Tags

The cookie value in the download endpoint is now optional and so is a query parameter rather than part of the path

Filtered out vehicle types with a detail_id=0 - since they were causing problems and are not needed

Added spatial filtering (rectangle) and updated temporal filtering

Added startYear and endYear query parameters to the download endpoint

Updated the download code to correct issues leading to over-looping and incorrect values in download file.  The output has now been validated against test data.

The legend endpoint of the api now returns pre-prepared json data about quantile classifications rather than calculating them on the fly.

Added cookie handling to download endpoint, which is needed to implement the spinner and disabled behaviour on the Download button of the app

Added logging to track status of download generation.
Also increased gevent timeout from 30s to 300s

The download netcdf file now has dynamically generated metadata to show what was used to create the file based on the user's selection.

Added erddap persistent volume to the helm chart so the api can get hold of the individual activity grids to build dynamic downloads.

Also upgraded flask to 2.0.1 to get past the error shown below that occurred during startup - the solution was on the stackoverflow: https://stackoverflow.com/questions/71189819/python-docker-importerror-cannot-import-name-json-from-itsdangerous

The error was:
ImportError: cannot import name 'json' from 'itsdangerous' (/usr/local/lib/python3.9/site-packages/itsdangerous/__init__.py)

The download endpoint now provides a netcdf file download for a pollutant.  The download for a pollutant is either taken from pre-prepared data or generated dynamically.  Pre-prepared data is used when a user requests a download for either Total Emissions or for a Snap Sector.  A dynamic download is prepared using raw Ricardo activity grids and emission factors when the user requests data that includes sub-Snap-Sector options - which is currently only for Road Transport (things like vehicle type and driving activity type).

Additional endpoints have been added to provide data to dropdowns on initialisation.  This allows easy changing of data as new pollutants and snap sectors are added.

Added a stack info endpoint to proved the data needed by the point source placemarker popups on the timeseries map of the web-app

Updated to support timeseries and legend of snap datasets, redundant code removed

Point source endpoint added that supports the web-app's leaflet-marker-clustering (https://jira.ceh.ac.uk/browse/DUKEMS-165).  When a user clicks a placemarker underlying a marker-cluster on the map, a request for point source data from the api is used to populate the popup that opens.

Added additional logging to legendemissions endpoint to try to identify where bottleneck is occurring

Adjusted property names on endpoint that currently provides vehicle type classes for the dropdown.  The requirements on the interface had changed with the use of a new multi-select dropdown.

Added logging to the endpoints so we can see when they are visited

Re-instated wsgi, but with gevent to monkey patch gunicorn with ansychronous super powers

Added the changes to deployment.yaml that make gunicorn use gevent in the run command - this was accidentally not added to the previous commit and corrects that.

- added gevent asynchronous request handling
- put pip installs into requirements.txt
- reduced the number of times a filter is applied to the flat data in download.py