poem-rater.sh

#!/bin/bash
set +o posix

JSON="foo.json"
CORPUS="/home/gwern/wiki/docs/ai/2019-03-06-gpt2-poetry-1000samples.txt"

encode() {
    TMP_FILE=$(mktemp /tmp/XXXX.txt)
    echo "$@" >> $TMP_FILE
    ENCODED=$(PYTHONPATH=src python encode2.py --model_name 2019-03-06-gwern-gpt2-poetry-projectgutenberg-network519407  $TMP_FILE)
    echo "$ENCODED"; }
generateJson() {
    echo "{\"query\": [], \"sample0\": $2, \"sample1\": $3, \"best\": $1}," >> $JSON; }


rm -rf /tmp/poem-samples/; mkdir /tmp/poem-samples/
cat "$CORPUS"  | fgrep -v '=' | split --lines=60 - /tmp/poem-samples/sample-

echo "[" >> $JSON
for POEM in `ls /tmp/poem-samples/sample-*`; do

    FIRST=$(head -30 $POEM)
    FIRST_ENCODED=$(encode "$FIRST")
    SECOND=$(tail -30 $POEM)
    SECOND_ENCODED=$(encode "$SECOND")


    echo "$FIRST"
    echo "============================================="
    echo "$SECOND"
    echo "" # print a newline to make output easier to read and divide from the foregoing

    echo "1: First wins| 2: Second wins"
    read -N 1 RATING
    case "$RATING" in

        $'\n')
        # skip
        ;;

        1)
            generateJson 0 "$FIRST_ENCODED" "$SECOND_ENCODED"
            ;;
        2)
            generateJson 1 "$FIRST_ENCODED" "$SECOND_ENCODED"
            ;;
    esac
done
echo "]" >> $JSON