Choose timezone
Your profile timezone:
HFLAV all-hands
→
Europe/Zurich
HFLAV all-hands
Meeting summary - according to Zoom AI ...
Quick recap
This was an all-hands meeting focused on presenting HLAB's new website project and discussing the migration to Zenodo for result storage. Mirco introduced the new website prototype, which is built on GitLab using Markdown and automatically renders tables and plots from Zenodo JSON files, replacing the current Drupal-based website due to CERN's discontinuation of Drupal. The new system aims to provide a single source of truth for all results, improve consistency across working groups, and allow automatic updates when new Zenodo records are uploaded. Ulrik presented the HFLAB API, which can search and access results through DOI references, and discussed improvements to the JSON schema including mandatory LaTeX names, comment fields for fits, and internal HFLAB identifiers for easier referencing. Working group members provided feedback on specific implementation challenges, particularly around naming conventions, handling complex observables like delta gamma S, and managing dependencies between different working groups' results. The discussion also covered how to handle PDG-specific results separately from general HFLAB averages, with considerations about whether to use separate repositories or tags for PDG content.
Next steps
Michael
- Contact Ulrik and Mirco privately to discuss obtaining a CERN account for accessing the GitLab project.
Mirco
- Finalize the new website project by September, ensuring migration from Drupal to GitLab-based site is complete.
- Implement a link from Zenodo records to the new website for human-readable display of results.
- Set up an e-group/mailing list for the new website project to collect feedback and coordinate work.
Olivier Schneider
- Work on extending the naming convention for measurements in the JSON schema.
Ulrik
- Develop a proposal for a unique internal HFLAV identifier for averages, possibly based on an automatic naming scheme.
- Consider and propose a solution for handling PDG-specific web pages (separate repository vs. tags).
- Extend the HFLAV API to allow reading records from the GitLab repository (work in progress).
Collaboration
- Each Working Group: Identify a person or few people to maintain and implement their section of the new website.
- Each Working Group: Review the new website prototype and provide feedback on functionality and improvements needed.
- Each Working Group: Work on updating their Zenodo JSON records to improve naming conventions, add LaTeX names, and include PDG codes where applicable.
Summary
HLAB Website Migration Project
Mirco welcomed new members to HLAB working groups and announced a new website project as a major focus for the year. The project aims to fully migrate results to Zenodo by October and launch the new website by November, with a September deadline imposed by CERN's discontinuation of the current Drupal framework. The new website will automatically generate tables and plots from a central repository, replacing the previous system with a GitLab-based solution.
Citation Policy and Data Presentation
The team discussed the citation policy and data presentation approach for their results. Mirco explained that a new website would be presented to make the data more usable and human-readable, with both a web interface and API available for automated access. Ulrik clarified that the Xenodu DOI links are permanent references guaranteed to stay available, while web pages may migrate over time, with metadata updates possible but not the entities themselves.
New Web Design Automation System
Mirco presented a new web page design using GitLab and Markdown that aims to improve automation and consistency across working groups. The new system automatically extracts data from Zenodo records to generate tables and plots, reducing manual copying and pasting while providing better organization and accessibility. The website structure focuses on direct access to results with minimal navigation, and Mirco demonstrated how different working groups can display their data through expandable tables and comparison features.
Scientific Results Display System Demo
Mirco demonstrated a new system for displaying scientific results and figures on web pages, emphasizing the need to automate data extraction and ensure consistency with source units. He showed examples of how different working groups could present their data, including tables, plots, and explanations, with specific formatting requirements for text and mathematical content. Mirco noted that while the system is flexible and can accommodate various display styles, working groups would have the option to modify the presentations to suit their needs.
Particle Physics Website Structure
Mirco presented a new website structure for particle physics data, explaining how it organizes information by three systems with channels and inputs from different measurements. He outlined the workflow for creating new editions, which involves working on Xenon updates first, then mirroring the structure and making necessary changes to links and tables. Mirco encouraged working group members to review the draft website, provide feedback on its functionality and flexibility, and suggested setting up a mailing list for further discussions.
Webpage Prototype Feedback Discussion
Mirco presented a prototype webpage with 37 observables, noting that other working groups successfully handle up to 150 observables. The current implementation has issues with rendering LaTeX names and displaying all relevant parameters, particularly for FIAS parameters which are not yet clickable. Mirco suggested that each working group identify specific people to serve as points of contact for feedback and maintenance rather than having all members interact directly with the code.
Subgroup Web Page Implementation
Olivier Schneider and Mirco discussed the implementation of web pages for subgroups, focusing on working with JSON files and Zenodo records. Mirco confirmed that web pages can work with JSON files that are not in Zenodo records, including local files, and explained that ordering and grouping of results should be handled through helper functions rather than in the Zenodo file itself. Mirco clarified that the grouping of direct and indirect measurements, as well as the ordering of results, is independent of the Zenodo record and can be customized at the webpage level using helper functions.
JSON Implementation Feedback Discussion
Mirco and Thomas discussed feedback on Mirco's work implementing JSON format and web functionality, with Mirco highlighting the need to distinguish between internal and public-facing comments in the code. They explored potential improvements including making plots downloadable with HLF logos and adding copy functionality for correlation matrices. Michael raised concerns about access requirements, noting that non-CERN account holders cannot currently access the private website, though Ulrik clarified that while public access requires CERN IT approval, users can build the website locally or contact Ulrik privately about obtaining CERN accounts.
Synodo Records Integration Updates
Ulrik presented updates on the Synodo records integration into PDD, highlighting that specific references are now included in the latest PDD version to ensure traceability of measurements. He discussed ongoing issues with the Synodo records, including incomplete data, unclear naming conventions, and the lack of an API, and proposed solutions such as extending naming conventions and making certain fields mandatory. Ulrik also introduced the new HFLAB FAIR client API, demonstrating its functionality for searching and retrieving data, and suggested its potential use for creating comparison tables with PDG results.
Internal H.Flav Identifier System Implementation
The team discussed implementing an internal H.Flav identifier system for averages to facilitate their export to LaTeX for their paper. Olivier explained that unique identifiers are needed to reference specific averages in LaTeX, particularly for subgroup variations like Delta Gamma S. Ulrik proposed automatically generating unique names based on the full path and parameter names in the JSON file, which would be more convenient than manually creating names. The discussion also covered how to handle references to averages from other working groups through the API, with Ulrik noting that a dependency tree system is still in development to address potential circular references.
HFLAB Results Management Discussion
The team discussed improving the use of HFLAB results and reducing reliance on external PTT results. Mirco suggested implementing staggered updates for working groups to better manage dependencies between results. The group also addressed the need to differentiate between general HFLAV results pages and PDG-specific pages, with Olivier Schneider raising questions about how to present PDG-recommended values on web pages. Ulrik proposed either using separate tags in the same Git repository or creating entirely separate repositories for PDG results, though he acknowledged needing more consideration on the best approach given the complexity of PDG codes referring to specific measurements and branching fractions.
There are minutes attached to this event.
Show them.