Dynamic error cdi; COUNTRY=jordan #1259

lowriech · 2024-05-28T20:35:41Z

@wadhwamatic check it out, pretty simple updates to parameterize a local hip service and to utilize a detail error message

lowriech · 2024-05-28T20:35:59Z

Goes with: https:/WFP-VAM/hip-service/pull/11

github-actions · 2024-05-29T11:47:13Z

Build succeeded and deployed at https://prism-1259.surge.sh
(hash d61911f deployed at 2024-07-02T18:23:29)

README.md

wadhwamatic · 2024-05-29T19:05:23Z

@lowriech - the date error message correctly appears when the latest date (11-May) is requested. But subsequent requests are not rendering when data is available even though the response is the expected geojson object. It would also be good to test this and #1253 together.

into dynamic-error-cdi

lowriech · 2024-05-31T00:42:50Z

@wadhwamatic I'm not seeing the error where "subsequent requests are not rendering when data is available even though the response is the expected geojson object". Can you clarify?

Merged in range update PR to this PR.

wadhwamatic · 2024-05-31T05:37:55Z

@lowriech - what I meant is:

If you request CDI on a date that doesn't have data, you see the error message you created
But if you make a second request one another date which does have CDI data, the geojson response is not rendered on the map even though I see in the dev console that the object was returned
It seems that the first error prevents a subsequent valid request from rendering

lowriech · 2024-06-04T16:16:41Z

Can you send the environment you're using to look at this? That's not what I see, I have updates as usual after the error. @wadhwamatic

wadhwamatic · 2024-06-18T22:24:46Z

@lowriech - I assigned two issues to @paololucchino to 1) modify the error message text and 2) to create an endpoint for available dates before a CDI layer is loaded.

What we have so far works, but is less than ideal since a user has no idea what dates are going to be valid. I'm wondering for now if we should just hack this a bit and assume the latest available CDI will be today - 1 month.

@ericboucher, what do you think about this approach given that we don't have a way to know available dates for CDI as of now. If this sounds good, maybe @gislawill can take on implementation.

frontend/src/utils/fetch-with-timeout.ts

ericboucher · 2024-06-19T14:17:01Z

@wadhwamatic can you share a case where the data is not available for CDI?

README.md

wadhwamatic · 2024-06-19T15:16:06Z

@wadhwamatic can you share a case where the data is not available for CDI?

@ericboucher - Jordan CDI on 1-June-2024 returns a date error. https://prism-1259.surge.sh/?hazardLayerIds=cdi_v1&date=2024-06-01

gislawill · 2024-07-01T20:36:16Z

@wadhwamatic, @lowriech, @ericboucher — I've made an attempt to improve this and would appreciate your thoughts. The latest available date for analysis currently is May 1st (exactly two months ago). To hack this before WFP-VAM/hip-service#12, I added an "expectedDataLagDays" option to the composite layer that shifts the default date.

Currently, the default date is the latest available date that that's returned from the server as SPI's latest available date.
This update, adds an expected lag of 55 days so the latest available date is the latest date returned from the server as SPI's latest available date and is more then 55 days in the past
After WFP-VAM/hip-service#12, this date should come from the backend and we can remove this logic

gislawill · 2024-07-01T20:50:43Z

frontend/src/utils/useDefaultDate.ts

 // TODO - use getPossibleDatesForLayer
+ // Update on 2024-07-01: getPossibleDatesForLayer would give us the first possible layer date, we want the _last_ possible date.


Note here, I don't think this to do is still relevant. Either that or we should refactor how composite layers are handled in getPossibleDatesForLayer

wadhwamatic · 2024-07-02T04:37:56Z

@gislawill - thanks for diving into this. The logic you described below all sounds good to me for our temporary solution until hip-service handles date availability.

