fulltextsearch

fullnextsearch is a replacement to nextant.

v0.10.4

• Upload/creating a .noindex file in a folder will prevent nextant to index its content (and subdirectories)
• --background will force the scan on next tick of the cron
• Live Index can use database instead of Semaphore
• bugfixes

• Upload/creating a .noindex file in a folder will prevent nextant to index its content (and subdirectories)
• --background will force the scan on next tick of the cron
• bugfixes.

v0.10.2

• To make nextant working fine on single-board computers, admin can now choose between 3 levels of performance: High, Low, Lower.
• compatible NC11
• bugfixes

v0.10.0

• nextant is working fine with Solr 6.3.0
• results are displayed using fileList.
• public sharelink are using same JS than private
• bugfixes:
  • local file that are considered not local by NC (ie. encrypted module enabled but not used).
  • default file filters configuration could make files not to be extracted.
  • other minor bugfixes. UI glitch.

v0.6.3

• search within link shared documents, both public and private (w/ key)
• search within federated documents
• admin can choose to add files and directory structure to improve search
• mid, mp3 and flac are now indexed.
• index/search now ignore case and accents
• minor bugfixes

v0.5.1

• bugfix: search allowed to non-admin
• bugfix: warning on deleted file with no extension

Well, still bugged but at least some neat features are here !

This update resets the current setup of your Nextant. Meaning that, after upgrading, you will need to visit the administration page to configure your Nextant (again), and index your files (again).

It is recommended to delete/recreate the core from your Solr Servlet before configuring your Nextant.

$ sudo -u solr /opt/solr/bin/solr delete -c nextant
$ sudo /etc/init.d/solr restart
$ sudo -u solr /opt/solr/bin/solr create -c nextant

Note: Indexing requiert a descent amount of ressources. I am aware that some of you are running nextcloud on single-board computer (Raspberry Pi, oDroid, ...) and I have no idea if Solr is running fine on that kind of equipment.

v0.4.2

• extract and index files from trash
• orphans documents are removed while indexing
• Force index unlock after 24h
• maximum size on extracted file
• new field nextant_path
• bugfixes

• Background jobs & index safe guard,
• New options,
• better UI,
• Highlighting result,
• bugfixes,
• more wiki

tl;dr: ./occ nextant:index is (a lot) better now.

The whole process of extracting document, indexing them and updating data relative to the original files has been completly reworked. The main result is that you can restart the indexation of your files from the point you stopped/crashed.

Also, fixed few bugs.

Note: as usual, after an upgrade or an installation, you need to index the files from your cloud. First, check that your Solr core has the correct schema by using ./occ nextant:check, then run the indexation: ./occ nextant:index.

tl;dr: You can now search within shared documents and you can index more file format.

Also the installation process is now easier as the schema of your Solr Core is managed by Nextant.

Note: if you're upgrading, please consider using ./occ nextant:check or click the Nextant/Test & Save button from the administration page of your cloud.

• More indexed file formats: opendocument, office, rtf, html, epub, ...
• Managing Solr Schema from Administration page or using ./occ nextant:check
• You can clear your Solr core using ./occ nextant:clear
• auto-check your solr setup on upgrade.
• 'configured' flag to verify that the app is well configured.
• bugfixes:
  • result being displayed twice.
  • escape few queries to Solr

What does it do right now:

• When uploaded to the cloud, text and pdf file are extracted to the Solr Server.
• It add an owner filter to solr documents so each user search in its own library.
• Use the Solr server when using the searchbox in the files App of your nextcloud.
• You can also extract your current files using ./occ nextant:scan

What does it do right now that it should not :

• results can be displayed behind an element so it can't be clicked.