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
We made some important updates to IF in the last fortnight.
I have been working on creating manifest files for our own websites, and also started trying a scope 3 LCA using Impact Framework. Generating manifests for real projects has been a good test of our product - I uncovered some fairly subtle but important bugs that led to some feature redesigns, and some tweaks to try to reduce friction for the manifest writer.
The most important change is the separation of the aggregation method for each parameter into components for time and component, making it possible to aggregate by averaging within a time series and then sum across multiple time series in a tree. This is really important for metrics such as SCI that are expressed as rates - if we have to aggregate using the same method within a time series and across multiple time series then we will always under or overestimate the true value. The 'split aggregation' fix resolves this issue and makes it much easier to configure SCI calculations for multi-component trees.
This change is merged into our main branch but not yet included in a release.
We also added support for inline arithmetic. This will help to make manifests more concise, since simple arithmetic operations don't necessarily require their own dedicated instance of a plugin anymore - instead you can do the arithmetic inline in the manifest instead, such as passing sums as arguments to plugins or as values in input data (e.g. a plugin parameter can now accept an expression such as some-value * 3 .
This will be available in the main branch very soon but it is currently awaiting an update in a dependency repository before we merge it in..
We also had a substantial feature added by @jamescrowley that allows a computed manifest to be re-run with new config and the results appended to the existing output data rather than overwriting it. This is going to be really useful for people that want to use the same manifest for repeated observations - maybe you want to observe daily but generate one manifest per month - well now you can just use the --append feature to build up your monthly report. I can imagine this being extra useful in combination with the if-merge feature.
The embodied carbon plugin got an overhaul too - now we mimic the CCF model for embodied carbon, taking some basic information about the server setup as arguments.
Other updates include reworking the explainer feature so that the response is primarily ordered by parameter rather than by plugin, which makes the list more easily searchable.
We are also putting the finishing touches on a new manifest visualizer that runs in the browser and allows to you more easily interrogate your output files.
Updates
Here's the links to more details about the most significant changes made over the last fortnight.
Who
Sponsor: @jawache (GSF)
Product owner: @jmcook1186 (GSF)
Leads: @navveenb (Accenture), @srini1978 (Microsoft)
Overview
We made some important updates to IF in the last fortnight.
I have been working on creating manifest files for our own websites, and also started trying a scope 3 LCA using Impact Framework. Generating manifests for real projects has been a good test of our product - I uncovered some fairly subtle but important bugs that led to some feature redesigns, and some tweaks to try to reduce friction for the manifest writer.
The most important change is the separation of the aggregation method for each parameter into components for
time
andcomponent
, making it possible to aggregate by averaging within a time series and then sum across multiple time series in a tree. This is really important for metrics such as SCI that are expressed as rates - if we have to aggregate using the same method within a time series and across multiple time series then we will always under or overestimate the true value. The 'split aggregation' fix resolves this issue and makes it much easier to configure SCI calculations for multi-component trees.This change is merged into our
main
branch but not yet included in a release.We also added support for inline arithmetic. This will help to make manifests more concise, since simple arithmetic operations don't necessarily require their own dedicated instance of a plugin anymore - instead you can do the arithmetic inline in the manifest instead, such as passing sums as arguments to plugins or as values in input data (e.g. a plugin parameter can now accept an expression such as
some-value * 3
.This will be available in the
main
branch very soon but it is currently awaiting an update in a dependency repository before we merge it in..We also had a substantial feature added by @jamescrowley that allows a computed manifest to be re-run with new config and the results appended to the existing output data rather than overwriting it. This is going to be really useful for people that want to use the same manifest for repeated observations - maybe you want to observe daily but generate one manifest per month - well now you can just use the
--append
feature to build up your monthly report. I can imagine this being extra useful in combination with theif-merge
feature.The embodied carbon plugin got an overhaul too - now we mimic the CCF model for embodied carbon, taking some basic information about the server setup as arguments.
Other updates include reworking the
explainer
feature so that the response is primarily ordered by parameter rather than by plugin, which makes the list more easily searchable.We are also putting the finishing touches on a new manifest visualizer that runs in the browser and allows to you more easily interrogate your output files.
Updates
Here's the links to more details about the most significant changes made over the last fortnight.
Issues
We have had a good fortnight, with no notable blockers.
Outlook
The text was updated successfully, but these errors were encountered: