Mercurial > ec-dotfiles
annotate moefetch.sh @ 180:8e6555aa8631
(none)
| author | edhoprima@gmail.com <edhoprima@gmail.com> |
|---|---|
| date | Fri, 05 Jun 2009 21:13:37 +0000 |
| parents | c132eb3ca841 |
| children | d3b7927bdb2b |
| rev | line source |
|---|---|
| 148 | 1 #!/bin/sh |
| 2 | |
| 3 # Copyright (c) 2009, edogawaconan <me@myconan.net> | |
| 4 # | |
| 5 # Permission to use, copy, modify, and/or distribute this software for any | |
| 6 # purpose with or without fee is hereby granted, provided that the above | |
| 7 # copyright notice and this permission notice appear in all copies. | |
| 8 # | |
| 9 # THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES | |
| 10 # WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF | |
| 11 # MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR | |
| 12 # ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES | |
| 13 # WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN | |
| 14 # ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF | |
| 15 # OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE. | |
| 159 | 16 # |
| 17 # Lots of bugs here. Use with care | |
| 148 | 18 # USE WITH CARE |
| 159 | 19 # |
| 20 # what it does: fetch every picture that has the specified TAGS. | |
| 148 | 21 # requirement: wget, libxslt, md5sum (or md5) |
| 22 | |
| 159 | 23 # program additional paths for: cut, sed, wc, MD5(sum), wget, xsltproc, grep |
| 24 ADDITIONAL_PATH= | |
| 148 | 25 |
| 159 | 26 # custom md5 path with arguments, expected output: <32digit md5><space(s)><filename> |
| 27 # Leave empty for "md5sum" (Linux, Solaris), "md5 -r" (*BSD) | |
| 28 MD5= | |
| 148 | 29 |
| 159 | 30 # default server address. Danbooru only! I do not take responsibility of stupidity. |
| 31 DEFAULT_SITE="moe.imouto.org" | |
| 148 | 32 |
| 33 # base directory. make sure it's writeable. I do not take responsibility if you don't own the folder and files as no check is done for this one. | |
| 159 | 34 # Structure is ${BASE_DIR}/<TAGS> |
| 35 # Absolute path only. | |
| 36 # Leave empty to use whatever folder you're running this at | |
| 37 BASE_DIR="" | |
| 148 | 38 |
| 39 # not user modifiable from here | |
| 40 | |
| 159 | 41 # useless welcome message. Also version |
| 42 Msg_Welcome() { | |
|
174
0948e76a57a1
added help. Bump to 0.1-beta2
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
173
diff
changeset
|
43 MOEFETCHVERSION="0.1-beta2" |
| 159 | 44 cat <<EOF |
| 45 moefetch ${MOEFETCHVERSION} | |
| 46 Copyright (c) 2009 edogawaconan <me@myconan.net> | |
| 47 | |
| 48 EOF | |
| 49 } | |
| 50 | |
| 51 # fatal error handler | |
| 52 Err_Fatal() { | |
| 53 echo "Fatal error: ${1}" | |
| 54 exit 1 | |
| 55 } | |
| 56 | |
| 57 # help message | |
| 58 Err_Help() { | |
| 59 cat <<EOF | |
| 177 | 60 moefetch.sh COMMAND [-s SITE_URL] TAGS |
|
174
0948e76a57a1
added help. Bump to 0.1-beta2
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
173
diff
changeset
|
61 |
|
176
3d2ae9417273
even more improvement
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
175
diff
changeset
|
62 COMMAND: |
|
3d2ae9417273
even more improvement
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
175
diff
changeset
|
63 (quick)fetch: do a complete update. Add prefix quick to skip file checking |
|
3d2ae9417273
even more improvement
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
175
diff
changeset
|
64 check: get list of new files, clean up local folder and print total new files |
|
175
5b7a154dbd21
cosmetics fix for help message
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
174
diff
changeset
|
65 |
| 177 | 66 -s SITE_URL: Specify URL of the Danbooru powered site you want to leech from. Default is ${DEFAULT_SITE} |
|
175
5b7a154dbd21
cosmetics fix for help message
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
174
diff
changeset
|
67 |
|
176
3d2ae9417273
even more improvement
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
175
diff
changeset
|
68 TAGS: Tags you want to download. Separated by spaces. Tag name follows standard Danbooru tagging scheme |
|
174
0948e76a57a1
added help. Bump to 0.1-beta2
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
173
diff
changeset
|
69 |
| 159 | 70 EOF |
| 71 exit 0 | |
| 72 } | |
| 73 | |
| 74 # generate link by transforming xml | |
| 75 Generate_Link() { | |
| 76 cd "${BASE_DIR}/temp" | |
| 148 | 77 echo |
| 78 echo "Fetching xml file" | |
| 159 | 79 wget "http://${SITE}/post/index.xml?tags=${TAGS}&offset=0&limit=100000" -O "${SITE_DIR}-${TARGET_DIR}-xml" -e continue=off |
| 148 | 80 echo "Processing XML file..." |
| 81 # xslt evilry | |
| 165 | 82 xsltproc - "${SITE_DIR}-${TARGET_DIR}-xml" <<EOF | sed 's/.*\(http.*\)\(\/[a-f0-9]\{32\}\).*\.\([jp][pn]g\)/\1\2.\3/g' | grep ^http > "${SITE_DIR}-${TARGET_DIR}-list" |
| 148 | 83 <xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0"> |
| 84 <xsl:output method="xml" indent="yes"/> | |
| 85 <xsl:template match="post"> | |
| 86 <xsl:value-of select="@file_url" /> | |
| 87 </xsl:template> | |
| 88 </xsl:stylesheet> | |
| 89 EOF | |
| 159 | 90 echo "`echo \`wc -l < \"${SITE_DIR}-${TARGET_DIR}-list\" \`` file(s) available on server" |
| 91 #output file: ${TARGET_DIR}-list | |
| 148 | 92 } |
| 93 | |
| 159 | 94 # check tools availability |
| 95 Check_Tools() { | |
| 96 # verify all programs required do indeed exist | |
| 97 #MD5 | |
| 98 if [ ! "${MD5}" ]; then | |
| 99 case `uname` in | |
| 100 *BSD) MD5="md5 -r";; | |
| 101 Linux|SunOS) MD5="md5sum";; | |
| 102 *) Fatal_Err "No known md5 tool for this platform. Please specify manually" | |
| 103 esac | |
| 104 fi | |
| 105 MD5_COMMAND=`echo ${MD5} | cut -d' ' -f1` | |
| 106 # basic tools | |
| 107 COMMANDS="cut sed wc wget xsltproc xargs rm mkdir chown comm grep ${MD5_COMMAND}" | |
| 108 for COMMAND in ${COMMANDS} | |
| 109 do | |
| 110 COMMAND_CHECK=`command -v "${COMMAND}"` | |
| 111 [ "${COMMAND_CHECK}" ] || Err_Fatal "${COMMAND} doesn't exist in ${PATH}" | |
| 112 done | |
| 113 | |
| 114 # grep checking | |
| 115 # originally created for workaround on solaris | |
| 116 #if [ `uname` = "SunOS" ]; then | |
| 117 FAIL="" | |
| 118 echo "blah" > superrandomtestfile | |
| 119 echo "blah" > superrandomtestfile.2 | |
| 120 grep -f superrandomtestfile.2 superrandomtestfile > /dev/null 2>&1 || FAIL=1 | |
| 121 rm -f superrandomtestfile superrandomtestfile.2 | |
| 122 [ "${FAIL}" ] && Err_Fatal "Your grep is not compatible. Please install or set path of correct grep" | |
| 123 } | |
| 124 | |
| 125 # verify required folders exist and writeable | |
| 126 Check_Folders(){ | |
| 127 [ -O "${BASE_DIR}" ] || Err_Fatal "You don't own ${BASE_DIR}. Please fix ${BASE_DIR}." | |
| 180 | 128 for FOLDER in temp trash deleted "${SITE_DIR}/${TARGET_DIR}"; do |
| 159 | 129 if [ ! -d "${BASE_DIR}/${FOLDER}" ]; then |
| 130 mkdir "${BASE_DIR}/${FOLDER}" || Err_Fatal "${FOLDER} folder creation failed" | |
| 131 fi | |
| 132 if [ ! -O "${BASE_DIR}/${FOLDER}" ]; then | |
| 167 | 133 echo "You don't own the ${BASE_DIR}/${FOLDER}, applying globally writeable permission on it" |
| 159 | 134 chmod -R u=rwX,g=rwX,o=rwX "${BASE_DIR}/${FOLDER}" || Err_Fatal "Error changing ownership. This shouldn't happen" |
| 135 fi | |
| 136 done | |
| 167 | 137 [ `echo \`ls "${BASE_DIR}/${SITE_DIR}/${TARGET_DIR}" | wc -l\`` -eq 0 ] && ISNEW=1 |
| 159 | 138 # let's move to workdir |
| 139 cd "${BASE_DIR}/temp" | |
| 140 for i in error ok list newlist; do | |
|
161
52877e2849bb
misc fix. These past commits wasn't actually tested
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
160
diff
changeset
|
141 touch "${SITE_DIR}-${TARGET_DIR}-${i}" || Fatal_Err "Error creating ${TARGET_DIR}-${i}. This shouldn't happen" |
| 159 | 142 done |
| 143 # | |
| 144 } | |
| 145 | |
| 146 # check files correctness | |
| 147 Check_Files() { | |
| 166 | 148 if [ ! "${ISNEW}" ]; then |
| 148 | 149 echo "Checking for errors..." |
| 150 # THE FILES | |
| 159 | 151 |
| 152 # current dir: ${BASE_DIR}/${SITE_DIR}/${TARGET_DIR} | |
| 153 cd "${BASE_DIR}/${SITE_DIR}/${TARGET_DIR}" | |
|
173
2b7c8c1ecdfe
better trash handler
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
172
diff
changeset
|
154 TRASH_DIR="${SITE_DIR}-${TARGET_DIR}-`date -u +%Y%m%d-%H.%M`" |
|
2b7c8c1ecdfe
better trash handler
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
172
diff
changeset
|
155 mkdir -p "${BASE_DIR}/trash/${TRASH_DIR}" |
|
179
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
156 # FIXME FIXME FIXME FIXME FIXME FIXME |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
157 for TRASH in * |
| 148 | 158 do |
|
179
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
159 if [ -d "${TRASH}" ]; then |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
160 mv -f "${TRASH}" "${BASE_DIR}/trash/${TRASH_DIR}" || Err_Fatal "Error deleting files" |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
161 echo "Moved ${TRASH} to ${BASE_DIR}/trash/${TRASH_DIR}" |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
162 else |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
163 TRASH="`echo ${TRASH} | sed -e 's/\([0-9a-f]\{32\}.*\)//g' | grep -v ^$`" |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
164 if [ "${TRASH}" ]; then |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
165 mv -f "${TRASH}" "${BASE_DIR}/trash/${TRASH_DIR}" || Err_Fatal "Error deleting files" |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
166 echo "Moved ${TRASH} to ${BASE_DIR}/trash/${TRASH_DIR}" |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
167 fi |
|
c132eb3ca841
improved folder checker
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
178
diff
changeset
|
168 fi |
| 159 | 169 done |
|
173
2b7c8c1ecdfe
