Skip to content

Update, review, tidy and fix across s/scripts and Docker setup#86

Open
pacharanero wants to merge 43 commits into
livefrom
update-review-fix-scripts-docker-compose
Open

Update, review, tidy and fix across s/scripts and Docker setup#86
pacharanero wants to merge 43 commits into
livefrom
update-review-fix-scripts-docker-compose

Conversation

@pacharanero
Copy link
Copy Markdown
Member

@pacharanero pacharanero commented Jun 2, 2026

Overview

  • BASED ON PR Who-validation #80 - I would recommend merging that first - it was not possible to base it on current live, as some of the changes in Who-validation #80 were ones I fixed here.
  • Had a couple of issues locally running scripts and in the process of debugging them I did a bit of a general housekeeping session and update to keep them all working and harmonised.
  • Changes are restricted to the s/scripts and the Docker/Compose setup.

…ameter which defaults to False, but applies linear interpolation across the board if true. Adds comparisons between who daily and uk_who calculated values
…, monthly for >5y as json, csv folder for the who csv files. adds a new function to parse the csv files either into individual json files or 3 combination files as before. deprecates old who json files
…ting chart function bmi against gold standard charts
- Updated comments for clarity regarding age thresholds.
- Simplified the logic for selecting WHO reference data for children aged 2.0 and below.
- Removed redundant checks and print statements related to reference selection.
- Ensured that the correct reference is selected based on the age and default_youngest_reference flag.
- Improved error handling for missing reference data by raising a LookupError with a clear message.
- fixed BMI failing tests with exception of +3.09
…male), bmi (male/female), length (male/female)
… test data out to separate file. now has over 5y tests for centile line calculation for length (boys/girls), weight (girls)
…easurement for plotting pass, when testing against the published WHO line coordinates. Oddly though they only pass if we DON'T use the convoluted calculations that apply a correction above +3 and below -3 SDS for BMI.
…tiles. Note that this involves disabling the WHO calculations >-3 and <-3 SD. We have no tests for measurement->SDS for this so these will need to be added
…d from WHO published anthro and anthroplus packages forked to RCPCH and amended to increase precision to 3dp. Chart functions passing (though >5y excluded from under 5s published curves). We are now including in <5s actually upto 1856 days, and taking WHO 2007 from 61 months (where the data starts from)
eatyourpeas and others added 13 commits February 21, 2026 18:21
- Create AGENTS.md with comprehensive guidance for AI/LLM agents
- Add LLM reference to README.md for discovery
- Fix docker-compose.yml: use background execution for JupyterLab
- Improve s/test script for container management
- Add s/notebook convenience script for JupyterLab launch
- Create docs/LIVE_DATASET_FAILED_TESTS_SUMMARY.md analyzing fixture differences
- Add deprecated fixture for regression testing reference
- Update test infrastructure for WHO validation branch

These changes ensure the who-validation branch is ready for development
with proper documentation for both human and AI developers, and a
working Docker-based development environment with convenient scripts.
- add WHO reference implementation notes and rationale
- rename the UK-WHO integration test file and test function
- update README, AGENTS, and failure-summary references
- rename the validation fixture to reflect the refactored dataset
… of fixing s/notebook - reviewed and tidied all the scripts while I'm here.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants