You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
fix(web): prevent writing large auspice json to local storage
After `dataset-json-url` param is used, the entire auspice json was stored in local storage. For large files this would be bigger that what browsers allow and the store operation would fail. Subsequent reads would retrieve either nothing or some fragment of data.
Here I add some logic to make sure we store auspice json in memory only. Sadly, this means that page reload wipes all the data and we currently don't have a good mechanism to store the previous dataset URL in order to re-fetch it. This will need to be implemented. Currently I opted for discarding the stored dataset if it was auspice dataset. This should be alright, as the feature was primarily designed to be used when navigating to Nextclade with url params set.
This PR only implements a workaround for something that's fundamentally poorly designed and broken. The full dataset description object is still saved to local storage. The definitive fix requires serious restructuring and altering many components. It will be implemented later (#1464)
But this workaround should allow most use-cases of `dataset-json-url` to work and makes this auspice dataset feature releasable.
Can be tested with a large tree like this:
```
?dataset-json-url=https://nextstrain.org/ncov/gisaid/global/all-time
```
0 commit comments