#!/bin/bash # Validate External Links by Iritscen # # Validates a list of external links in CSV format. The resulting logs are produced in three formats: # - TXT (for easy diffing with an earlier log) # - RTF (for reading as a local file with clickable links) # - HTML (for reading as a web page) # Call script with "--help" argument for documentation. Also see Read Me.rtf for critical notes. # # Recommended rule: # |----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----- ----| # # Table of contents (sections of script in order of appearance, not execution): # • Globals # • Help Output # • Setup # • Utility Functions # • Summary Output # • Initialization # • Data Sourcing # • Config Output # • Legend Output # • Main Loop # Set separator token to newline IFS=" " ### GLOBALS ### # Settings -- these will be changed from their defaults by the arguments passed in to the script LINKS_URL="" # download external link CSV from this location (can use "file://" protocol) EXCEPT_URL="" # location of wiki page with a list of exceptions for NG results OUTPUT_DIR="" # place reports and all other output in a folder inside this existing folder RECORD_OK_LINKS=0 # record response code to the log even when it's a value in OK_CODES SHOW_SLASH=0 # record issue when a slash is added to the end of a URL SHOW_HTTPS=0 # record issue when "http" is upgraded to "https" SHOW_YT_RD=0 # record redirection for a youtu.be URL expanding to the full URL SUGGEST_SNAPSHOTS_NG=0 # query the Internet Archive for a possible snapshot URL for each NG page SUGGEST_SNAPSHOTS_OK=0 # query the Internet Archive for an existing snapshot of each OK page CHECK_ARCHIVE_LINKS=0 # check URLs on archive.org and archive.is TAKE_PAGE_SHOT=0 # take a screenshot of each OK page TIMEOUT=10 # time to wait for a response when querying a site CHROME_PATH="" # path to a copy of Google Chrome that has the command-line screenshot feature URL_START=1 # start at this URL in LINKS_FILE URL_LIMIT=0 # if non-zero, stop at this URL in LINKS_FILE UPLOAD_INFO="" # path to a file on your hard drive with the login info needed to upload a report # Fixed strings -- see the occurrences of these variables to learn their purpose AGENT="Mozilla/5.0 (Macintosh; Intel Mac OS X 10_15_7) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/103.0.5060.134 Safari/537.36" ARCHIVE_API="http://archive.org/wayback/available" ARCHIVE_GENERIC="https://web.archive.org/web/*" ARCHIVE_OK_CODES="statuscodes=200&statuscodes=203&statuscodes=206" CHROME_SCREENSHOT="screenshot.png" EXCEPT_FILE_NAME="exceptions.txt" EXPECT_SCRIPT_NAME="val_expect_sftp.txt" WIKI_CURL="https://wiki.oni2.net/Validate_External_Links/Curl_codes" WIKI_HTTP="https://wiki.oni2.net/Validate_External_Links/HTTP_codes" WIKI_MAIN="https://wiki.oni2.net/Validate_External_Links" WIKI_ME="http://iritscen.oni2.net" THIS_DIR=$(cd $(dirname $0); pwd) WORKING_DIR=$(pwd) WIKI_PATH="wiki.oni2.net" # These are parallel arrays of the IDs and names of OniGalore's current namespaces declare -a NS_IDS=(-2 -1 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 100 101 102 103 104 105 108 109 110 111) declare -a NS_NAMES=("Media" "Special" "Main" "Talk" "User" "User_talk" "OniGalore" "OniGalore_talk" "File" "File_talk" "MediaWiki" "MediaWiki_talk" "Template" "Template_talk" "Help" "Help_talk" "Category" "Category_talk" "BSL" "BSL_talk" "OBD" "OBD_talk" "AE" "AE_talk" "Oni2" "Oni2_talk" "XML" "XML_talk") # These arrays tell the script which suffixes at the ends of URLs represent files and which are pages. # This determines whether the script tries to take a screenshot of the URL or just gets its HTTP code. declare -a HTTP_FILES=(3ds 7z avi BINA blend bsl cpp csv dae dll dmg doc east exe fbx first63 flv gamedata gif jpg last32 log m4a mhm mov mp3 mp4 oni ONWC pdf png psd py rar tga TRMA txt vbs wav wmv xaf xcf xlsx xml zip) declare -a HTTP_TLDS_AND_PAGES=(abstract action ars asp aspx cfm cgi com css de do full htm html it js jsp net org pgi php php3 phtml pl ru shtml stm uk x) # These arrays tells us which HTTP response codes are OK (good), which are RD (redirections), and which # are NG (no good). Pages that return OK codes will be screenshotted. Remember to update http_codes.txt # if you add a new code. declare -a OK_CODES=(200 401 405 406 418 501) declare -a RD_CODES=(301 302 303 307 308) declare -a NG_CODES=(000 400 403 404 410 429 500 502 503 504 530) # Characters not allowed in a URL. Curly braces are sometimes used on the wiki to build a link using # transcluded text, and if the transclusion fails, then the braces show up in the URL ILLEGAL_CHARS="{ }" # The shortest URL possible, used for sanity-checking some URLs: http://a.co MIN_URL_LENGTH=11 # These are parallel arrays giving the prefixes that can be used in place of normal external links to # some wikis and other sites; based on https://wiki.oni2.net/Special:Interwiki declare -a INTERWIKI_PREFIXES=(commons metawikimedia mw wikibooks wikidata wikimedia wikinews wikiquote wikisource wikispecies wikiversity wikivoyage wikt wp) declare -a INTERWIKI_DOMAINS=(commons.wikimedia.org meta.wikimedia.org mediawiki.org wikibooks.org wikidata.org wikimediafoundation.org wikinews.org wikiquote.org wikisource.org species.wikimedia.org wikiversity.org wikivoyage.org wiktionary.org wikipedia.org) # Variables for keeping track of main loop progress and findings LINK_NUM=0 EI_LINKS=0 IW_LINKS=0 OK_LINKS=0 RD_LINKS=0 NG_LINKS=0 SKIP_UNK_NS=0 SKIP_JS_PAGE=0 SKIP_BAD_URL=0 SKIP_NON_ASCII=0 SKIP_UNK_SUFFIX=0 SKIP_UNK_CODE=0 SKIP_EXPECT_NG=0 SKIP_EXPECT_RD=0 SKIP_EXPECT_EI=0 SKIP_EXPECT_IW=0 SKIP_HTTPS_UP=0 SKIP_SLASH_ADD=0 SKIP_YOUTU_BE=0 SKIP_ARCHIVES=0 FILE_LINKS=0 PAGE_LINKS=0 SKIPPED_HEADER_ROW=0 FINISHED_LIST="no" START_RUN=0 END_RUN=0 ### HELP OUTPUT ### # A pseudo-man page. Here is the 80-character rule for the page text: # 234567890123456789012345678901234567890123456789012345678901234567890123456789 function printHelp() { cat << EOF NAME Validate External Links SYNOPSIS validate_external_links.sh --help validate_external_links.sh --links URL --output DIR [--exceptions URL] [--record-ok-links] [--show-added-slashes] [--show-https-upgrades] [--show-yt-redirects] [--suggest-snapshots] [--check-archive-links] [--take-screenshots