[theHenks]

• I have reviewed all the code in this pull request and deem it of the standard expected for this project.

Summary

runinfo of a dataset

• For a dataset other than :default, runinfo holds what the run lists of the dataset in datasets/runlists name: one column per listed category, and an entry of a category only for the runs listed for it. Any other entry has missing start and end keys, no DAQ cycle keys, a NaN livetime and is_analysis_run = false. Selecting a category that the dataset does not have throws an ArgumentError. :default keeps every run and category of datasets/runinfo.
• get_exposure skips the runs without a start key of the category, and with is_analysis_run = true (the default) the runs that are no analysis runs.
• read_ldata of a period, partition or run table skips the runs whose entry of the category has no DAQ cycle keys, also with search_disk = true, and throws if no run is left. Reading a single run is unchanged.
• find_filekey(data, ts) searches the DAQ cycle keys of every run in the metadata, so a timestamp in a cycle outside the dataset resolves to its own cycle.

Conveniences

• filter(f, ds::DataSet) returns a DataSet, e.g. filter(!in(bad_filekeys(l200)), DataSet(l200, runinfo(l200).phy)). The display of a set built from runinfo shows per category the summed runinfo livetime of the runs with keys in the set, and the runtime from the first to the last key.
• find_filekey.(data, timestamps) broadcasts over timestamps.
• A PropsDB of a LegendData takes a Unitful.Time as validity selection, e.g. l200.par.ppars.aoe(t), and selects the file key of the DAQ cycle that contains t.
• read_ldata(data, tier, t) and read_ldata(data, tier, t, det) with a single Unitful.Time read as [t].
• read_ldata(data, tier, ds::DataSet) and read_ldata(data, tier, ds, det) read all file keys of the set.

Behavior changes

• Under a named dataset, a run listed only for some of its categories has no keys, livetime or analysis flag in the others, and categories the run lists do not name are no columns.
• get_exposure with the default keywords counts only analysis runs. Under :default, runs not in the valid run list no longer count.
• is_lrun and is_analysis_run throw for a category that the dataset does not have.

Since results change, I suggest releasing this as 0.7.0. p_process_skm_phy in JuLeAna needs to skip runs without a phy start key before JuLeAna moves to this version.

Tests

New testsets "runinfo of a dataset", "DataSet" and "selection by a time" (test_dataprod_config.jl), "Exposure of the runs of a dataset" (test_exposure.jl), and "runs of a dataset" plus single-timestamp reads (test_ext_legendhdf5io.jl). Pkg.test() passes locally (732 passed, 1 broken as on main).

Parts of this pull request, including code, tests and this description, were written with the help of generative AI.

[codecov]

Codecov Report

❌ Patch coverage is 92.72727% with 4 lines in your changes missing coverage. Please review.
✅ Project coverage is 63.75%. Comparing base (d854b60) to head (15606a7).

Files with missing lines	Patch %	Lines
ext/LegendDataManagementLegendHDF5IOExt.jl	77.77%	2 Missing ⚠️
src/props_db.jl	85.71%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main     #246      +/-   ##
==========================================
+ Coverage   63.25%   63.75%   +0.50%     
==========================================
  Files          27       27              
  Lines        2814     2842      +28     
==========================================
+ Hits         1780     1812      +32     
+ Misses       1034     1030       -4     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.