Generating RSS from a directory every time it changes
Translation of the original in Catalan. Read it in: CatalàCastellano
For internal matters of the eBosc network, I was interested in generating an RSS file every time a directory changed its content. Next I explain how I solved the matter.
For internal matters of the eBosc network, I was interested in generating an RSS file every time a directory changed its content. Next I explain how I solved the matter.
So I installed dnotify which allowed me to execute a command every time a directory was modified and I could leave it running as a daemon (-b parameter from version 0.18 onwards).
dnotify -b -C /directory -e /path/generarRSS.sh
If it detects that a new file has been created in the directory, the indicated command is launched, in this case it is a small shell script that I have made myself, with the following content:
#!/bin/shls --color=no --sort=time -1 /path/ | /root/bin/rsspipe.py --limit 20 --url "http://localhost/" --title="Llistat del directori" --link="http://localhost/" --description="Descripció del contingut del directori" /path/fitxer.xml
As you can see the only thing I do is a listing sorted by time and then I send the output of the directory listing to the RSSPipe program. In reality this is a version that I have modified myself, since the original version did not do everything I wanted. Even though I have no idea about python without looking at any manual I have been able to modify the code, since it was easy to deduce what had to be touched for it to do what I wanted.
Code of my version of the RSSPipe:
#! /usr/bin/python## rsspipe.py v0# <http://rentzsch.com/code/rsspipe># Copyright (c) 2004 Jonathan 'Wolf' Rentzsch.# Some rights reserved: <http://creativecommons.org/licenses/by/2.0/>"""Reads lines from stdin, outputting an RSS 0.92 file containing the last N lines as items.usage example: tail -f /var/log/httpd/access_log|python rsspipe.py --title 'http traffic' httpd_access.xml""" import sys, cgifrom optparse import OptionParser parser = OptionParser()parser.add_option("-t", "--title", dest="title", help="feed title", default="rsspipe")parser.add_option("-l", "--link", dest="link", help="feed link", default="http://rentzsch.com/code/rsspipe")parser.add_option("-d", "--description", dest="description", help="feed description", default="rsspipe")parser.add_option("-i", "--limit", dest="limit", help="number of lines to export as RSS items", default=100)parser.add_option("-u", "--url", dest="url", help="la url on es pot descarregar el fitxer", default="ftp://localhost/")(options, args) = parser.parse_args() outputFile = open(args[0],'w')lineQueue = []i=0;while i<int(options.limit): readline = sys.stdin.readline() if readline != '': i=i+1; if len(lineQueue) >= int(options.limit): del lineQueue[0] lineQueue.append(readline.rstrip()) outputString = '<?xml version="1.0"?>\n<rss version="0.92">\n<channel>\n\t<title>' + options.title + '</title>\n\t<link>' + options.link + '</link>\n\t<description>' + options.description + '</description>\n'; for theLine in lineQueue: encodedLine = cgi.escape(theLine) outputString = outputString + '<item>\n\t<title>' + encodedLine + '</title>\n\t<link>' + options.url + encodedLine + '</link>\n\t<description>' + encodedLine + '</description>\n</item>\n'; outputString = outputString + '\n</channel>\n</rss>' outputFile.seek(0) outputFile.truncate(0) outputFile.write(outputString) outputFile.flush()
The changes I have made to the RSSpipe are the following:
- I have added a new parameter (-url) that serves to indicate the path of the file. I need it to reference how to download the file for example via HTTP or FTP.
- Instead of making the program run infinitely I have told it that it only runs to generate the lines it has to, not infinitely, as was originally planned.
- Inside each item the link tag now references the URL of the -url parameter plus the item in question, so the link is already usable.
- I have cleaned the output of the RSS file which was very ugly. Everything was on a single line, I have put the line breaks and indented it.