out
+e1612db0634d0eff467c0e6dcf287755  outputs/4.out
diff --git a/covid-mts/hashes/outputs_small.md5sum b/covid-mts/hashes/outputs_small.md5sum
new file mode 100644
index 00000000..d227e44c
--- /dev/null
+++ b/covid-mts/hashes/outputs_small.md5sum
@@ -0,0 +1,4 @@
+7f8dd6c33d3a13de844bcdef61ecce4d  outputs_small/1.out
+011d767cb5741eab175c31462ff63b54  outputs_small/2.out
+645c86dfc08a2233edf50613327adaa3  outputs_small/3.out
+79fec1cbbb6e608413ffce247120d859  outputs_small/4.out
diff --git a/covid-mts/hashes/small/1.hash b/covid-mts/hashes/small/1.hash
deleted file mode 100644
index 9b74f4e6..00000000
--- a/covid-mts/hashes/small/1.hash
+++ /dev/null
@@ -1 +0,0 @@
-e2bce195c4833af9f5be080efa1be87436e6c1597cb4a1a4e63775b3bfbb9e6d
diff --git a/covid-mts/hashes/small/2.hash b/covid-mts/hashes/small/2.hash
deleted file mode 100644
index 4b47e1c7..00000000
--- a/covid-mts/hashes/small/2.hash
+++ /dev/null
@@ -1 +0,0 @@
-c0b74e21ec4202a01dd19d6e7a656d0ee39eee3f1ec772db6951a6ad0a079aed
diff --git a/covid-mts/hashes/small/3.hash b/covid-mts/hashes/small/3.hash
deleted file mode 100644
index 693f040b..00000000
--- a/covid-mts/hashes/small/3.hash
+++ /dev/null
@@ -1 +0,0 @@
-e518c50210e67c19440a5300ed819061a6b791008db5f01ac35abe9f1d950dff
diff --git a/covid-mts/hashes/small/4.hash b/covid-mts/hashes/small/4.hash
deleted file mode 100644
index 83ee9473..00000000
--- a/covid-mts/hashes/small/4.hash
+++ /dev/null
@@ -1 +0,0 @@
-a278628a3d911043466beb9d2eb49b0d194273ddf1a0b194bd5e56960b877b0c
diff --git a/covid-mts/input.sh b/covid-mts/input.sh
new file mode 100755
index 00000000..e24dfadd
--- /dev/null
+++ b/covid-mts/input.sh
@@ -0,0 +1,11 @@
+#!/bin/bash
+
+REPO_TOP=$(git rev-parse --show-toplevel)
+eval_dir="${REPO_TOP}/covid-mts"
+input_dir="${eval_dir}/input"
+
+mkdir -p "$input_dir"
+
+curl --insecure 'https://atlas-group.cs.brown.edu/data/covid-mts/in.csv.gz' | gunzip > "$input_dir/in.csv"
+
+curl --insecure 'https://atlas-group.cs.brown.edu/data/covid-mts/in_small.csv.gz' | gunzip > "$input_dir/in_small.csv"
diff --git a/covid-mts/input/.gitignore b/covid-mts/input/.gitignore
new file mode 100644
index 00000000..8b137891
--- /dev/null
+++ b/covid-mts/input/.gitignore
@@ -0,0 +1 @@
+
diff --git a/covid-mts/inputs.sh b/covid-mts/inputs.sh
deleted file mode 100755
index a40333e9..00000000
--- a/covid-mts/inputs.sh
+++ /dev/null
@@ -1,15 +0,0 @@
-#!/bin/bash
-
-cd "$(realpath $(dirname "$0"))"
-mkdir -p inputs
-cd inputs
-
-if [ ! -f ./in.csv ]; then
-    curl -f 'https://atlas-group.cs.brown.edu/data/covid-mts/in.csv.gz'> in.csv.gz
-    gzip -d in.csv.gz
-fi
-
-if [ ! -f ./in_small.csv ]; then
-    curl -f 'https://atlas-group.cs.brown.edu/data/covid-mts/in_small.csv.gz' > in_small.csv.gz
-    gzip -d in_small.csv.gz
-fi
diff --git a/covid-mts/run.sh b/covid-mts/run.sh
index d85ad017..d60c69ac 100755
--- a/covid-mts/run.sh
+++ b/covid-mts/run.sh
@@ -1,47 +1,22 @@
 #!/bin/bash
 
-export SUITE_DIR=$(realpath $(dirname "$0"))
-export TIMEFORMAT=%R
-cd $SUITE_DIR
-
-if [[ "$1" == "--small" ]]; then
-    echo "Using small input"
-    input_file="$SUITE_DIR/inputs/in_small.csv"
-else
-    echo "Using default input"
-    input_file="$SUITE_DIR/inputs/in.csv"
+REPO_TOP=$(git rev-parse --show-toplevel)
+eval_dir="${REPO_TOP}/covid-mts"
+input_dir="${eval_dir}/input"
+outputs_dir="${eval_dir}/outputs"
+scripts_dir="${eval_dir}/scripts"
+
+suffix=""
+if [[ "$@" == *"--small"* ]]; then
+    suffix="_small"
 fi
 
-mkdir -p "outputs"
-all_res_file="./outputs/covid-mts.res"
-> $all_res_file
-
-# time_file stores the time taken for each script
-# mode_res_file stores the time taken and the script name for every script in a mode (e.g. bash, pash, dish, fish)
-# all_res_file stores the time taken for each script for every script run, making it easy to copy and paste into the spreadsheet
-covid-mts() {
-    mkdir -p "outputs/$1"
-    mode_res_file="./outputs/$1/covid-mts.res"
-    > $mode_res_file
-
-    echo executing covid-mts $1 $(date) | tee -a $mode_res_file $all_res_file
-
-    for number in `seq 4` ## initial: FIXME 5.sh is not working yet
-    do
-        script="${number}"
-        script_file="./scripts/$script.sh"
-        output_dir="./outputs/$1/$script/"
-        output_file="./outputs/$1/$script.out"
-        time_file="./outputs/$1/$script.time"
-        log_file="./outputs/$1/$script.log"
-
-        if [[ "$1" == "bash" ]]; then
-            (time bash $script_file $input_file > $output_file ) 2> $time_file
-        fi
+input_file="$input_dir/in$suffix.csv"
+output_scoped="$outputs_dir/outputs$suffix"
+mkdir -p "$output_scoped"
 
-        cat "${time_file}" >> $all_res_file
-        echo "$script_file $(cat "$time_file")" | tee -a $mode_res_file
-    done
-}
+"$scripts_dir/1.sh" "$input_file" > "$output_scoped/1.out"
+"$scripts_dir/2.sh" "$input_file" > "$output_scoped/2.out"
+"$scripts_dir/3.sh" "$input_file" > "$output_scoped/3.out"
+"$scripts_dir/4.sh" "$input_file" > "$output_scoped/4.out"
 
-covid-mts "bash"
diff --git a/covid-mts/verify.sh b/covid-mts/verify.sh
index 09c8e391..83521719 100755
--- a/covid-mts/verify.sh
+++ b/covid-mts/verify.sh
@@ -1,65 +1,22 @@
 #!/bin/bash
 
-# Exit immediately if a command exits with a non-zero status
-# set -e
-
-cd "$(realpath $(dirname "$0"))"
-
-mkdir -p hashes/small
+REPO_TOP=$(git rev-parse --show-toplevel)
+eval_dir="${REPO_TOP}/covid-mts"
+outputs_dir="${eval_dir}/outputs"
+scripts_dir="${eval_dir}/scripts"
+hashes_dir="${eval_dir}/hashes"
 
+suffix=""
 if [[ "$@" == *"--small"* ]]; then
-    hash_folder="hashes/small"
-else
-    hash_folder="hashes"
+    suffix="_small"
 fi
 
 if [[ "$@" == *"--generate"* ]]; then
-    # Directory to iterate over
-    directory="outputs/bash"
-
-    # Loop through all .out files in the directory
-    for file in "$directory"/*.out
-    do
-        # Extract the filename without the directory path and extension
-        filename=$(basename "$file" .out)
-
-        # Generate SHA-256 hash
-        hash=$(shasum -a 256 "$file" | awk '{ print $1 }')
-
-        # Save the hash to a file
-        echo "$hash" > "$hash_folder/$filename.hash"
-
-        # Print the filename and hash
-        echo "File: $hash_folder/$filename.hash | SHA-256 Hash: $hash"
-    done
+    # give relative paths to md5sum
+    (cd "$outputs_dir"; md5sum "outputs$suffix"/* > "$hashes_dir/outputs$suffix.md5sum")
+    exit 0
 fi
 
-# Loop through all directories in the parent directory
-for folder in "outputs"/*/
-do
-    # Remove trailing slash
-    folder=${folder%/}
-
-    echo "Verifying folder: $folder"
-
-    # Loop through all .out files in the current directory
-    for file in "$folder"/*.out
-    do
-        # Extract the filename without the directory path and extension
-        filename=$(basename "$file" .out)
-
-        if [ ! -f "$folder/$filename.hash" ]; then
-            # Generate SHA-256 hash
-            hash=$(shasum -a 256 "$file" | awk '{ print $1 }')
-
-            # Save the hash to a file
-            echo "$hash" > "$folder/$filename.hash"
-        fi
-
-        # Compare the hash with the hash in the hashes directory
-        diff "$hash_folder/$filename.hash" "$folder/$filename.hash"
-
-        # Print the filename and hash
-        echo "File: $folder/$filename | SHA-256 Hash: $(cat "$folder/$filename.hash")"
-    done
-done
+# give relative paths to md5sum
+(cd "$outputs_dir"; md5sum --check --quiet --status "$hashes_dir/outputs$suffix.md5sum")
+echo covid-mts$suffix $?