-
Notifications
You must be signed in to change notification settings - Fork 31
Conversation
[Resolves fecgov#289]
Borrowed from fecgov#302.
7c80af2
to
0ac41a7
Compare
Note: I'm keeping lots of the utility code that was used for this filter, since we're going to be using it for other views.
…/committee-file-date
This has been updated per Slack / GitHub discussion @noahmanger @LindsayYoung. |
Borrowed from defunct code in fecgov#312.
Pinging @noahmanger @LindsayYoung for review. After we talked this through last week, the patch became very simple and should be quick to review--it basically just adds a sortable column on first file date, and a date formatter helper for the frontend. There's also some utility code for calculating the current month, quarter, and year, which isn't used here but will be in #302. |
Looks good, but the problem is that there's no way to see the most recent because there's a bunch of committees with blanks in that column. How should we handle this? |
We could change the sorting logic to filter out |
We could also have the database treat |
Any preferences @noahmanger @LindsayYoung? I think omitting records with null values on the sorted column is reasonable, but maybe there are better solutions. |
Based on discussion in fecgov#312.
@noahmanger the point you brought up about empty rows will be resolved once #322 is merged. |
There's quite a few JS libs added in this code. I know I hadn't completely set this up yet before I moved to less time on this project, but ideally, when theres JS libs added, I'd like to see what the impact would be. Like how many kbs added, and how that effects things like time to first paint, speed index, or custom metrics. Although I understand this is hard to do now, just wanted to comment as a good place to move to. |
Good looking out @msecret . What do you all think about hiding the "Treasurer Name" column? @LindsayYoung I think this was a suggestion you made a while back? |
The libraries added here should each be < 30 kB, minified and gzipped. If we're looking to optimize load times, I'm guessing we'd get the most traction from splitting the browserify bundle into pieces--loading d3, handlebars, and datatables on the homepage (which uses none of the above). |
Should we make a separate issue for looking at performance and making these sorts of changes? |
We have an issue for splitting code bundles: https://github.com/18F/openFEC/issues/908. We also have performance tracking on dev, which we should be able to track with New Relic (is that right @msecret?)--maybe we need an issue for gathering / evaluating those performance numbers. |
I'd like to check out New Relic to see what performance tracking they have implemented. So the thing with performance that I've gathered: the best metric to really get a sense of how your users are experiencing the site is SpeedIndex. It's also the hardest to collect, essentially requiring a webpagetest.org instance running to capture a video of the site loading. This means that often an easier to measure metric, that also comes closer emulates the SpeedIndex metric is a custom metric. For us on FEC, this would be different on each view/page, and could be things like, "time till the whole results table shows", "time till search bar shows", "time till search results appear", etc. I think defining these custom metrics and coming up with what our goal for them would be a good step forward for tracking performance. |
I have mixed feelings on Treasurer name- it is very important since these are the people legally responsible, but it is not necessary everywhere. |
Yeah, I share that. If we're going to keep it we should also have a filter On Mon, Jul 13, 2015 at 10:58 AM, Lindsay Young [email protected]
Noah Manger |
Based on discussion in fecgov#312.
This looks good to me, but @msecret should probably review the js / python if you haven't already. |
…/committee-file-date
This should be good to merge @msecret / @noahmanger --it's just adding a column and updating some tests. |
Add file date filter and column.
[Resolves #289]
This is working as expected, but probably needs more discussion; see #289.