ccs-amsterdam / amcat4client Goto Github PK

Get list of (python) script plugins from server, allow user to pick script, get options from server, let user fill in options, get results, display appropriately (e.g. list / table / qraph / custom html / custom figure)

Allow user to choose aggregation axes and possibly bins, show result as table

Allow users to pick graph type ( line/bar and/or presets like "results over time"), allow user to pick axes, show results as graph

When opening the client (summary page), it seems to also query an aggregate per day (perhaps to get total # of hits?) next to the proper aggregate per

When authentication is enabled and one tries access a specific index with guest access (e.g., http://localhost/h/localhost%2Famcat/i/state_of_the_union/query), this screen is shown:

Clicking the button does not do anything (the browser console also does not show any errors). To resolve it, the user has to manually navigate to the amcat4client address (in this case http://localhost/), log in as guest and can then access the original URL.

For existing non-text columns, offer a useful filter (e.g. date range, medium picker dropdown). Probably some sort of 'add filter' button next to the query bar?

In the filter bar, if a keyword field has more than N options it shows only the top-N

Check if 'our' old plugin still works so we get hits per documents for a query

Please check out the bug caused by the index wp3 . I have no clue about react, but now the screen is completely white, and the browser console reads as follows:

I cannot reproduce this error with the python client

To reproduce:

• login to host 1 (e.g. amcat4.labs.vu.nl/amcat)
• paste a link to a resource on anohter host (e.g. http://localhost:3000/h/http%3A%2F%2Flocalhost%3A5000/i/x/query)
• it tells you to login to localhost:5000 (correctly), but middlecat then logs you in to amcat4.labs.vu.nl, creating a circle

we need them!

It would be nice to allow an index to be used without login (presumably as reader or metareader) eg to expose/share a dataset

After changing a query it now immediately kiss the results, but if that takes longer than eg 1 second it's not obvious that it's actually peeing the new query. Some sort of loading indicator would be nice!

Currently there are some interactions between the queryscript business logic and the display of results. I think it would be good if results are 'self contained' in that they contain the metadata needed to display them and interact with them, including table structure, clicking to document details (#5), and pagination.

So every result should be structured to have 'metadata' and 'results' keys, where the metadata contains all needed information on columns, pagination, possible actions, etc.

replacement for amcat3 articlesets. tags are a normal elasticsearch field, but it would be good to give them some interface support:

• create drawer with picker list https://material-ui.com/demos/drawers/ / https://material-ui.com/demos/lists/

Add extra query bar for e.g. line graph with multiple queries

Presumably because the default time interval is per week?

Solution maybe to change the default time interval to year, or to make it dynamic based on the data range (like in the summary)

In the documents result screen, click on an article to get a (modal screen with?) the article details and text

replacement for amcat3 articlesets. tags are a normal elasticsearch field, but it would be good to give them some interface support:

• add/remove from set in article list [and 'add whole query to set']

Allow user to save a query as a stable url. Most useful if we can guarantee immutable sets