FILE] [--timeout NUM] [--start-url NUM] [--end-url NUM] [--upload FILE] DESCRIPTION This script parses a list of external links found in the OniGalore wiki (which is dumped by the Oni2.net server periodically in a particular format), validates them using the Unix tool 'curl', and produces a report of which links were "OK" (responded positively to an HTTP query), which were "RD" (responded with a 3xx redirect code), which could be "IW" (interwiki) links, which are "EI" (external internal) links and could be intrawiki links, and which were "NG" (no good; a negative response to the query). This report can then be automatically uploaded to the location of your choice. The script can also suggest Internet Archive snapshots for "NG" links, and take screenshots of "OK" links for visual verification by the reader that the page in question is the one intended to be displayed. You must pass this script the URL at which the list of links is found (--links) and the path where the directory of logs should be outputted (--output). All other arguments are optional. OPTIONS --help Show this page. --links URL (required) URL from which to download the CSV file with external links. Note that this URL can be a local file if you supply a file:// path. --output DIR (required) Unix path to directory in which Val should place its reports. --exceptions URL In order to remove links from the report which Val finds an issue with but which you regard as OK, list those desired exceptions on a wiki page. See the sample file "exceptions.pdf" for the required format of the page. Note that this URL can point to a local file if you supply a path beginning with "file://". --record-ok-links Log a link in the report even if its response code is "OK". --show-added-slashes Report on redirects that simply add a '/' to the end of the URL. --show-https-upgrades Report on redirects that simply upgrade a "http://" URL to a "https://" URL. --show-yt-redirects Report on redirects that expand a youtu.be URL. --suggest-snapshots-ng Query the Internet Archive for a possible snapshot URL for each "NG" page. --suggest-snapshots-ok Query the Internet Archive for a snapshot of each "OK" page just to make sure it's available. Note that this will add a tremendous amount of time to the script execution because there is a rate limit to the Archive API. Note that this option does nothing unless you also use the --record-ok-links argument. --check-archive-links Check links that are already pointing to a page on the Internet Archive or archive.is (AKA archive.today). In theory these links should be totally stable and not need validation. --take-screenshots FILE Call the Google Chrome binary at this path to take screenshots of each "OK" page. --timeout NUM Wait this many seconds for a site to respond. The default is 10. Important note: Val will attempt to reach each URL three times, so the time taken to ping an unresponsive site will be three times this setting. --start-url NUM Start at this link in the links CSV file. --end-url NUM Stop at this link in the links CSV file. --upload FILE Upload report using the credentials and path given in this local text file. See sftp_login.txt for template. BUGS The script cannot properly parse any line in the external links file which contains a comma in the name of the wiki page containing a link. Commas in the link itself are not an issue. EOF } ### SETUP ### # If first argument is a help request, or if nothing was passed in at all, print help page and quit if [ "$#" -eq 0 ] || [ "$1" == "--help" ]; then printHelp | less exit 0 fi # Parse arguments as long as there are more arguments to process while (( "$#" )); do case "$1" in --links ) LINKS_URL="$2"; shift 2;; --exceptions ) EXCEPT_URL="$2"; shift 2;; --output ) OUTPUT_DIR="$2"; shift 2;; --record-ok-links ) RECORD_OK_LINKS=1; shift;; --show-added-slashes ) SHOW_SLASH=1; shift;; --show-https-upgrades ) SHOW_HTTPS=1; shift;; --show-yt-redirects ) SHOW_YT_RD=1; shift;; --suggest-snapshots-ng ) SUGGEST_SNAPSHOTS_NG=1; shift;; --suggest-snapshots-ok ) SUGGEST_SNAPSHOTS_OK=1; shift;; --check-archive-links ) CHECK_ARCHIVE_LINKS=1; shift;; --take-screenshots ) TAKE_PAGE_SHOT=1; CHROME_PATH="$2"; shift 2;; --timeout ) TIMEOUT=$2; shift 2;; --start-url ) URL_START=$2; shift 2;; --end-url ) URL_LIMIT=$2; shift 2;; --upload ) UPLOAD_INFO=$2; shift 2;; * ) echo "Invalid argument '$1' detected. Aborting."; exit 1;; esac done # If the required arguments were not supplied, print help page and quit if [ -z $LINKS_URL ] || [ -z $OUTPUT_DIR ]; then echo "Error: I did not receive one or both required arguments. Run me with the \"--help\" argument for documentation." exit 2 fi # If user wants screenshots, make sure path to Chrome was passed in and is valid if [ $TAKE_PAGE_SHOT -eq 1 ]; then if [ ! -f "$CHROME_PATH" ]; then echo "Error: You need to supply a path to the Google Chrome application in order to take screenshots." exit 3 fi fi # Check that UPLOAD_INFO exists, if this argument was supplied if [ ! -z $UPLOAD_INFO ] && [ ! -f "$UPLOAD_INFO" ]; then echo "Error: The file $UPLOAD_INFO supplied by the --upload argument does not appear to exist. Aborting." exit 4 fi # Check that OUTPUT_DIR is a directory if [ ! -d "$OUTPUT_DIR" ]; then echo "Error: The path $OUTPUT_DIR supplied by the --output argument does not appear to be a directory. Aborting." exit 5 fi # Make timestamped folder inside OUTPUT_DIR for this session's log and screenshots SAFE_TIME=$(date "+%Y-%m-%d--%H-%M-%S") # for use in path names NICE_TIME=$(TZ=":UTC" date "+%a, %d %b %Y %k:%M:%S GMT") # matches format of LINKS_DATE, which is extracted from the extlinks file OUTPUT_FOLDER="ValExtLinks ($SAFE_TIME)" OUTPUT_PATH="$OUTPUT_DIR/$OUTPUT_FOLDER" SHOT_PATH="$OUTPUT_PATH/Screenshots" LOG_NAME="ValExtLinks report" LOG_NAME_TXT="$LOG_NAME.txt" LOG_NAME_RTF="$LOG_NAME.rtf" LOG_NAME_HTM="$LOG_NAME.htm" LOG_PATH="$OUTPUT_PATH/$LOG_NAME" LOG_PATH_TXT="$LOG_PATH.txt" LOG_PATH_RTF="$LOG_PATH.rtf" LOG_PATH_HTM="$LOG_PATH.htm" mkdir "$OUTPUT_PATH" if [ $TAKE_PAGE_SHOT -eq 1 ]; then mkdir "$SHOT_PATH" fi # Check that 'mkdir' succeeded if [ ! -d "$OUTPUT_PATH" ]; then echo "Error: I could not create the folder \"$OUTPUT_FOLDER\" inside the directory $OUTPUT_PATH. Aborting." exit 6 fi # Get date on the file at LINKS_URL and print to log LINKS_DATE=$(curl --silent --head $LINKS_URL | grep "Last-Modified") if [ -z "$LINKS_DATE" ]; then echo "Error: I could not find the external links file at the path \"$LINKS_URL\" supplied by the --links argument. Aborting." exit 7 fi LINKS_DATE=${LINKS_DATE#Last-Modified: } ### UTILITY FUNCTIONS ### # Writes a plain-text header to TXT log file function printTXTheader() { valPrint t "Validate External Links report" valPrint t "generated $NICE_TIME" valPrint t "from data of $LINKS_DATE" valPrint t "script by Iritscen (contact: $WIKI_ME)" valPrint t "" } # Writes the RTF header to RTF log file, then a centered title, then sets text to left-justified function printRTFheader() { valPrint r "{\rtf1\ansi\ansicpg1252\cocoartf1504\cocoasubrtf820 {\fonttbl\f0\fswiss\fcharset0 Helvetica;} {\colortbl;\red255\green255\blue255;} {\*\expandedcolortbl;;} \margl1440\margr1440\vieww12600\viewh12100\viewkind0 \pard\tx720\tx1440\tx2160\tx2880\tx3600\tx4320\tx5040\tx5760\tx6480\tx7200\tx7920\tx8640\li360\pardirnatural\qc\partightenfactor0 \f0\fs28 \cf0 \b1 Validate External Links report \b0\\ generated $NICE_TIME\\ from data of $LINKS_DATE\\ script by Iritscen ({\field{\*\fldinst{HYPERLINK \"$WIKI_ME\"}}{\fldrslt contact}}) \\ \pard\tx720\tx1440\tx2160\tx2880\tx3600\tx4320\tx5040\tx5760\tx6480\tx7200\tx7920\tx8640\li360\pardirnatural\partightenfactor0 \cf0 " } # Closes the RTF markup of the RTF log file function printRTFfooter() { valPrint r "}" } # Writes the HTML header to HTML log file function printHTMheader() { valPrint h " Validate External Links report

Validate External Links report

generated $NICE_TIME
from data of $LINKS_DATE
script by Iritscen (contact)

" } # Closes the HTML markup of the HTML log file function printHTMfooter() { valPrint h " " } # The central logging function. The first parameter is a string composed of one or more characters that # indicate which output to use: 'c' means console, 't' means the TXT log, 'r' means the RTF log, and # 'h' means the HTML log. 'n' means "Don't print a newline at the end of the line." 's' means "Print an # extra newline at the end." 'w' means "Don't pass console output through 'fmt'" ("fmt" fits the output # to an 80-column CLI but can break special formatting and the 'n' option). function valPrint() { if [[ "$1" == *c* ]]; then if [[ "$1" == *n* ]]; then echo -n "$2" elif [[ "$1" == *w* ]]; then echo "$2" elif [[ "$1" == *s* ]]; then echo -e "$2\n" else echo "$2" | fmt -w 80 fi fi if [[ "$1" == *t* ]]; then if [[ "$1" == *n* ]]; then echo -n "$2" >> "$LOG_PATH_TXT" elif [[ "$1" == *s* ]]; then echo -e "$2\n" >> "$LOG_PATH_TXT" else echo "$2" >> "$LOG_PATH_TXT" fi fi if [[ "$1" == *r* ]]; then if [[ "$1" == *n* ]]; then echo "$2" >> "$LOG_PATH_RTF" elif [[ "$1" == *s* ]]; then echo "$2\line\line" >> "$LOG_PATH_RTF" else echo "$2\line" >> "$LOG_PATH_RTF" fi fi if [[ "$1" == *h* ]]; then if [[ "$1" == *s* ]]; then echo "$2 " >> "$LOG_PATH_HTM" elif [[ "$1" == *n* ]]; then echo "$2" >> "$LOG_PATH_HTM" else echo "$2
" >> "$LOG_PATH_HTM" fi fi } # Pluralize the string in parameter 1 if the number in parameter 2 is not 1 function pluralCheckNoun() { if [ $2 -ne 1 ]; then if [[ $1 =~ x$ ]]; then echo $1es else echo $1s fi else echo $1 fi } # Output "is" if parameter 1 is 1, otherwise "are" function pluralCheckIs() { if [ $1 -ne 1 ]; then echo "are" else echo "is" fi } # Output "was" if parameter 1 is 1, otherwise "were" function pluralCheckWas() { if [ $1 -ne 1 ]; then echo "were" else echo "was" fi } # Output "a " if parameter 1 is 1, otherwise nothing function pluralCheckA() { if [ $1 -eq 1 ]; then echo "a " fi } # Output "an " if parameter 1 is 1, otherwise nothing function pluralCheckAn() { if [ $1 -eq 1 ]; then echo "an " fi } # Upload the reports using info specified in the --upload argument. ONLY USE "valPrint c" here, as the # reports being saved to disk have already been closed. function uploadReport() { valPrint c "Uploading reports..." SCRIPT_PATH="$THIS_DIR/$EXPECT_SCRIPT_NAME" SFTP_USER_NAME_MARKER="user:" SFTP_PASSWORD_MARKER="pw:" SFTP_PORT_MARKER="port:" SFTP_PATH_MARKER="path:" SFTP_USER_NAME=$(grep $SFTP_USER_NAME_MARKER $UPLOAD_INFO) SFTP_USER_NAME=${SFTP_USER_NAME#$SFTP_USER_NAME_MARKER} SFTP_PASSWORD=$(grep $SFTP_PASSWORD_MARKER $UPLOAD_INFO) SFTP_PASSWORD=${SFTP_PASSWORD#$SFTP_PASSWORD_MARKER} SFTP_PORT=$(grep $SFTP_PORT_MARKER $UPLOAD_INFO) SFTP_PORT=${SFTP_PORT#$SFTP_PORT_MARKER} SFTP_PATH=$(grep $SFTP_PATH_MARKER $UPLOAD_INFO) SFTP_PATH=${SFTP_PATH#$SFTP_PATH_MARKER} for SUFFIX in htm rtf txt; do expect "$SCRIPT_PATH" "$LOG_PATH.$SUFFIX" $SFTP_USER_NAME $SFTP_PASSWORD $SFTP_PORT $SFTP_PATH "$LOG_NAME.$SUFFIX" if [ "$?" -ne 0 ]; then valPrint c "Error $? occurred when attempting to upload $LOG_NAME.$SUFFIX!" else valPrint c "Report in `echo $SUFFIX | tr [:lower:] [:upper:]` format was uploaded." fi done } # Prints session summary when script is done function wrapupAndExit() { # Get off progress line on console, drop down a line from last link in log, and close HTML table valPrint ctr "" valPrint h "
" # If we didn't finish processing the last URL, then the iterator is one too high if [ $FINISHED_LIST != "yes" ]; then let LINK_NUM-=1 if [ $FINISHED_LIST == "no" ]; then valPrint ctrh "The session was canceled by the user." fi fi # Generate string with elapsed time END_RUN=$(date +%s) ELAPSED=$(echo $(($END_RUN - $START_RUN)) | awk '{printf "%d min. %d sec. elapsed", int($1/60), int($1%60)}') # Do some math on results of session LINKS_PROCESSED=$((LINK_NUM-URL_START+1)) TRIVIAL_RDS=$((SKIP_SLASH_ADD+SKIP_HTTPS_UP+SKIP_YOUTU_BE)) LINK_ERRORS=$((SKIP_UNK_NS+SKIP_JS_PAGE+SKIP_BAD_URL+SKIP_NON_ASCII+SKIP_UNK_SUFFIX+SKIP_UNK_CODE)) LINKS_EXCEPTED=$((SKIP_EXPECT_NG+SKIP_EXPECT_RD+SKIP_EXPECT_EI+SKIP_EXPECT_IW)) LINK_PROBLEMS_TOTAL=$((NG_LINKS+RD_LINKS+EI_LINKS+IW_LINKS)) LINK_PROBLEMS_NG=$((NG_LINKS-SKIP_EXPECT_NG)) LINK_PROBLEMS_RD=$((RD_LINKS-SKIP_EXPECT_RD)) LINK_PROBLEMS_EI=$((EI_LINKS-SKIP_EXPECT_EI)) LINK_PROBLEMS_IW=$((IW_LINKS-SKIP_EXPECT_IW)) LINK_PROBLEMS_NET=$((LINK_PROBLEMS_NG+LINK_PROBLEMS_RD+LINK_PROBLEMS_EI+LINK_PROBLEMS_IW)) # Print something in the Links section if no link issues were printed if [ $LINK_PROBLEMS_NET -eq 0 ]; then valPrint h "No link problems to report! See the RTF or TXT report for a list of links with issues that were not reported." fi if [ $LINK_PROBLEMS_TOTAL -eq 0 ]; then valPrint t "No link problems to report!" valPrint r "\i1 No link problems to report! \i0" fi ## SUMMARY OUTPUT ## valPrint ct "Summary ($ELAPSED):" valPrint r "\b1 Summary \b0 ($ELAPSED)" valPrint hn "

Summary ($ELAPSED)

" valPrint ctrh "I finished processing $LINKS_PROCESSED of $LINK_COUNT $(pluralCheckNoun link $LINK_COUNT) (there $(pluralCheckWas $FILE_LINKS) $FILE_LINKS file $(pluralCheckNoun link $FILE_LINKS) and $PAGE_LINKS page $(pluralCheckNoun link $PAGE_LINKS))." # Print processed link totals if [ $LINKS_PROCESSED -gt 0 ]; then valPrint ctrh "$LINKS_PROCESSED processed $(pluralCheckNoun link $LINKS_PROCESSED):"; fi if [ $LINK_ERRORS -gt 0 ]; then valPrint ctrh "- $LINK_ERRORS $(pluralCheckNoun link $LINK_ERRORS) could not be processed"; fi if [ $SKIP_ARCHIVES -gt 0 ]; then valPrint ctrh "- $SKIP_ARCHIVES archive.org/archive.is $(pluralCheckNoun link $SKIP_ARCHIVES) were not checked"; fi if [ $LINK_PROBLEMS_TOTAL -gt 0 ]; then valPrint ctrh "- $LINK_PROBLEMS_TOTAL processed $(pluralCheckNoun link $LINK_PROBLEMS_TOTAL) had $(pluralCheckAn $LINK_PROBLEMS_TOTAL)$(pluralCheckNoun issue $LINK_PROBLEMS_TOTAL)"; fi if [ $LINKS_EXCEPTED -gt 0 ]; then valPrint ctr " (excepted $LINKS_EXCEPTED link $(pluralCheckNoun issue $LINKS_EXCEPTED) from report)"; valPrint h "  (excepted $LINKS_EXCEPTED $(pluralCheckNoun link $LINKS_EXCEPTED) from report)"; fi if [ $OK_LINKS -gt 0 ]; then valPrint ctrh "- $OK_LINKS processed $(pluralCheckNoun link $OK_LINKS) $(pluralCheckWas $OK_LINKS) OK"; fi if [ $TRIVIAL_RDS -gt 0 ]; then valPrint ctr " (counted $TRIVIAL_RDS trivial $(pluralCheckNoun redirection $TRIVIAL_RDS) as OK)"; valPrint h "  (counted $TRIVIAL_RDS trivial $(pluralCheckNoun redirection $TRIVIAL_RDS) as OK)"; fi # Print errored link totals if [ $LINK_ERRORS -gt 0 ]; then valPrint c "$LINK_ERRORS link $(pluralCheckNoun error $LINK_ERRORS) (see RTF or TXT report for specific links):" valPrint h "$LINK_ERRORS link $(pluralCheckNoun error $LINK_ERRORS) (see RTF or TXT report for specific links):" valPrint rt "$LINK_ERRORS link $(pluralCheckNoun error $LINK_ERRORS):" fi if [ $SKIP_UNK_NS -gt 0 ]; then valPrint ctrh "- $SKIP_UNK_NS missing/unknown $(pluralCheckNoun namespace $SKIP_UNK_NS)"; fi if [ $SKIP_JS_PAGE -gt 0 ]; then valPrint ctrh "- $SKIP_JS_PAGE $(pluralCheckNoun link $SKIP_JS_PAGE) on $(pluralCheckA $SKIP_JS_PAGE)JavaScript $(pluralCheckNoun page $SKIP_JS_PAGE)"; fi if [ $SKIP_BAD_URL -gt 0 ]; then valPrint ctrh "- $SKIP_BAD_URL illegal $(pluralCheckNoun URL $SKIP_BAD_URL)"; fi if [ $SKIP_NON_ASCII -gt 0 ]; then valPrint ctrh "- $SKIP_NON_ASCII non-ASCII $(pluralCheckNoun URL $SKIP_NON_ASCII)"; fi if [ $SKIP_UNK_SUFFIX -gt 0 ]; then valPrint ctrh "- $SKIP_UNK_SUFFIX unknown URL $(pluralCheckNoun suffix $SKIP_UNK_SUFFIX)"; fi if [ $SKIP_UNK_CODE -gt 0 ]; then valPrint ctrh "- $SKIP_UNK_CODE unknown response $(pluralCheckNoun code $SKIP_UNK_CODE)"; fi # Print excepted link totals if [ $LINKS_EXCEPTED -gt 0 ]; then valPrint c "$LINKS_EXCEPTED link $(pluralCheckNoun problem $LINKS_EXCEPTED) excepted (see RTF or TXT report for specific links):" valPrint h "$LINKS_EXCEPTED link $(pluralCheckNoun problem $LINKS_EXCEPTED) excepted (see RTF or TXT report for specific links):" valPrint rt "$LINKS_EXCEPTED link $(pluralCheckNoun problem $LINKS_EXCEPTED) excepted:" fi if [ $SKIP_EXPECT_NG -gt 0 ]; then valPrint ctrh "- $SKIP_EXPECT_NG/$NG_LINKS NG $(pluralCheckNoun link $NG_LINKS)"; fi if [ $SKIP_EXPECT_RD -gt 0 ]; then valPrint ctrh "- $SKIP_EXPECT_RD/$RD_LINKS $(pluralCheckNoun redirection $RD_LINKS)"; fi if [ $SKIP_EXPECT_EI -gt 0 ]; then valPrint ctrh "- $SKIP_EXPECT_EI/$EI_LINKS external internal $(pluralCheckNoun link $EI_LINKS)"; fi if [ $SKIP_EXPECT_IW -gt 0 ]; then valPrint ctrh "- $SKIP_EXPECT_IW/$IW_LINKS potential intrawiki $(pluralCheckNoun link $IW_LINKS)"; fi # Perform exceptions audit EXCEPTION_ISSUES=0 valPrint ctrh "Exceptions list audit:" for ((i = 0; i < ${#EXCEPT_ARRAY[@]}; ++i)); do EXCEPT_LINE="${EXCEPT_ARRAY[$i]}" EXCEPT_LINE=$(echo "$EXCEPT_LINE" | sed 's/\&/\&/g') # copied from exception-matching code if [ ${EXCEPT_FOUND[$i]} -eq 0 ]; then EXCEPT_URL="${EXCEPT_LINE#*,}" EXCEPT_URL="${EXCEPT_URL%,*}" EXCEPT_PAGE="${EXCEPT_LINE##*,}" EXCEPT_PAGE="${EXCEPT_PAGE%% *}" if [ "$EXCEPT_PAGE" == "*" ]; then valPrint tr "- The link '$EXCEPT_URL' did not return an error or is not on any page." else valPrint tr "- The link '$EXCEPT_URL' did not return an error or is not on page '$EXCEPT_PAGE'." fi let EXCEPTION_ISSUES+=1 elif [ ${EXCEPT_USED[$i]} -eq 0 ]; then EXCEPT_URL="${EXCEPT_LINE#*,}" EXCEPT_URL="${EXCEPT_URL%,*}" EXCEPT_CODE=${EXCEPT_LINE%%,*} valPrint tr "- The link '$EXCEPT_URL' did not return error code $EXCEPT_CODE." let EXCEPTION_ISSUES+=1 fi done if [ $EXCEPTION_ISSUES -eq 0 ]; then valPrint ctrh "- No issues found." else valPrint c "$EXCEPTION_ISSUES exception list $(pluralCheckNoun issue $EXCEPTION_ISSUES) detected (see RTF or TXT report for details)." valPrint h "$EXCEPTION_ISSUES exception list $(pluralCheckNoun issue $EXCEPTION_ISSUES) detected (see RTF or TXT report for details)." fi # Print checked link totals if [ $LINK_PROBLEMS_NET -gt 0 ]; then valPrint ctrh "$LINK_PROBLEMS_NET link $(pluralCheckNoun issue $LINK_PROBLEMS_NET):"; fi if [ $LINK_PROBLEMS_NG -gt 0 ]; then valPrint ctrh "- $LINK_PROBLEMS_NG NG $(pluralCheckNoun link $LINK_PROBLEMS_NG)"; fi if [ $LINK_PROBLEMS_RD -gt 0 ]; then valPrint ctrh "- $LINK_PROBLEMS_RD $(pluralCheckNoun redirection $LINK_PROBLEMS_RD)"; fi if [ $LINK_PROBLEMS_EI -gt 0 ]; then valPrint ctrh "- $LINK_PROBLEMS_EI $(pluralCheckNoun link $LINK_PROBLEMS_EI) that could be intrawiki"; fi if [ $LINK_PROBLEMS_IW -gt 0 ]; then valPrint ctrh "- $LINK_PROBLEMS_IW $(pluralCheckNoun link $LINK_PROBLEMS_IW) that could be interwiki"; fi # Close the log files' markup valPrint trh "ValExtLinks says goodbye." printRTFfooter printHTMfooter # Upload report if this was requested if [ ! -z $UPLOAD_INFO ] && [ $FINISHED_LIST != "no" ]; then uploadReport fi # Really quit now valPrint c "ValExtLinks says goodbye." exit 0 } trap wrapupAndExit INT ### INITIALIZATION ### # Print opening message to console and log files valPrint cw "$(tput bold)--Validate External Links--$(tput sgr0)" printTXTheader printRTFheader printHTMheader ## DATA SOURCING ## valPrint t "Startup:" valPrint r "\b1 Startup \b0" valPrint hn "

Startup

" # Attempt to download file at LINKS_URL, then check that it succeeded valPrint cwtrhn "Downloading list of external links from $LINKS_URL..." LINKS_FILE_NAME=$(echo "$LINKS_URL" | sed 's/.*\///') LINKS_FILE="$OUTPUT_PATH/$LINKS_FILE_NAME" curl --silent -o "$LINKS_FILE" $LINKS_URL if [ ! -f "$LINKS_FILE" ]; then echo -e "\nThe download of $LINKS_URL appears to have failed. Aborting." wrapupAndExit else valPrint ctrh " success." fi # Attempt to download file at EXCEPT_URL, then check that it succeeded if [ ! -z $EXCEPT_URL ]; then valPrint cwtrhn "Downloading list of reporting exceptions from $EXCEPT_URL..." EXCEPT_DATA=$(curl --silent $EXCEPT_URL) if [ -z "$EXCEPT_DATA" ]; then echo -e "\nThe download of the exceptions data from '$EXCEPT_URL' appears to have failed. Aborting." wrapupAndExit else valPrint ctrh " success." fi EXCEPT_DATA=${EXCEPT_DATA%END LIST*} EXCEPT_DATA=${EXCEPT_DATA#*BEGIN LIST} EXCEPT_FILE="$OUTPUT_PATH/$EXCEPT_FILE_NAME" # Store on disk for debugging purposes echo "$EXCEPT_DATA" > "$EXCEPT_FILE" # Transfer to array for easy searching later declare -a EXCEPT_ARRAY=($(echo "$EXCEPT_DATA")) # Create parallel arrays for marking which exceptions get used later declare -a EXCEPT_USED=() declare -a EXCEPT_FOUND=() for ((i = 0; i < ${#EXCEPT_ARRAY[@]}; ++i)); do EXCEPT_USED+=(0) EXCEPT_FOUND+=(0) done fi # Pipe 'cat' to 'wc' because passing LINKS_FILE to 'wc' would print the file name after the line count LINK_COUNT_STRING=$(cat "$LINKS_FILE" | wc -l) # Number of URLs is number of lines minus one (first line is column header row for the CSV) LINK_COUNT=$(echo "${LINK_COUNT_STRING}" | tr -d '[:space:]') let LINK_COUNT-=1 valPrint ctrh "Found $LINK_COUNT links to process." valPrint trh "" ## CONFIG OUTPUT ## valPrint t "Config:" valPrint r "\b1 Config \b0" valPrint hn "

Config

" valPrint ctrhn "Links to consider: " if [ $URL_LIMIT -ne 0 ] && [ $URL_LIMIT -lt $LINK_COUNT ]; then valPrint ctrh "$((URL_LIMIT-URL_START+1)), from link $URL_START to $URL_LIMIT" elif [ $URL_START -ne 1 ]; then valPrint ctrh "$((LINK_COUNT-URL_START+1)), from link $URL_START to $LINK_COUNT" else valPrint ctrh "$LINK_COUNT" fi valPrint ctrh "Site query timeout: $TIMEOUT seconds" valPrint ctrhn "Show OK links: " if [ $RECORD_OK_LINKS -eq 1 ]; then valPrint ctrh "Yes"; else valPrint ctrh "No"; fi valPrint ctrhn "Take screenshots: " if [ $TAKE_PAGE_SHOT -eq 1 ]; then valPrint ctrh "Yes"; else valPrint ctrh "No"; fi valPrint ctrhn "Suggest archive.org snapshots for NG pages: " if [ $SUGGEST_SNAPSHOTS_NG -eq 1 ]; then valPrint ctrh "Yes"; else valPrint ctrh "No"; fi valPrint ctrhn "Suggest archive.org snapshots for OK pages: " if [ $SUGGEST_SNAPSHOTS_OK -eq 1 ]; then valPrint ctrh "Yes"; else valPrint ctrh "No"; fi valPrint ctrhn "Ignore slash-adding redirects: " if [ $SHOW_SLASH -eq 1 ]; then valPrint ctrh "No"; else valPrint ctrh "Yes"; fi valPrint ctrhn "Ignore HTTPS-upgrading redirects: " if [ $SHOW_HTTPS -eq 1 ]; then valPrint ctrh "No"; else valPrint ctrh "Yes"; fi valPrint ctrhn "Ignore youtu.be redirects: " if [ $SHOW_YT_RD -eq 1 ]; then valPrint ctrh "No"; else valPrint ctrh "Yes"; fi valPrint ctrhn "Check archive.org and archive.is links: " if [ $CHECK_ARCHIVE_LINKS -eq 1 ]; then valPrint ctrh "Yes"; else valPrint ctrh "No"; fi valPrint tr "A summary of my findings will be found at the bottom of the report." valPrint h "A summary of my findings will be found at the bottom of the report." valPrint trh "" ## LEGEND OUTPUT ## valPrint t "Legend:" valPrint r "\b1 Legend \b0" valPrint hn "

Legend

" valPrint t "(For guidance in fixing these links, see $WIKI_MAIN. The exceptions list is at $EXCEPT_URL.)" valPrint r "(For guidance in fixing these links, see {\field{\*\fldinst{HYPERLINK \"$WIKI_MAIN\"}}{\fldrslt here}}. The exceptions list is {\field{\*\fldinst{HYPERLINK \"$EXCEPT_URL\"}}{\fldrslt here}}.)" valPrint h "(For guidance in fixing these links, see here. The exceptions list is here.)" valPrint trh "OK = URL seems to be working" valPrint trh "NG = URL no longer seems to work" valPrint trh "RD = URL is redirecting to this new URL" valPrint trh "EI = URL is an external link to an internal page and should be converted to an intrawiki link using the suggested markup" valPrint trh "IW = URL is an external link to a fellow wiki and should be converted to an interwiki link using the suggested markup" valPrint t "(xxx) = Unix tool 'curl' obtained this HTTP response status code (see here for code reference: $WIKI_HTTP)" valPrint r "(xxx) = Unix tool 'curl' obtained this HTTP response status code (see {\field{\*\fldinst{HYPERLINK \"$WIKI_HTTP\"}}{\fldrslt here}} for code reference)" valPrint h "(xxx) = Unix tool 'curl' obtained this HTTP response status code (see here for code reference)" valPrint t "(000-xx) = 'curl' did not get an HTTP response code, but returned this exit code (see here for code reference: $WIKI_CURL)" valPrint r "(000-xx) = 'curl' did not get an HTTP response code, but returned this exit code (see {\field{\*\fldinst{HYPERLINK \"$WIKI_CURL\"}}{\fldrslt here}} for code reference)" valPrint h "(000-xx) = 'curl' did not get an HTTP response code, but returned this exit code (see here for code reference)" valPrint trh "IA suggests = Last available snapshot returned by the Internet Archive" valPrint trh "Try browsing = The Archive failed to return a snapshot URL, so check for a snapshot manually using this link" valPrint trh "" ### MAIN LOOP ### valPrint t "Links:" valPrint r "\b1 Links \b0" valPrint hn "

Links

" START_RUN=$(date +%s) # Process each line of the .csv in LINKS_FILE for LINE in `cat "$LINKS_FILE"`; do START_LINK=$(date +%s) let LINK_NUM+=1 # First line is the column header row for the CSV, so let's verify that the format hasn't changed if [ $SKIPPED_HEADER_ROW -eq 0 ]; then if [ $LINE == "namespace,title,target" ]; then SKIPPED_HEADER_ROW=1 LINK_NUM=0 # this line is not a link, so reset the link counter valPrint hn "" continue else valPrint ctrh "Did not find expected header row in $LINKS_FILE. Aborting." wrapupAndExit fi fi # Skip this link if we are not at URL_START yet if [ $LINK_NUM -lt $URL_START ]; then continue fi # Stop if we are at the limit declared for testing purposes if [ $URL_LIMIT -gt 0 ] && [ $LINK_NUM -gt $URL_LIMIT ]; then FINISHED_LIST="limit" wrapupAndExit fi # Print progress to screen if [ $LINK_NUM -gt 1 ]; then printf "\e[1A\n" # erase previous progress message so that new one appears in its place fi valPrint cn "Evaluating URL $LINK_NUM/$LINK_COUNT..." # The number of the namespace is the element before the first comma on the line NS_ID=${LINE%%,*} # Find namespace number in NS_IDS and use it to look up namespace's name in NS_NAMES NS_NAME="" a=0 while [ "x${NS_IDS[$a]}" != "x" ]; do # once this evaluates to "x", the array is done if [ $NS_ID == "NULL" ]; then break elif [ $NS_ID -eq ${NS_IDS[$a]} ]; then NS_NAME="${NS_NAMES[$a]}" break fi let a+=1 done if [ "$NS_NAME" == "" ]; then if [ $NS_ID == "NULL" ]; then valPrint trs "Skipping line $LINK_NUM ('$LINE') because the namespace (and probably the page too) is 'NULL'. Probably the link is no longer in existence on the wiki." else valPrint trs "Skipping line $LINK_NUM ('$LINE') because I could not find a name for namespace ID $NS_ID." fi let SKIP_UNK_NS+=1 let PAGE_LINKS+=1 continue fi # The name of the page is everything between the namespace ID and the next comma on the line (commas # in page names will break this) PAGE_NAME=${LINE#$NS_ID,} PAGE_NAME=${PAGE_NAME%%,*} # Build longer wiki page URLs from namespace and page names FULL_PAGE_PATH=https://$WIKI_PATH/$NS_NAME:$PAGE_NAME LOCAL_PAGE_PATH=$NS_NAME:$PAGE_NAME # Namespace "Main:" cannot be a part of the path; it's an implicit namespace, and naming it # explicitly breaks the link if [ $NS_ID -eq 0 ]; then FULL_PAGE_PATH=https://$WIKI_PATH/$PAGE_NAME LOCAL_PAGE_PATH=$PAGE_NAME fi # We don't want to consider wiki pages ending in .js, as the MW parser cannot reliably isolate URLS # in JavaScript code, so it returns erroneous links PAGE_NAME_SUFFIX=$(echo $PAGE_NAME | sed 's/.*\.//') if [ $PAGE_NAME_SUFFIX == "js" ]; then valPrint trs "Skipping URL '${LINE#$NS_ID,$PAGE_NAME,}' on line $LINK_NUM because it was found on JavaScript page '$LOCAL_PAGE_PATH'." let SKIP_JS_PAGE+=1 let PAGE_LINKS+=1 continue fi # The URL being linked to is everything after the previous two fields (this allows commas to be in # the URLs, but a comma in the previous field, the page name, will break this) URL=${LINE#$NS_ID,$PAGE_NAME,} # Scan for illegal characters if [[ $URL == *[$ILLEGAL_CHARS]* ]]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because it contains characters illegal in a URL." let SKIP_BAD_URL+=1 let PAGE_LINKS+=1 continue fi # If we're skipping archive links, see if this is one if [ $CHECK_ARCHIVE_LINKS -eq 0 ] && [[ ( $URL == *web.archive.org* || $URL == *archive.is* ) ]]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I have not been asked to check archive links." let SKIP_ARCHIVES+=1 let PAGE_LINKS+=1 continue fi # Now we need to know if the URL is for a file or a web page. First step is to determine if the # URL ends in a suffix HAS_SUFFIX=0 # If the URL ends in something like ".php?foo=bar", strip everything from the '?' onward CLEAN_URL=${URL%%\?*} # If the URL ends in something like "#section_15", strip everything from the '#' onward CLEAN_URL=${CLEAN_URL%%\#*} # 'sed' cannot handle Unicode in my Bash shell, so skip non-ASCII URL and make reader check it if [[ $CLEAN_URL == *[![:ascii:]]* ]]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I cannot handle non-ASCII characters." let SKIP_NON_ASCII+=1 let PAGE_LINKS+=1 continue fi # Isolate the characters after the last period and after the last slash POST_DOT=$(echo "$CLEAN_URL" | sed 's/.*\.//') POST_SLASH=$(echo "$CLEAN_URL" | sed 's/.*\///') # If the last period comes after the last slash, then the URL ends in a suffix POST_DOT_LENGTH=$(echo | awk -v input=$POST_DOT '{print length(input)}') POST_SLASH_LENGTH=$(echo | awk -v input=$POST_SLASH '{print length(input)}') if [ $POST_DOT_LENGTH -lt $POST_SLASH_LENGTH ]; then HAS_SUFFIX=1 else HAS_SUFFIX=0 fi # Now we need to know if the URL is for a file. If the URL ends in a suffix, compare it against the # known URL endings supplied in HTTP_FILES and HTTP_TLDS_AND_PAGES IS_FILE=-1 if [ $HAS_SUFFIX -eq 0 ]; then IS_FILE=0 else # Turn off case sensitivity while we compare suffixes shopt -s nocasematch # Special case: URLs ending in something like "productID.304297400" are pages, not files, so if # the URL's suffix is all numbers, we are looking at the end of a web page URL if [[ $POST_DOT =~ ^-?[0-9]+$ ]]; then IS_FILE=0 fi # Special case: URLs ending in a parens, e.g. "ms537113(v=vs.85)", are pages if [[ $POST_DOT =~ ^.*[\(\)]$ ]]; then IS_FILE=0 fi # Special case: URLs containing a '%', e.g. "10.1007%2FBF00329055", are pages if [[ $POST_DOT == *%* ]]; then IS_FILE=0 fi # If we did not identify this URL as a web page above, we need to compare the suffix against known # file extensions if [ $IS_FILE -eq -1 ]; then for EXTENSION in "${HTTP_FILES[@]}"; do if [[ $EXTENSION == $POST_DOT ]]; then # double brackets respect the nocasematch setting IS_FILE=1 break fi done fi # If we did not identify this URL as a file above, we need to compare the suffix against known # pages and TLDs in order to rule out that this is an unknown file suffix or TLD that the user # needs to add to the HTTP_FILES or HTTP_TLDS_AND_PAGES array if [ $IS_FILE -eq -1 ]; then for PAGE_OR_TLD in "${HTTP_TLDS_AND_PAGES[@]}"; do if [[ $PAGE_OR_TLD == $POST_DOT ]]; then IS_FILE=0 break fi done fi # Turn case sensitivity back on in Bash shopt -u nocasematch fi # If this suffix escaped identification as either a file, page or TLD, inform the reader STR_TYPE="" if [ $IS_FILE -eq -1 ]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I encountered the