I am trying to build a chatbot in Rasa/Dialogflow, the problem i ham facing is to convert English to SQL query so that what user write in English can be converted into SQL fetch data from MYSQL database and display result to use.
Can someone suggest me how to do it?
Ideally this is only possible through solutions like SEQ2SQL(Link here for reference).
But I implemented it in a workaround fashion:-
I got the json using tracker.latest_message .
After which I processed the json to make our own structured json like:
[{'column_name':'a',
'operator': '=',
'value':'100'},
{'column_name':'b',
'operator': '>',
'value':'100'}]
Above structure was used to form the where clause of the query.
Same way I made a custom json for Select Part as well :-
[{sum:column1},{count:column2}]
5.Then I looped through the json I had created and made our queries.
Note:- This json Structure will not be able to cover all possible scenarios but worked decently for me.
Related
Ok, what im hoping is possible here is that given an input document with some json in it, I was considering creating a schema document using this json and then output it to a string which I can then use as a reference for future validation. (we have a LOT of json passed around in our software, so creating a schema from what we have will allow me to catch in CI when changes have been made so we can review for privacy violations).
I realise this isn't at all what the schema validator tools are for. but it looks like they do create internally the schema, does anyone have any idea how I might be able to do this?
Thanks
In our project we created several useful queries on log analytics that we deploy as a "savedSearch" (Microsoft.OperationalInsights/workspaces/savedSearches#2020-08-01).
Now when we load the query in the editor we can export it to excel, which can be nicely refreshed to view current data.
However this link is created to the query that is in the editor and not the stored/deployed query. The alternative is to export to Power Bi (M query) which generates a script that you can then use in excel.
In both cases the query itself seems to be in the connection, so it does not get updated when we deploy a new version. Does anyone know of a way to make this connection to a stored/deployed query?
I feel like this should be as straightforward as a connection to resource so that not only the data, but also the query itself gets updated.... I must be missing something
One way I can think of is to leverage Functions in log queries.
You can first save your query as a function, then export it to excel that would create a connection but execute the function, instead of the raw query.
You can tweak your query later if needed and save/overwrite to the same function, and the refresh should still be able to pull in the latest results since the changes are now neatly abstracted away via the function. :)
I am new to NodeJS, MongoDB and express.
I am able to insert, delete and update data but I dont really know how to work with it.
I want to choose some data(e.g. from a dropdown form) and then pass that data to some functions and calculate results. How is it possible to show the results to the client in the webpage(e.g. in a table)?
For example: I have different JSON objects stored in my database and now I want to select one by the name. Then do some calculations with it and show the results to the client.
Also: Is it possible to tell the Database to only accept valid JSON data?
Yes you can get specified data from the database and it will work a lot better with mongoose. You can define schema in mongoose and you can focus more on the data instead of how your data look like. And you can do validation to accept the valid data.
I want to fetch the Hollywood movies data. I have referred many APIs like IMDB, omdb, etc but all these APIs are taking a query string as a required parameter.
But i want some dummy data of movies without passing query string as a parameter with all details of movies like title, poster_image etc.
So please tell me something so that I can move further on this.
This is not a direct answer, however, have you tried googling "imdb dataset"
Here is what i have found, you need to register on their site and download CSV. Which you can later import to your DB.
IMDB Dataset - Kaggle
I try to get an extract of the text with the searched words highlighted on a JSON collection.
My search syntax is:
qb.word(qb.field('doc_text'),vartxt)
With 'doc_text' declared as field
(field type: root, include root: false, includes: doc_text), in an Node.js application.
The search works well, and it is done well on this field ...
But in txt[0].results[kl].matches[0]['match-text'], I find the first 3 properties of the JSON,
and not an extract from 'doc_text' with the words found.
I have another application in which the highlights work correctly, but it is based on XML.
Did I forget something in the field declaration, or is the operation different between JSON
and XML data, or is the highlight system not running on JSON via Node.js and QueryBuilder ?
Kind regards
Fields do not work quite the same way in XML and JSON. I think you're running into this limitation:
http://docs.marklogic.com/guide/app-dev/json#id_24090
The value of a field in XML can be the concatenation of all the text nodes, but the same does not apply in JSON.
I think I understood !!!
This query gives a correct snippet, with the extract and the words highlighted :
mkcq.and(mkcq.collection('document'), mkcq.word(mkcq.field('doc_text'), 'connaitre'))
On the other hand, this query gives the first 3 fields of the JSON :
mkcq.and(mkcq.collection('document'), mkcq.word(mkcq.field('doc_text'), 'connaitre'), mkcq.value(mkcq.element('','doc_user'), 'mbp'))
I do not know if this is normal or not, but it should be able to be corrected either by a simplified query and a selection on the returned records, evening by a particular snippeter.
Kind regards