terranodo / angular-search Goto Github PK

Similar to keywords as portrayed in @gsolat's mockups.

• Ask Ben Lewis for link to JDA application (Ariel)
• Look at heatmap display when there is just one item, when there are a lot of items, and look at how contrast / colors are implemented (Mila)
• Decide whether to stay or change the background layer (Mila)
• Create tests to maintain behavior in the future. (Mila)

DatePicker should be reset

see related here:
#29 (diff)

Know the dates of the first and last tweet to set dynamically the datepicker limits.

Once I type a user, there is no way to filter it out anymore, unless I refresh the page.

Not on filter box.

Implement suggested keywords below search

• When clicking on the keywords they should get added to the search terms
• Possibility to delete keywords by clicking on a close button in the search box
• Refresh list when zooming or filtering by other means.

port is 3001 not 3000.

npm run server

> [email protected] server /Users/panchicore/www/heatmaps/angular-search
> parallelshell 'npm run live-reload -s' 'npm run serve -s' 'grunt dev-watch'

Starting up http-server, serving ./
Available on:
  http://127.0.0.1:3001

Screenshots:

Including:

• Tweet location on the map - in a way that does not get hidden by the heatmap (just like layer's bbox on registry)
• Shows embedded pictures, videos, and a scroll with the tweet replies.
• In a panel that is not a popup but has a location fixed in the layout. It can appear just on hover but it should have an specific location, like the layer detail in registry.

@gsolat, can you share the design here?

• Auto update the url based on state of the search client.
• Auto update the client fields based on url.

The best way to show off is to create a query and share the resulting url via slack with another team member, that team member should be able to click on the url and have the same status show up on his computer.

Currently the keywords are passed as a string like this "Test 2" "Test" "Test Test" and is splitted up and stored as an array. but needs to be constructed and destructed everytime a keyword is added or removed.

Suggestions:
Create a service out of it, that stores the keywords as an array and can create a search term out of them.

1. Distance-scale indicator would be more visible if it appears inside the bounding box, and so it doesn't get hidden below the export-button.

Remember this is done with q_geo vs a_hh_filter.

It resizes to the right location when zooming in but not when zooming out.

Reset function should be executed as a general action to clear all the search. It should be inside the export-button section, and, it has to appear at the bottom of the tweet list.

• have just the Export button that opens the modal (to save space)
• have date picker open up widget on mouse click in text input (as opposed to having to click the small button beside said text box) for from and to
• enable the time selection to 'pan' a given timespan; assuming start/end are set, allow for sliding the entire range
• put map into view (gets cut off at least on a 13" MacBook)
• update empty instruction modal
• allow for left hand result list to scroll independently without scrolling the entire page

The time facets doesn't work when the range is not small.

q.time = [2010-01-01 TO 2013-04-01T00:00:00]
a.time.limit = 1

Response:

{
  "code": 500,
  "message": "Gap P1D is too small for this range [2010-01-01 TO 2013-04-01T00:00:00]"
}

Query link: http://54.158.101.33:8080/bopws/tweets/search?q.time=%5B2010-01-01%20TO%202013-04-01T00%3A00%3A00%5D&q.geo=%5B-90%2C-180%20TO%2090%2C180%5D&d.docs.limit=5&d.docs.sort=score&a.time.limit=1&a.time.gap=P1D&a.hm.limit=0&a.text.limit=0&a.user.limit=0

Make the search heatmap window always have 20px of padding on each side, when zooming in, out or panning.

Disable manual controls (but don't remove the functionality). We will need it again in the future once we figure out a good UI, but for now let's leave the query as a bit smaller window than the map size.

-a

Currently the keyword deletion process is a little bit complicated and not easy to understand.

#62 (diff)

We need to discuss a better way of doing it.