unknown URL suffix '$POST_DOT'. Please add this suffix to the appropriate array in this script, HTTP_FILES or HTTP_TLDS_AND_PAGES." let SKIP_UNK_SUFFIX+=1 continue elif [ $IS_FILE -eq 1 ]; then STR_TYPE="file" let FILE_LINKS+=1 else STR_TYPE="page" let PAGE_LINKS+=1 fi # Get response code using 'curl' to see if this link is valid; the --insecure option avoids an # issue with sites that require HTTPS CURL_CODE=$(curl -o /dev/null --silent --insecure --compressed --head --user-agent '$AGENT' --max-time $TIMEOUT --retry 2 --write-out '%{http_code}\n' $URL) CURL_ERR=$(echo $?) CURL_RESULT=$CURL_CODE # Tack 'curl' exit code onto result string if 'curl' returned "000" (no HTTP response) if [ $CURL_CODE == "000" ]; then CURL_RESULT="$CURL_RESULT-$CURL_ERR" fi # Begin to determine our status code for this URL (EI, IW, OK, RD, or NG) STATUS="??" NEW_URL="" INTERWIKI_INDEX=-1 # First make sure that this isn't an "external internal" link to our own wiki that can be replaced # by "[[page_name]]". If it uses a special access URL beginning with "/w/", let it pass, as it # probably cannot be replaced by "[[ ]]" markup if [[ $URL == *$WIKI_PATH* ]] && [[ $URL != *$WIKI_PATH/w/* ]]; then STATUS="EI" let EI_LINKS+=1 fi # If it's not, check if this is a link to a domain that we have an interwiki prefix for (also make # sure that it's not an archive.org link to a page from an interwiki domain) if [ $STATUS == "??" ] && [[ $URL != *web.archive.org* ]]; then for ((i = 0; i < ${#INTERWIKI_DOMAINS[@]}; ++i)); do if [[ $URL == *${INTERWIKI_DOMAINS[$i]}* ]] && [[ $URL != *${INTERWIKI_DOMAINS[$i]}/w/* ]]; then STATUS="IW" let IW_LINKS+=1 INTERWIKI_INDEX=$i break fi done fi # If we didn't match an interwiki domain, see if the status code is in our "OK" codes list if [ $STATUS == "??" ]; then for CODE in "${OK_CODES[@]}"; do if [[ $CODE == $CURL_CODE ]]; then STATUS="OK" let OK_LINKS+=1 # If this is a YouTube link, we have to look at the actual page source to know if the video # is good or not; override the link's info if it's actually NG if [[ $URL == *www.youtube.com* ]]; then PAGE_TEXT=$(curl --silent --insecure --user-agent '$AGENT' --max-time $TIMEOUT $URL | grep "\"simpleText\":\"Video unavailable\"") if [ ! -z "$PAGE_TEXT" ]; then STATUS="NG" CURL_RESULT=404 let OK_LINKS-=1 let NG_LINKS+=1 fi fi break fi done fi # If we didn't get a match with the "OK" codes, check it against the "RD" codes if [ $STATUS == "??" ]; then for CODE in "${RD_CODES[@]}"; do if [[ $CODE == $CURL_CODE ]]; then # Get URL header again in order to retrieve the URL we are being redirected to NEW_URL=$(curl -o /dev/null --silent --insecure --head --user-agent '$AGENT' --max-time $TIMEOUT --write-out '%{redirect_url}\n' $URL) # Adjust the old and new URLs to both use HTTP for comparison purposes, so we can filter # those changes out if the user didn't ask for them URL_HTTP=$(echo $URL | sed -E 's/^https:/http:/') NEW_URL_HTTP=$(echo $NEW_URL | sed -E 's/^https:/http:/') # Sometimes 'curl' fails to get the redirect_url due to time-out or bad web site config NEW_URL_LENGTH=$(echo | awk -v input=$NEW_URL_HTTP '{print length(input)}') if [ $NEW_URL_LENGTH -lt $MIN_URL_LENGTH ]; then NEW_URL_HTTP="[new URL not retrieved]" fi # Remove slash at end of new URL, if present, so we can filter out the redirects that # merely add an ending slash if the user didn't ask for them NEW_URL_NO_SLASH=$(echo $NEW_URL_HTTP | sed -E 's:/$::') # Detect if this is a youtu.be link simply being expanded by YouTube to the full # youtube.com address YOUTU_BE=0 if [[ $URL_HTTP == http://youtu.be* ]] && [[ $NEW_URL_HTTP == http://www.youtube.com* ]]; then YOUTU_BE=1 fi # If the URLs match besides HTTP being upgraded to HTTPS, then the link is OK (unless user # wants those to be reported) if [ $SHOW_HTTPS -eq 0 ] && [ $URL_HTTP == $NEW_URL_HTTP ]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I have not been asked to show http->https upgrades, and I was redirected to '$NEW_URL'." STATUS="OK" let OK_LINKS+=1 let SKIP_HTTPS_UP+=1 # If the URLs match besides an added ending slash, then the link is OK (unless user wants # those to be reported) elif [ $SHOW_SLASH -eq 0 ] && [ $URL_HTTP == $NEW_URL_NO_SLASH ]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I have not been asked to show added trailing slashes, and I was redirected to '$NEW_URL'." STATUS="OK" let OK_LINKS+=1 let SKIP_SLASH_ADD+=1 elif [ $YOUTU_BE -eq 1 ]; then # We have to look at the actual page source to know if a YouTube video is good or not PAGE_TEXT=$(curl --silent --insecure --user-agent '$AGENT' --max-time $TIMEOUT $NEW_URL | grep "\"simpleText\":\"Video unavailable\"") if [ ! -z "$PAGE_TEXT" ]; then STATUS="NG" let NG_LINKS+=1 else if [ $SHOW_YT_RD -eq 0 ]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I have not been asked to show redirects for youtu.be links, and I was redirected to '$NEW_URL'." STATUS="OK" let OK_LINKS+=1 let SKIP_YOUTU_BE+=1 else STATUS="RD" let RD_LINKS+=1 fi fi else STATUS="RD" let RD_LINKS+=1 fi break fi done fi # If we didn't get a match with the "RD" codes, check it against the "NG" codes if [ $STATUS == "??" ]; then for CODE in "${NG_CODES[@]}"; do if [[ $CODE == $CURL_CODE ]]; then STATUS="NG" let NG_LINKS+=1 break fi done fi # If we didn't match a known status code, advise the reader if [ $STATUS == "??" ]; then valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because I encountered the unknown response code $CURL_CODE." let SKIP_UNK_CODE+=1 continue fi # Check problem links against exceptions list before proceeding FOUND_EXCEPT=0 if [ $STATUS != "OK" ] && [ ! -z "$EXCEPT_URL" ]; then # The code we expect to find in the exceptions file is either the 'curl' result or "EI"/"IW" EXPECT_CODE="$CURL_RESULT" if [ $STATUS == "EI" ]; then EXPECT_CODE="EI" elif [ $STATUS == "IW" ]; then EXPECT_CODE="IW" fi # Look for link in exceptions list and make sure the listed result code and wiki page also match for ((i = 0; i < ${#EXCEPT_ARRAY[@]}; ++i)); do { EXCEPT_LINE="${EXCEPT_ARRAY[$i]}" # Undo any HTML-encoding from the wiki page; for now we just worry about the ampersand, as most # other HTML-encoded characters are not found in URLs EXCEPT_LINE=$(echo "$EXCEPT_LINE" | sed 