better trash handler
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
172
diff
changeset
|
170 [ "`ls "${BASE_DIR}/trash/${TRASH_DIR}"`" ] || rmdir "${BASE_DIR}/trash/${TRASH_DIR}" |
| 159 | 171 printf "" > "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}-error" |
| 172 for FILE in * | |
| 173 do | |
| 170 | 174 if [ "`${MD5} "${FILE}" | cut -d ' ' -f1 -`" != "`echo "${FILE}" | cut -d '.' -f1`" ] |
| 148 | 175 then |
| 159 | 176 echo |
| 177 echo "${FILE}" >> "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}-error" | |
| 178 echo "Error: ${FILE}" | |
| 148 | 179 fi |
| 180 printf "." | |
| 181 done | |
| 182 echo | |
|
178
3f5ee8b2791f
error when restructuring
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
177
diff
changeset
|
183 TOTAL_ERROR=`echo \`wc -l < "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}-error"\`` |
|
3f5ee8b2791f
error when restructuring
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
177
diff
changeset
|
184 echo "${TOTAL_ERROR} file(s) error" |
| 170 | 185 echo "Removing error files" |
| 186 if [ "${TOTAL_ERROR}" -eq 0 ]; then | |
| 187 echo "No error file. 0 file removed" | |
| 188 else | |
| 189 cat "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}-error" | xargs rm | |
| 190 echo "${TOTAL_ERROR} file(s) removed" | |
| 191 fi | |
| 192 echo "`echo \`ls | wc -l\`` file(s) available locally" | |
| 193 | |
| 159 | 194 # current dir: ${BASE_DIR}/temp |
| 195 cd ${BASE_DIR}/temp | |
| 148 | 196 |
| 197 echo "Generating list of new files..." | |
| 198 # THE FILES | |
| 159 | 199 #ls "../${TARGET_DIR}" | grep -vf "${TARGET_DIR}-error" > "${TARGET_DIR}-ok" |
|
156
d3b002fd944e
fix: my attempt at speeding up things failed. reverting back to trusty grep -vf
edhoprima
parents:
155
diff
changeset
|
200 # |
| 159 | 201 ls "${BASE_DIR}/${SITE_DIR}/${TARGET_DIR}" | comm -1 -3 "${SITE_DIR}-${TARGET_DIR}-error" - > "${SITE_DIR}-${TARGET_DIR}-ok" |
| 202 cat "${SITE_DIR}-${TARGET_DIR}-list" | grep -vf "${SITE_DIR}-${TARGET_DIR}-ok" > "${SITE_DIR}-${TARGET_DIR}-newlist" | |
| 203 echo "`echo \`wc -l < \"${SITE_DIR}-${TARGET_DIR}-newlist\"\`` file(s) to be downloaded" | |
| 204 | |
| 148 | 205 else |
| 159 | 206 if [ "${ISQUICK}" ]; then |
| 152 | 207 echo "quick mode selected. Skipping check" |
| 208 else | |
| 209 echo "Empty local repository" | |
| 210 fi | |
| 159 | 211 cd "${BASE_DIR}/temp" |
| 166 | 212 cat "${SITE_DIR}-${TARGET_DIR}-list" > "${SITE_DIR}-${TARGET_DIR}-newlist" |
| 148 | 213 fi |
| 214 } | |
| 215 | |
| 159 | 216 # start downloading the images |
| 217 Fetch_Images() { | |
|
160
68227a30d0b3
forgot to fix Fetch_Images to reflect new folder naming scheme
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
159
diff
changeset
|
218 cd "${BASE_DIR}/temp" |
|
68227a30d0b3
forgot to fix Fetch_Images to reflect new folder naming scheme
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
159
diff
changeset
|
219 if [ `echo \`wc -l < "${SITE_DIR}-${TARGET_DIR}-newlist"\`` -eq 0 ]; then |
| 148 | 220 echo "No new file" |
| 221 else | |
| 222 echo "Starting wget" | |
|
160
68227a30d0b3
forgot to fix Fetch_Images to reflect new folder naming scheme
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
159
diff
changeset
