I have a User class and store the user's submitted full name as a single String:
class User {
private String mFullName;
}
There's no partial string matching available in app engine though. I am looking through current options, looks like the nonrel-search project has come up with a pretty good solution.
I'm using java though. I don't think there's a java port of that project. What are other java users doing?
I'm thinking of just writing user info to a separate MySQL database on account creation, and just hitting that database with a %like% search when users need it (this should be pretty infrequent).
Duplicating user info to a separate MySQL database would be a temporary solution until something comes built-in for app-engine java, but I haven't heard of anything like this on the roadmap.
Any info/ideas would be great!
Thanks
On the python side of things, I've seen one or two full text search implementations mentioned. The way they work is to tokenize the text to be searched and then create App Engine indexes that can be used for queries. I'm not sure if anything comparable has been created for Java at this time.
If you can make do with searches that are equivalent to 'string%' or '%string', it shouldn't be too difficult to make your own index and do a range scan. In the case of '%string', you would want to store the strings to be searched in reverse, e.g., 'gnirts%'.
Related
I'm new to Azure Cognitive Services, and while I'm pretty sure it can help me solve my problem, I don't quite understand which part of it to use for it...
Here's what I want to do:
We have blog posts, say ~1k, and those blog posts all have categories and tags (multiple each). What I want to do, is to "guess" the right categories/tags for each article based on the content, and then present that to the editor as a suggestions at the time of input ("looks like this article is about: health, well-being, ..."). The ~1k articles we already have in the system are currently correctly tagged/categorized, so I'd like to use these a data source for this "guessing".
I've used Azure Search before, and it seems like some combination of EntityRecognition and KeyPhraseExtraction might be a way in the right direction? Azure Cognitive Services also seems to have an API that supports TextAnalytics that would do something similar. I'm a bit confused about why these are two different things (or are they not?)
This also seems like an entirely common problem (matching text against pre-defined categories based on other text that is categorized), so I'm wondering if I'm just missing an obvious solution here?
Thanks in advance.
I think the Azure Cognitive Text Analytics API is your best bet as you are looking for real-time analysis prior to tagging/categorizing for storage.
Text Analytics could return a list of named entities that you could map to your available tags/categories and present to the user.
Azure Cognitive Search requires an indexer and skillset to process target text with an end result of storing the processed results to an index specifically for searching.
I've been trying my very best not to ask any nosy question here in stackoverflow, but it has been almost one week since I got stuck in this problem and I couldn't find any solution.
I already have my working website built with CakePHP 3.2. What the website basically does is scrape Twitter for tweets containing a given search term, check if it's already in my database, and store it if it doesn't yet exist. Twitter's JSON response has this "tweet_id" property, and I've been using that value to check for whether I should ignore or append a specific tweet to my DB. While this might be okay while my database is small, I suspect it's going to slow things down considerably when my tables grow bigger. Thus my need for ElasticSearch.
My ElasticSearch server is running on my Arch Linux install, and I've configured my app to point to the said server. Also, I have my "Type" object named the same way as my "Tweets" table (I followed the documentation until the overview part http://book.cakephp.org/3.0/en/elasticsearch.html). This craps out an "Unknown method "alias" error, and following Google searches led me to creating an alternate pagination class since that was what some found to be the cause of the error (https://github.com/lorenzo/audit-stash/issues/4), which still doesn't fix things.
I'm not sure if I got this right. I installed the ElasticSearch plugin with the assumption that all I have to do is name the Types the same name as my tables, since to me the documentation "implies" that this should be done on top of the Blog Tutorial they did to "improve query performance".
TLDR, how is this supposed to work? Is my above assumption right? Do I name the Types differently and index everything myself? I'm not sure if there's just too much automagic, or I'm just poor at these sort of things. And yes, I'm new to frameworks (but not PHP, among other languages)
Thanks in advance!
I'm trying to provide a chat like environment game for kids to detect if the person speaking to them is human or not (similar to what Eugene did).
I'm looking for an SDK or api that would provide me with replies for sentences (either with context of the entire conversation or without). An alternative would be a an entire database of sentences that I'll be able to search for myself and implement my own logic.