Currently, the default date is the latest available date that that's returned from the server as SPI's latest available date.
This update, adds an expected lag of 55 days so the latest available date is the latest date returned from the server as SPI's latest available date and is more then 55 days in the past
After WFP-VAM/hip-service#12, this date should come from the backend and we can remove this logic

Ideally, we would also use the date lag definition to hide unavailable dates in the timeline. In this case, we'd want to show available date ticks that are more than 55 days in the past for the CDI layer. But if that is going to add a lot of extra work, we can skip it for now.

gislawill · 2024-07-02T14:46:18Z

@wadhwamatic sure thing, I added that update here: #1295

I separated it from this PR because I'm not confident the improvement in user experience is worth the risk of bugs + complications to maintainance we'd be introducing. If https:/WFP-VAM/hip-service/issues/12 is coming soon and this UX improvement, I believe that solution (in #1295) is functional and addresses the issue. It's just pretty unintuitive code so I wouldn't want it in the codebase for now

gislawill · 2024-07-02T18:02:35Z

Heads up, after discussing with Amit, I pulled all of my commits from this branch to #1295

That allows it to be used for demos without adding this date logic complication to the repo

gislawill · 2024-07-02T18:15:28Z

This should be ready for review again @wadhwamatic and @ericboucher

ericboucher · 2024-07-03T15:30:24Z

ok to merge into the other PR but I would prefer to remove the REACT_APP_USE_LOCAL_HIP_SERVICE logic before merging to main

* Updating to use month range * Fix date propagation in CDI * Update composite_data.ts * Use getFormattedDate * Update composite_data.ts * Update composite_data.ts * add mozambique cdi for testing * Update index.tsx * Dynamic error cdi; COUNTRY=jordan (#1259) * Updating to use local hip service * no console * Clean up * Clean up readme --------- Co-authored-by: Christopher J Lowrie <[email protected]> Co-authored-by: ericboucher <[email protected]> Co-authored-by: Amit Wadhwa <[email protected]> Co-authored-by: Will Gislason <[email protected]> * Removing local hip service env var * Error handling fixes and end date support form the lyer * Initial support for seasonal CDI * Remove console log, update configs for tests to pass * Appease eslint * Fixing seasonal fetch - using first of next season as 'end' * Support for validity areas in composite layers * Reducing date deduping * Clean up * refactor datesAreEqualWithoutTime * Fix datesAreEqualWithoutTime * Adding 'season' validity * Skip the tests * Fixing query date for seasonal range * Updating CDI config in cambodia and mozambique * Update tests * undo log update * update mozambique CDI for testing * Update composite monthly date range * Handling overlap selection for dates without overlapping validity * fix test * rename scale to period * Updates to date picking * Allow composite layers with validity or without * Update the query display matching to be more flexible * Updating event handlers in time selection * Test fic * Better test fix * Break out visible layer dates and selectable layer dates * don't fetch CDI for invalid dates * fix tests * Fixing availability date indicator * Ensuring messaging fires * fix lint * Reuse clickDate for draggable item * Update index.tsx * Update datasetStateSlice.ts * Update index.tsx * Adding back no date overlapping handling * Ensuring that date selection warning is displayed when the date jumps to valid date * Remove check intersection, leveraging logic in layer-utils * using in datesAreEqualWithoutTime in findDateIndex * Update layers-utils.tsx * Only remove non-boundary layers * Handling AA dates with fewer limitations * Add comment to code * Fix extra visible dates for AA * Update logic in useLayer's useEffect to ensure we don't exit early * Fix AA parseFloat * Ensuring group is writable in formatLayersCategories * fixing eslint --------- Co-authored-by: Christopher J Lowrie <[email protected]> Co-authored-by: ericboucher <[email protected]> Co-authored-by: Amit Wadhwa <[email protected]> Co-authored-by: Chris Lowrie <[email protected]>

Updating to use local hip service

75cbbce

no console

1920311

ericboucher reviewed May 29, 2024

View reviewed changes

README.md Outdated Show resolved Hide resolved

ericboucher changed the base branch from master to cdi-range-update May 29, 2024 15:12

