git @ Cat's Eye Technologies yastasoti / develop-0.3

Tree @develop-0.3 (Download .tar.gz)


Version 0.3

Yet another script to archive stuff off teh internets.

It's not a spider that automatically crawls previously undiscovered webpages — it's intended to be run by a human to make backups of resources they have already seen and recorded the URLs of.

It was split off from Feedmark, which doesn't itself need to support this function.


  • input is a JSON list of objects containing links (such as those produced by Feedmark)
  • output is a JSON list of objects that could not be retrieved, which can be fed back into the script as input
  • checks links with HEAD requests by default. --archive-to causes each link to be fetched with GET and saved to the specified directory. --archive-via specifies an archive router which causes each link to be fetched, and saved to a directory which is selected based on the URL of the link.
  • tries to be idempotent and not create a new local file if the remote file hasn't changed
  • handles links that are local files; checks if the file exists locally
  • can log its actions verbosely to a specified logfile
  • source code is a single, public-domain file with a single dependency (requests)


feedmark --output-links article/*.md | yastasoti --article-root=article/ - | tee results.json

This will make only HEAD requests to check that the resources exist. It will not fetch them. The ones that could not be fetches will appear in results.json, and you can run yastasoti on that again to re-try:

yastasoti --article-root=article/ results.json | tee results2.json

Archive stuff off teh internets

cat >links.json << EOF
        "url": ""
yastasoti --archive-to=downloads links.json

Override the filename the stuff is archived as

By default, the subdirectory and filename to which the stuff is archived are based on the site's domain name and the stuff's path. The filename, however, can be overridden if the input JSON contains a dest_filename field.

cat >links.json << EOF
        "url": "",
        "dest_filename": "home_page.html"
yastasoti --archive-to=downloads links.json

Categorize archived materials with a router

An archive router (used with --archive-via) is a JSON file that looks like this:

    "*": "/dev/null",
    "*": "footube/",
    "*": "archive/"

If a URL matches more than one pattern, the longest pattern will be selected. If the destination is /dev/null it will be treated specially — the file will not be retrieved at all. If no pattern matches, an error will be raised.

To use an archive router once it has been written:

yastasoti --archive-via=router.json links.json


Tested under Python 2.7.12. Seems to work under Python 3.5.2 as well, but this is not so official.

Requires requests Python library to make network requests. Tested with requests version 2.21.0.

If tqdm Python library is installed, will display a nice progress bar.


  • Archive youtube links with youtube-dl.
  • Handle failures (redirects, etc) better (detect 503 / "connection refused" better.)
  • Allow use of an external tool like wget or curl to do fetching.
  • Multiple --article-roots.