Q&A on KB Articles using NLP - nlp

I have a large set of KB Articles (software support), and I am looking for ways to extract answers based on a question asked against these KB articles. I am looking for guidance to understand the following -
Are there libraries available (maybe Python) that could allow me to get a potential answer or a set of ranked/scored answers based on a question posed?.
I had a look at IBMs Watson Service and that does a pretty good job, but these articles are not allowed to be interpreted through a cloud service.
PS: I know questions requesting an open-ended guidance are not usually part of StackOverflow's ecosystem, but the domain this question covers is too wide to get a good direction anywhere else.

Yes, its possible.
My team at Bluetick Consultants has been working on this since 6 months and we have figured it out. We have tested it on dropbox KB.
Our solution uses multiple python libraries, tensorflow and custom ML models.

You can check cdQA python library which allows you to build custom Closed Domain Question Answering Systems without using the cloud: https://github.com/cdqa-suite/cdQA

Related

Liferay 7 - Knowledge base versus Web contents

I'm working on Liferay 7 and I have a question on using the portlet Knowledge Base or Web contents. I'm wondering which one is the best to do a FAQ.
Web contents can now be organized in folders, have tags and plenty of informations. KB seems to offer the same capabilities but the content is limited to Markdown script.
Does someone used the KB and have feedback on it ?
Any advice appreciated !
Julien
If you have a small set of FAQ and you require a simple implementation, the web content will suit you as they are simple.
The knowledge base is good fit for web-book kind of things.
If you have dynamic FAQ where people can add edit the answers or where multiple answers are required, you can go for message boards. It has an application display template(ADT) for Q&A as well. I am not very sure but a message broad with Q&A template can be limited to one answer only (with GUI configuration) to function like FAQ.
The other way is to download the message broad Q&A ADT from Liferay sources and tweak it to display as FAQ. This will not require much effort and this will provide for the dynamism in FAQ functionality.

Open source audio pattern recognition (finger printing) [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 6 years ago.
Improve this question
Is there any open source audio fingerprinting solution that provides a "Shazam like" solution?
Shazam uses the commercial solution from Landmark digital but I am looking for an alternative (even if inferior in accuracy and performance) open source solution.
Any one knows anything about such implementation or even of a published not implemented algorithm?
echoprint looks like a solid alternative to commercial libs.
for speech recognition:
CMUSphinx
OpenEars
we only tested echoprint and so far so good.
I know the answer is a bit late but as this page pops up in google, we may as well improve it over time :)
edited as CMU is speech oriented
There are a few Open Source audio recognition projects out there, although you can forget the quality/performances of commercial audio identification services such as Shazam, SoundHound, etc. It also depends on what kind of recognition you want to do (audio files fingerprinting/tagging, real-time identification, OTA (Over-The-Air) recognition, etc.)
Open Surce
AcoustID/MusicBrainz Provides tools to perform audio fingerprinting and metadata association (fingerprinter and server). You can put up your own fingerprint/metadata server or use MusicBrainz's service. Good solution for audio file fingerprinting and recognition, not suitable for real-time high-performances applications (i.e. OTA). We have also tested it for audio stream monitoring (broadcast monitoring) but the results were quite poor.
Echoprint Provides the same tools as AcoustID (fingerprinter and server+metadata) and can also be used for audio stream monitoring as it can recognize snippets taken from anywhere within the audio with decent accuracy (but i wouldn't use it for serious commercial applications). They also claim it is suitable for OTA applications but the performances are nowhere suitable for production uses.
Last.fm Not fully open source, although they have open sourced the fingerprinting module. You have to query their identification and metadata provider services as they do not provide a full fledged solution.
ok. different search keywords, all the results I looked for.
http://musicbrainz.org/doc/Audio_Fingerprint
Thanks for anyone viewing
If you are looking for a solution in .NET check SoundFingerprinting library.
It's open source and built on top of Content Fingerprinting Using Wavelets research paper.
The algorithm is different from Shazaam's, but the general idea is similar: extract most prominent coefficients from the spectrum, then use them to build the fingerprints for later retrieval.
Description of the algorithm can be found here.
If you are looking for JAVA library then go for MusicG, I have used it in my one of past project and its working fine.
http://code.google.com/p/musicg/

Simple self-hosted website monitoring [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 3 years ago.
Improve this question
I'm looking for a simple self hosted website monitoring tool.
It should be somthing similar to watchmouse.com or pimgdom.com, with a nice UI, colorful charts and so on (Customers like that :)).
At the moment we use Zabbix also for HTTP monitoring, but since now our hoster care about the hardware and software monitoring on the machine directly, we don't need Zabbix anymore.
For pure http-monitoring zabbix or an other monitoring suite is really an overkill.
So what I'm not looking for is:
Zabbix
Nagios
Hyperic
...
Sadly but the truth, after some hours of researching I wasn't able to find a fitting application. My hope is now on you.
I realize this is an old question but I was looking for something like this today and came across Cabot which is self hosted and free, and according to the project's description: "provides some of the best features of PagerDuty, Server Density, Pingdom and Nagios".
Hope this helps someone in the future.
I found this a while ago for my purposes. Nice and simple and self hosted.
You do need shell access to setup cron jobs for it so it probably won't work in a shared environment.
php Server Monitor
Hope this helps.
Peter
I had a lot of success with Groundwork in the past, It's a BEAST and does just about everything imaginable and can be configured in so many ways. It might be overkill if you are just looking for something to schedule some http responses then graph the logs.
Groundwork is more for enterprise level deployments and has both Paid and Community editions with a pretty active community behind it too.
Not sure if you have already found a solution to this or not but give a shot to Apica System's Synthetic Monitoring. You can use the full SaaS, full on-premise, or hybrid model of this system. Take a look at the free trial and if you like what you see, the full portal as well as monitoring agents (with tons of more features than the trial) can be hosted behind your firewall in your own network. As per for monitoring, you can monitor websites/mobile apps, API endpoints, DNS, etc. You can also run complex use cases and see how the web app responds using Selenium or ZebraTester scripts.
If all you want to monitor is website uptime/downtime and response time, I'd have a look at TurboMonitor - it doesn't have all the bells and whistles provided by some other monitoring websites but it's quick and accurate for those two things.
Price-wise, I wouldn't take what they have on their website too seriously. I only actually found out about them when I met them in person and they were very happy to give me a "professional" account for free, supposedly like 5€/month or something on their website.

