acidtv / octosearch Goto Github PK

requests.exceptions.ConnectionError: ('Connection aborted.', RemoteDisconnected('Remote end closed connection without response',))

For example in pseudocode ('simplematch' => user_query, 'match_url' => user_query)

So search can still match by the filename

And make it possible to configure files to be parsed with the fallback parser

last_seen is passed as a timestamp in seconds since epoch to elastic, should this be a regular datetime?

With a browser plugin or local application?

purge files that have been skipped with the last update?

record last_seen with entries, so all files last_seen before indexing started can be purged.

It also makes configuring exceptions a bit awkward:

[parser]
application/vnd.ms-excel=octosearch:application/octet-stream

It doesn't show results for the non-auth'ed sources

For stuff like 'you've logged in' or 'saved your stuff'

http://flask.pocoo.org/docs/0.12/patterns/flashing/

Use a lightweight css framework like PureCSS, Spectre or Kube?

Mostly what is on https://www.elastic.co/guide/en/elasticsearch/reference/current/query-dsl-simple-query-string-query.html, but a little more user-friendly.

Traceback (most recent call last):
File "octosearch.py", line 63, in
app.start(args)
File "octosearch.py", line 38, in start
index_job.index(indexer_conf)
File "/usr/src/octosearch/octosearch/indexer.py", line 32, in index
for file in indexer.index(conf):
File "/usr/src/octosearch/octosearch/indexers/mountedcifs/mountedcifs.py", line 42, in index
for file in super().index(conf):
File "/usr/src/octosearch/octosearch/indexers/localfs/localfs.py", line 26, in index
yield LocalFile(full_path)
File "/usr/src/octosearch/octosearch/indexer.py", line 180, in init
self._set_properties(path)
File "/usr/src/octosearch/octosearch/indexer.py", line 186, in _set_properties
statdata = os.stat(path)
OSError: [Errno 22] Invalid argument: ''

Because there can only be one auth provider configured.

Change it to check_read_permissions or something like that.

Make sure this works properly for non-public files

--index, --webserver and --truncate are all optional now. if nothing is specified the program connects to elasticsearch and then exits without an error or message.

Find out if it's possible to specify a main command that has a number of possible options attached, like the mercurial cmdline for example.

Indexer supplied data like title should probably have more priority than what the parser returns, because the indexer has more domain knowledge.

http://docs.paramiko.org/en/2.4/api/sftp.html

To see if that actually works, and if the indexer *File classes can be used.

Instead of a local search app.

For example octosearch://smb://server/path or octosearch://smb/server/path