Another option is to integrate some kind of open source solution if possible.
Does anyone know of such a service \ solution can could point me to the correct direction?
If you want some database of discusions then try to found some Corpus of dialogs this is the first one which i found. It is the corpus of movie dialogs:
http://www.mpi-sws.org/~cristian/Cornell_Movie-Dialogs_Corpus.html
And if you want some chat-bot to use in your program than is probably easiest way to use some public available api of one. There is again something which could work for you:
http://blog.program-o.com/chatbot-api/
Good Luck
I'm working on the front end of an app that uses Solr for data storage. Currently I have an empty index, but it'd (understandably) be a lot easier for me if some dummy data was returned so I could make sure that it's output correctly on the front end.
If I was working with and RDBMS (let's say postgres) I'd open up a GUI (e.g. pgadmin) and type data manually into a few rows to achieve this goal. I have access to the Solr web interface, but I can't see any obvious call to action saying INSERT YOUR DATA HERE. The closest thing I can find to an answer on the web is this SO thread, but it's still not quite the droids easy GUI-based solution I'm looking for.
So, my question is: Is the a way to quickly and easily insert some data equivalent to the RDBMS method mentioned above?
Make sure you have defined a schema in schema.xml.
SOLR does indeed have a (limited) html GUI, which on a local installation is probably found at localhost:8983/solr (default). If you can get to the base admin page, then on the left there is a small combobox where you can select a core/collection. If you click on THAT, then you get a list of options that emerges, and you can pick 'documents' to get a similar GUI to what I think you expect from postgres/RDBMS/whatnot.
http://localhost:8983/solr/#/collection1/documents is the URL on a default SOLR installation that I have. This should work as long as you don't have default cores. (Replace collection1 with your collection name and localhost:8983 with wherever your solr is hosted/the port).
Background:
I'm building a poetry site with user submitted content. The relevant user actions for my questions are that users can:
a. Go to fancysitename.com/view to see all poems so far
b. Go to fancysitename.com/submit to submit your own poem.
c. Go to fancysitename.com/apoemid to view a particular poem you've bookmarked before.
d. Go to fancysitename.com/search to enter a word to search for in all the poems.
All the poems are stored as text fields in a database and referenced by a poem id. So the "apoemid" in step c will be the primary key of the tuple and I'll just pull up the text after getting the key from the url.
Question:
The poems exist nowhere except in a database. My webapp is literally 4 html files. Will this approach affect my search engine rankings?
Is there a more efficient way to do 'd' rather than do a Select * on the db and manually parsing the text on the server? Each poem will be at the most 10 lines long, so I would imagine using a full text search engine like Lucerne will probably be overkill.
Caveat
I'm running this on the google app engine for now, so my database customization options are pretty limited. So while I'd certainly be interested in hearing about the ideal way to do this, this is a pet side project so my budget is limited :(
Thanks!
Edit: Apparently I don't google so well at 7am. I've since found a solution for question 2 here so please disregard question 2.
AppEngine currently doesnt support full text indexing, they do have a better than nothing SearchableModel.
Some details of SearchableModel can be found here:
http://groups.google.com/group/google-appengine/browse_thread/thread/f64eacbd31629668/8dac5499bd58a6b7?lnk=gst&q=searchablemodel
Regarding search engine ranking, yes having all your poems in the datastore can affect your ranking. This is generally overcome through the use of a sitemap. Here is an article about how StackOverflow uses a sitemap to help its search ranking.
http://www.codinghorror.com/blog/archives/001174.html
In most database engines, you can accomplish this kind of searching. For example MysQL does have full text searching. I am not sure how app engine works but you can always have a stored procedure does this search.
Where you store your data will not affect your site's ranking, only how you serve it up (on what URLs, etc). There's absolutely no way for an arbitrary search spider to tell where you store your data, and no reason for it to care, either.
Regardless of the length of your text, you will need full-text searching if you want to search inside a string. As Sam points out, SearchableModel ought to work just fine for that.