Looking for a web host that can handle Haskell [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
We don’t allow questions seeking recommendations for books, tools, software libraries, and more. You can edit the question so it can be answered with facts and citations.
Closed 1 year ago.
Improve this question
I'm looking for a web host that will let me run a Haskell web application. VPS's seem attractive to me because you can run essentially anything you want. But some of the cloud hosts offer really nice scalability in terms of hard disk space and bandwidth.
Does anyone know of a host that will let me run exotic languages like Haskell but can also seamlessly scale up the hard disk space/RAM/bandwidth/CPU available to my host?
If you just want very simple hosting with CGI, NearlyFreeSpeech.net supports Haskell and some other less common languages. I personally also like their overall nonsense-free approach and sensible pricing model (pre-pay metered charges, instead of the usual model of a fixed monthly charge, oversold server capacity, and absurd overage fees).
There are a few caveats however, mainly that they don't permit standalone servers or persistent daemons, only things invoked via CGI from Apache. This might be a problem for some Haskell web app frameworks.
Maybe this is obvious, but you can always use Amazon EC2. You'll have full control, and definitely meets your requirement for seamlessly scaling up.
This may be a very late answer but I found that hosting on Heroku with its Cedar stack is the easiest. Yesod has a very clear explanation.
Apparently, it's possible to get ghc running on Webfaction. There are also threads about it in the Webfaction support forums, and the admins/techs are quite willing to make an effort to make it work, though it's clearly not something that is supposed to be available out of the box.
EDIT, 2011-08-23: Fixed link.
In theory all you need is CGI/FastCGI support. I've had some luck playing around with Happstack on a very basic Dreamhost account by following these instructions:
While non-trivial to get running, this
web experiment proves that it is at
the very least possible to run
Happstack applications on cheap
hosting providers such as Dreamhost
with little more than a shell account
and CGI support.
I've only tried this with toy applications, and don't know how it would scale.
Looks like you can also run Haskell in Azure Functions.
If you are using IHP (Integrated Haskell Platform), you can use their free cloud hosting service at https://ihpcloud.com/.

Document Management [closed]

Closed. This question does not meet Stack Overflow guidelines. It is not currently accepting answers.
Questions asking us to recommend or find a tool, library or favorite off-site resource are off-topic for Stack Overflow as they tend to attract opinionated answers and spam. Instead, describe the problem and what has been done so far to solve it.
Closed 9 years ago.
Improve this question
We looking for a simple, open source, web based document management system for Linux. With document management I mean the ability to store a set of files (minimally doc, xls and pdf) as a document. Associate meta data with the document like owner and version. Update and delete documents. Ability to index and search content. Authentication and the ability to authorize at least read, and possible write. If possible I would like to avoid implementations in Java or PHP, and as we use MySQL already that would work especially well for meta-data storage.
We have used Google Applications in the past but the lack of support for PDF makes it a poor fit. Other downsides include their service losing some of our spreadsheets, no concept of company owning information opposed to individual accounts, and some of our information is sensitive and we prefer keeping it in-house (passwords, contracts etc).
MediaWiki was not a good fit either as our documents is really a set opposed to be structured content (i.e. not looking for a content management system), and at least the version we had installed did not deal well with attachments.
Based on review of past questions I plan on looking into KnowledgeTree. Any other projects that we should consider?
I've been using KnowledgeTree now for a few months developing an ASP.Net application and I only have good things to say about it. Our product uses it for PDF storage/retrieval and it really couldn't be easier to deal with. The basic install gives you a simple environment with almost endless amounts of configuration for meta-data, document groups, and various security options. Also, the KnowledgeTree staff have been very helpful and have provided us with sample code when we have run into 'how are we going to do that?' moments.
I'll second the recommendation for KnowledgeTree. Have been using it for a couple years and have roughly 1K documents indexed. Sometime last year, I wrote a short script which monitors KT's transaction table (in MySQL) and notifies users of new or updated documents via Twitter, Identica, and/or Jabber. The Twitter/Identica feeds can then be monitored with a RSS reader.
Look for something that will index all your document formats and keep them searchable.
I solved this in my office using Coldfusion. It has verity search engine built in. This indexes files on your network (doc/xls/pdf, etc) to make the text in them searchable (like google).
An instant search engine for all my files, for upto 150,000 or so is built in for free with Coldfusion so it suits my purpose.. Something like this would allow you to save your files on a network how/wherever and you'd be able to extract the rest of the information about owners, modification dates through libraries available in java / .net.
I am sure you could replicate this with another language, but probably a bit more effort. I am presently wishing I could use the Google Docs API as a wysiwyg editor in my own wiki in-house.. that would solve most of my problems then because everything would be intranet based.
Try https://www.mayan-edms.com, written on Django, db agnostic
You can consider GroupDocs as they have got storage, conversion and few more features.

Resources