accept-new-implementations.sh

#!/bin/bash

googledownload="Implementations (Responses) - Form Responses 1.tsv"
tsv=implementations.tsv

# Make sure there's exactly one argument and that it's an integer:
if [ $# -eq 1 ] && [ "$1" -eq "$1" ] 2>/dev/null; then
    num_new=$1
    echo Accepting $num_new implementations from "$googledownload"
else
    echo Usage:
    echo $0 n
    echo where \"n\" is the number of new implementations to take from
    echo the Google TSV download into the site\'s official source list
    echo of implementations.
    exit -1
fi


# First confirm assumptions about fields
# echo Confirming assumptions about fields in $tsv

expectedFieldsFile=implementations-google-column-names.tsv

head -1 "$googledownload" | diff -w $expectedFieldsFile -

if [ ! $? -eq 0 ] ; then
    echo Oh no\!  Field names were not as expected by $0 - diff above
    echo exiting
    exit -2
fi


# Make sure the file has at least enough lines for the header row plus $num_new
numlines=`wc -l "$googledownload" | awk '{print $1}'`

if (( $numlines < 1 + $num_new )); then
    echo Downloaded TSV file \"$googledownload\"
    echo has only $numlines line\(s\) - not enough for a header row plus $num_new new implementations
    exit -3;
fi

# The heart of it:
# Remove fields including timestamp, email...
# Append to end of implementations
# `cut` is adding weird ^M newlines, hence the calls to `tr`
tail -$num_new "$googledownload" | cut -f 1-3,5-21,23-24 | tr '\015' '\012' | tr -s '\012' >> $tsv


# echo check for duplicates.  Field 4 is \"Name\"
duplicates=`cut -f 4 $tsv | sort | uniq -d`

if [ ! -z "$duplicates" ] ; then
    echo Uh oh - there now appear to be multiple rows with the same name:
    echo "$duplicates"
    echo For each you probably want to choose the better one or merge the two.
    echo If you regret having accepted these new implementations you could undo with:
    echo "      " git checkout implementations.tsv
fi


# Check if any of the submitters are new site contributors and if so
# add them to the list.

contributors=contributors.txt
new=/tmp/newcontribs

# Here's what's happening:

# We want all names alphabetical regardless of whether they have
# websites listed, so special kludge: sort on the second character
# (to ignore the "[" for links) and put a space before the
# non-website ones (so the name starts on the second character of the
# line).

# Breaking down this huge chain of pipes:
#  Only the desired number | just contributor's name and possible website |
#  XXX temp kludge involving Jeremy |
#  convert to markdown (kludgily handing missing websites) |
#  sort | uniq 

tail -$num_new $tsv | cut -f 2,3 |\
    fgrep -v "Jeremy Wagner" |\
    awk  -F'\t' '{if (length($2)==0) print "", $1; else printf("[%s](%s)\n",$1,$2);}' | \
    sort -k 1.2 | uniq > $new

newbies=`comm -13 $contributors $new`

# XXX need to better handle contributors without a website
if [ ! -z "$newbies" ] ; then
    echo Adding the following new contributors to $contributors:
    echo "    $newbies"

    # Sort in the new ones, re-using $new
    rm -f $new
    echo "$newbies" | cat - $contributors | sort -k 1.2  | uniq > $new
    cp -f $new $contributors
fi
rm -f $new 


echo XXX Need to download any images...

exit 0