Open Semantic Search Server - Bad File URL

[blion-tec]

Hello! I've just indexed a bunch of files and realized that all the files are indexed as in the Desktop version, with file://localhost/ type URL, so they aren't accessible in any way. Is there a way to correct the references so the files are accessible outside of the box, from any machine web browser? Example:

[YoannMR]

Hi,

Not sure if this is what you need but see below my notes on how to be able to open files from a remote server (EC2 instance) in a web browser:

How to enable opening files on EC2 instance from a web browser:

• OSS comes with an apache2 web server which can access files located in /var/www/html/
• Create a logical link of the folder that holds the indexed files (ex. /root/ folder)
  • cd /var/www/html/
  • ln -s /root/
  • enable access right ex: chmod 777 /root/ or chmod 400 /root/
  • Modify OSS indexing configuration file to change path to url instead of folder path
    • nano /etc/opensemanticsearch/connector-files
    • Comment this line:

    • config['mappings'] = { "/": "file:///" }

    • Un-comment and modify this line:

    • config['mappings'] = { "/var/www/documents/": "http://www.opensemanticsearch.org/documents/" }

    • config['mappings'] = { "/var/www/html/": "http://your.ec2.ip.address/" }
  • Restart solr ?:
    • service solr stop
    • service solr start
  • Need to re-index all documents…
    • While indexing a folder, the path syntax matters (it needs to start with a ‘/’ otherwise the html href won’t start with a ‘/’ and it will add ‘/search/’ before the path which breaks it…). The trick is to put all folders to be indexed at the root (ex.: for /my_folder/, they should be at /my_folder/ not /root/my_folder/)

Also a disturbing thing about setting up the IP address in the config file: I set it up once on an old instance. I now use the same file, with same old IP address on a different instance with different IP but it still works (ie: I can open a pdf file within the webbrowser).

Disclaimer: I'm just a random user of OSS, so don't take my advice for anything more than just what I found out through errors and trials.

Hope that helps! Yoann

[rafael844]

Did everything as you said, but its not working. Im using OSS Desktop. Shared the folder sged in the VM created the link: ln -s /home/user/Documents/sf_sged/ and chmod 777 /home/user/Documents/sf_sged/

the /etc/opensemanticsearch/connector-files: config['mappings'] = { "/var/www/html/": "http://10.200.52.140/" }

the apache2.conf Listen 10.200.52.140:80

but it not works, It indexes, and I can see the preview and Tagging & annotation fine, but when I try to open the file it retrives: Not Found The requested URL /home/user/Documents/sf_sged/83/id_127_File.pdf was not found on this server.

Whats could be wrong ?

[Olydia]

I'm having the same issue, I did everything still I can't open a file on OSS server ? Any idea on how to fix the problem Thank you ! :)

[barrydegraaff]

yeah this is definitively annoying, as changing the setting config['mappings'] requires a re-index. And that takes days. As a new user installing the opensemanticsearch server package I was not aware of the fact that this used to be a desktop application. Putting file:// links in a web application is IMHO not that useful, as that means it is a single-user/single-computer web application.

It would be nice if the path override could be configured to just change the links in the web-ui, so re-indexing would not be necessary. In the mean time, I created a browser extension with some javascript that re-writes my opensemanticsearch urls.

[barrydegraaff]

In the end all I needed was: config['mappings'] = { "/": "https://mydomain.com/" }

And I purged my entire installation and reindexed.

Since some traces are left behind after purging, I also had a background task to make sure no ETL process could start before I could copy a pre-configured config file into place, basically cp'ing /root/connector-files /etc/opensemanticsearch/connector-files all the time.

[ufukayyildiz]

Did everything as you said, but its not working. Im using OSS Desktop. Shared the folder sged in the VM created the link: ln -s /home/user/Documents/sf_sged/ and chmod 777 /home/user/Documents/sf_sged/

the /etc/opensemanticsearch/connector-files: config['mappings'] = { "/var/www/html/": "http://10.200.52.140/" }

the apache2.conf Listen 10.200.52.140:80

but it not works, It indexes, and I can see the preview and Tagging & annotation fine, but when I try to open the file it retrives: Not Found The requested URL /home/user/Documents/sf_sged/83/id_127_File.pdf was not found on this server.

Whats could be wrong ?

worked for me. thanks