|
223 cd "${BASE_DIR}/${SITE_DIR}/${TARGET_DIR}" |
|
68227a30d0b3
forgot to fix Fetch_Images to reflect new folder naming scheme
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
159
diff
changeset
|
224 wget -e continue=on -bi "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}-newlist" -o "${BASE_DIR}/temp/${SITE_DIR}-${TARGET_DIR}.log" |
| 148 | 225 fi |
| 226 } | |
| 227 | |
| 159 | 228 # initialize base variables and initial command check |
| 229 Init(){ | |
| 230 # Get base folder - current folder or fallback to ${HOME} | |
| 231 [ "${BASE_DIR}" ] || BASE_DIR="${PWD}" | |
| 232 [ "${BASE_DIR}" ] || BASE_DIR="{$HOME}" | |
| 233 [ "`echo ${BASE_DIR} | cut -c1 | grep \/`" ] || BASE_DIR="/${BASE_DIR}" | |
| 234 # path initialization | |
| 235 [ "${ADDITIONAL_PATH}" ] && PATH=${ADDITIONAL_PATH}:${PATH} | |
| 236 export PATH | |
|
158
cba73f6a96bb
grep check. OpenSolaris' default grep doesn't support -f
edhoprima
parents:
157
diff
changeset
|
237 |
| 159 | 238 # misc variables |
| 166 | 239 ISQUICK= |
| 240 ISNEW= | |
|
158
cba73f6a96bb
grep check. OpenSolaris' default grep doesn't support -f
edhoprima
parents:
157
diff
changeset
|
241 |
| 159 | 242 [ $# -lt 2 ] && Err_Help |
| 243 case "$1" in | |
|
174
0948e76a57a1
added help. Bump to 0.1-beta2
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
173
diff
changeset
|
244 check|fetch|quickfetch) |
| 159 | 245 echo "Starting..." |
| 246 JOB="$1" | |
| 247 ;; | |
| 248 *) | |
| 249 Err_Help | |
| 250 ;; | |
| 251 esac | |
| 252 shift | |
| 253 SITE= | |
| 254 case "$1" in | |
| 255 -s|--site) | |
| 256 shift | |
| 257 SITE="$1" | |
|
163
e2149ba6ab9c
shift placed at wrong place.
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
162
diff
changeset
|
258 shift |
| 159 | 259 ;; |
| 260 *) | |
|
161
52877e2849bb
misc fix. These past commits wasn't actually tested
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
160
diff
changeset
|
261 SITE="${DEFAULT_SITE}" |
| 159 | 262 ;; |
| 263 esac | |
| 166 | 264 TAGS="$@" |
| 159 | 265 echo "Tags: ${TAGS}" |
| 266 # slash is not wanted for folder name | |
|
162
1f937c2e8b3f
tags doesn't get parsed :(
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
161
diff
changeset
|
267 TARGET_DIR="`echo "${TAGS}" | sed -e 's/\//_/g'`" |
|
1f937c2e8b3f
tags doesn't get parsed :(
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
161
diff
changeset
|
268 SITE_DIR="`echo "${SITE}" | sed -e 's/\/$//g;s/\//_/g'`" |
| 159 | 269 } |
| 148 | 270 |
| 159 | 271 Msg_Welcome |
| 272 Init "$@" | |
| 273 Check_Tools | |
| 274 Check_Folders | |
|
158
cba73f6a96bb
grep check. OpenSolaris' default grep doesn't support -f
edhoprima
parents:
157
diff
changeset
|
275 |
| 148 | 276 |
| 159 | 277 # let's do the job! |
| 278 case "${JOB}" in | |
|
174
0948e76a57a1
added help. Bump to 0.1-beta2
edhoprima@gmail.com <edhoprima@gmail.com>
parents:
173
diff
changeset
|
279 check) |
| 159 | 280 Generate_Link |
| 281 Check_Files | |
| 148 | 282 ;; |
| 159 | 283 fetch) |
| 284 Generate_Link | |
| 285 Check_Files | |
| 286 Fetch_Images | |
| 287 ;; | |
| 288 quickfetch) | |
| 289 ISNEW=1 | |
| 290 ISQUICK=1 | |
| 291 Generate_Link | |
| 292 Check_Files | |
| 293 Fetch_Images | |
| 148 | 294 ;; |
| 295 esac |