Christopher J Lowrie added 2 commits May 30, 2024 17:36

Adding in range update

5ecbc6c

Merge branch 'dynamic-error-cdi' of https:/WFP-VAM/prism-app

da034b8

into dynamic-error-cdi

Merge branch 'cdi-range-update' into dynamic-error-cdi

1992d99

Merge branch 'cdi-range-update' into dynamic-error-cdi

60a4bce

wadhwamatic requested a review from gislawill June 18, 2024 22:19

gislawill reviewed Jun 18, 2024

View reviewed changes

frontend/src/utils/fetch-with-timeout.ts Outdated Show resolved Hide resolved

ericboucher changed the title ~~Dynamic error cdi~~ Dynamic error cdi; COUNTRY=jordan Jun 19, 2024

Merge branch 'cdi-range-update' into dynamic-error-cdi

f9be7db

ericboucher reviewed Jun 19, 2024

View reviewed changes

README.md Outdated Show resolved Hide resolved

Merge branch 'cdi-range-update' into dynamic-error-cdi

df5a171

gislawill self-assigned this Jun 25, 2024

gislawill reviewed Jul 1, 2024

View reviewed changes

gislawill mentioned this pull request Jul 2, 2024

Restricting the available days when data lags expected; COUNTRY=jordan #1295

Closed

7 tasks

gislawill force-pushed the dynamic-error-cdi branch from d902875 to df5a171 Compare July 2, 2024 17:54

gislawill added 2 commits July 2, 2024 13:13

Clean up

20450e7

Clean up readme

d61911f

wadhwamatic requested a review from ericboucher July 2, 2024 18:27

ericboucher merged commit f078b81 into cdi-range-update Jul 3, 2024
6 checks passed

ericboucher deleted the dynamic-error-cdi branch July 3, 2024 15:30

gislawill mentioned this pull request Aug 6, 2024

Hack: restrict CDI date selection for lag; COUNTRY=jordan #1326

Draft

7 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Dynamic error cdi; COUNTRY=jordan #1259

Dynamic error cdi; COUNTRY=jordan #1259

lowriech commented May 28, 2024

lowriech commented May 28, 2024

github-actions bot commented May 29, 2024 •

edited

Loading

wadhwamatic commented May 29, 2024

lowriech commented May 31, 2024

wadhwamatic commented May 31, 2024

lowriech commented Jun 4, 2024

wadhwamatic commented Jun 18, 2024 •

edited

Loading

ericboucher commented Jun 19, 2024

wadhwamatic commented Jun 19, 2024

gislawill commented Jul 1, 2024

gislawill Jul 1, 2024

wadhwamatic commented Jul 2, 2024 •

edited

Loading

gislawill commented Jul 2, 2024

gislawill commented Jul 2, 2024

gislawill commented Jul 2, 2024

ericboucher commented Jul 3, 2024

		// TODO - use getPossibleDatesForLayer
		// Update on 2024-07-01: getPossibleDatesForLayer would give us the first possible layer date, we want the _last_ possible date.

Dynamic error cdi; COUNTRY=jordan #1259

Dynamic error cdi; COUNTRY=jordan #1259

Conversation

lowriech commented May 28, 2024

lowriech commented May 28, 2024

github-actions bot commented May 29, 2024 • edited Loading

wadhwamatic commented May 29, 2024

lowriech commented May 31, 2024

wadhwamatic commented May 31, 2024

lowriech commented Jun 4, 2024

wadhwamatic commented Jun 18, 2024 • edited Loading

ericboucher commented Jun 19, 2024

wadhwamatic commented Jun 19, 2024

gislawill commented Jul 1, 2024

gislawill Jul 1, 2024

Choose a reason for hiding this comment

wadhwamatic commented Jul 2, 2024 • edited Loading

gislawill commented Jul 2, 2024

gislawill commented Jul 2, 2024

gislawill commented Jul 2, 2024

ericboucher commented Jul 3, 2024

github-actions bot commented May 29, 2024 •

edited

Loading

wadhwamatic commented Jun 18, 2024 •

edited

Loading

wadhwamatic commented Jul 2, 2024 •

edited

Loading