's/\&/\&/g') # Check for URL match EXCEPT_URL="${EXCEPT_LINE#*,}" EXCEPT_URL="${EXCEPT_URL%,*}" if [ "$EXCEPT_URL" != "$URL" ]; then continue fi # Check for page name match EXCEPT_PAGE="${EXCEPT_LINE##*,}" EXCEPT_PAGE="${EXCEPT_PAGE%% *}" if [ "$EXCEPT_PAGE" == "*" ] || [ "$EXCEPT_PAGE" == "$LOCAL_PAGE_PATH" ]; then let EXCEPT_FOUND[$i]+=1 valPrint trs "Found exception '$URL' on page '$LOCAL_PAGE_PATH'." # Check for result code match EXCEPT_CODE=${EXCEPT_LINE%%,*} if [ "$EXCEPT_CODE" == "$EXPECT_CODE" ]; then FOUND_EXCEPT=1 let EXCEPT_USED[$i]+=1 valPrint trs "Skipping URL '$URL' (found on page '$LOCAL_PAGE_PATH') because its expected result, '$EXPECT_CODE', is in the exceptions list." if [ $STATUS == "EI" ]; then let SKIP_EXPECT_EI+=1 elif [ $STATUS == "IW" ]; then let SKIP_EXPECT_IW+=1 elif [ $STATUS == "RD" ]; then let SKIP_EXPECT_RD+=1 else let SKIP_EXPECT_NG+=1 fi break fi fi } done fi if [ $FOUND_EXCEPT -eq 1 ]; then continue fi # If appropriate, record this link to the log, with clickable URLs when possible if [ $STATUS != "OK" ] || [ $RECORD_OK_LINKS -eq 1 ]; then # Prepare 'curl' result in parentheses to print after status code, unless this is an "IW" or "EI" # link, in which case showing the status code doesn't make sense. Adjust spacing after string to # ensure TXT and RTF reports have aligned columns of results. CURL_STR_H=" ($CURL_RESULT)" CURL_STR_T="$CURL_STR_H" CURL_STR_R="$CURL_STR_H " if [ $STATUS == "IW" ] || [ $STATUS == "EI" ]; then CURL_STR_H="" CURL_STR_T=" " CURL_STR_R=" " fi # Record link and its wiki page in TXT, RTF, and HTML markup valPrint t "${STATUS}${CURL_STR_T} $STR_TYPE $URL" valPrint t " linked from $FULL_PAGE_PATH" valPrint r "${STATUS}${CURL_STR_R}${RTF_TABS}$STR_TYPE {\field{\*\fldinst{HYPERLINK \"$URL\"}}{\fldrslt $URL}}" valPrint r " linked from {\field{\*\fldinst{HYPERLINK \"$FULL_PAGE_PATH\"}}{\fldrslt $LOCAL_PAGE_PATH}}" valPrint hn "" valPrint hn "" # Place vertical space here since we won't be printing anything more about this link if [ $STATUS == "OK" ] && [ $SUGGEST_SNAPSHOTS_OK -eq 0 ]; then valPrint tr ""; valPrint hs ""; fi # Record redirect URL if one was given by a 3xx response page if [ $STATUS == "RD" ]; then valPrint ts " Server suggests $NEW_URL" valPrint rs " Server suggests {\field{\*\fldinst{HYPERLINK \"$NEW_URL\"}}{\fldrslt $NEW_URL}}" valPrint hs "" fi # Notify reader if we can use an intrawiki link for this URL if [ $STATUS == "EI" ]; then INTRA_PAGE=${URL#*://*/} valPrint ts " Just use [[$INTRA_PAGE]]" valPrint rs " Just use [[$INTRA_PAGE]]" valPrint hs "" fi # Notify reader if we can use an interwiki prefix for this URL if [ $STATUS == "IW" ]; then INTER_PAGE=$(echo "$URL" | sed 's/.*\///') valPrint ts " You can use [[${INTERWIKI_PREFIXES[$INTERWIKI_INDEX]}:$INTER_PAGE]]" valPrint rs " You can use [[${INTERWIKI_PREFIXES[$INTERWIKI_INDEX]}:$INTER_PAGE]]" valPrint hs "" fi # Query Internet Archive for latest "OK" snapshot for "NG" page if [[ ( $STATUS == "NG" && $SUGGEST_SNAPSHOTS_NG -eq 1 ) || ( $STATUS == "OK" && $SUGGEST_SNAPSHOTS_OK -eq 1 ) ]]; then # We need to watch out for the rate limit or we'll get locked out; look at how much time has # elapsed and then wait the remainder between that and how long of a wait we think is needed # to avoid the dreaded "Too Many Requests" response. 5 seconds is just a guess. CUR_TIME=$(date +%s) WAIT_REMAINDER=$((5 - $CUR_TIME + $START_LINK)) if [ $WAIT_REMAINDER -gt 0 ]; then valPrint t "Waiting $WAIT_REMAINDER second(s) to conform to Archive.org rate limit." sleep $WAIT_REMAINDER fi # Issue query to the API ARCHIVE_QUERY=$(curl --silent --max-time $TIMEOUT "$ARCHIVE_API?url=$URL&$ARCHIVE_OK_CODES") # Notify reader if we hit the rate limit and just keep going if [[ "$ARCHIVE_QUERY" == "*Too Many Requests*" ]]; then valPrint t " IA has rate-limited us!" valPrint r " IA has rate-limited us!" valPrint hs "" # If a "closest" snapshot was received, inform reader elif [[ "$ARCHIVE_QUERY" == *\"closest\":* ]]; then # In case the URL has a shebang in it (like mega.nz links do), escape the '!' to break it ARCHIVE_QUERY=$(echo "$ARCHIVE_QUERY" | sed 's/#!/#\\!/') # ...isolate "url" property in the response that follows the "closest" tag SNAPSHOT_URL=${ARCHIVE_QUERY##*\"closest\":} # everything after '"closest":' SNAPSHOT_URL=${SNAPSHOT_URL#*\"url\": \"} # everything after '"url": "' SNAPSHOT_URL=${SNAPSHOT_URL%%\"*} # everything before '"' # Remove the port 80 part that IA often adds to the URL, as it's superfluous SNAPSHOT_URL=$(echo $SNAPSHOT_URL | sed 's/:80//') # Inform the reader of the snapshot URL valPrint ts " IA suggests $SNAPSHOT_URL" valPrint rs " IA suggests {\field{\*\fldinst{HYPERLINK \"$SNAPSHOT_URL\"}}{\fldrslt $SNAPSHOT_URL}}" valPrint hs "" else # Otherwise give a generic Wayback Machine link for this URL, which might work valPrint ts " Try browsing $ARCHIVE_GENERIC/$URL" valPrint rs " Try browsing {\field{\*\fldinst{HYPERLINK \"$ARCHIVE_GENERIC/$URL\"}}{\fldrslt $ARCHIVE_GENERIC/$URL}}" valPrint hs "" fi fi fi # If this is a page and it seems to be there, take screenshot of it using headless Google Chrome if [ $IS_FILE -eq 0 ] && [ $TAKE_PAGE_SHOT -eq 1 ] && [ $STATUS == "OK" ]; then # Sanitize URL by removing "http(s)://" and converting any other colons or slashes to underscores SHOT_NAME=$(echo "$URL" | sed 's/https*\:\/\///' | sed 'y/:\//__/') SHOT_FILE="$SHOT_PATH/$SHOT_NAME.png" # Don't take screenshot if we already encountered this page and screenshotted it if [ ! -f "$SHOT_FILE" ]; then "$CHROME_PATH" --headless --disable-gpu --screenshot --window-size=1500,900 $URL > /dev/null 2>&1 if [ -f "$WORKING_DIR/$CHROME_SCREENSHOT" ]; then mv -n "$WORKING_DIR/$CHROME_SCREENSHOT" "$SHOT_FILE" else valPrint trhs "Screenshot of URL $URL seems to have failed!" fi else valPrint trhs "Skipping screenshot of URL '$URL' because file '$SHOT_FILE' already exists." fi fi done FINISHED_LIST="yes" wrapupAndExit
${STATUS}${CURL_STR_H}$STR_TYPE$URL
linked from$LOCAL_PAGE_PATH
Server suggests$NEW_URL
Just use[[$INTRA_PAGE]]
You can use[[${INTERWIKI_PREFIXES[$INTERWIKI_INDEX]}:$INTER_PAGE]]
IA suggests(hit the API rate limit!)
IA suggests$SNAPSHOT_URL
Try browsing$ARCHIVE_GENERIC/$URL