Issues/names other issues

[keatonwilson]

⏺ Summary

• Integrate meta table (authorship metadata) into all four collate_*() functions via left_join on dataset_name, making author names and ORCID IDs
  available in collated output
• Improve collate_soil() categorization — use sensor_location from data_description table instead of heuristics based on individual_id/plot_id
  values, making the individual/plot/study split more robust and explicit
• Push filters into the database — collate_met() now uses .env$dataset_name for proper tidy evaluation and performs joins lazily inside DuckDB
  before a single collect(), improving performance
• Prevent duplicate columns — detect and drop overlapping columns between study_site and meta before joining to avoid .x/.y suffixes
• Suppress lutz warnings — wrap tz_lookup_coords() calls in suppressWarnings() across all collate functions
• Fix test database paths — use test_path("../..") in tests so get_db_path() resolves correctly regardless of working directory
• Update vignettes — document the meta table, update core table count from 6 to 7, and note dataset_name column in flag tables
• Clean up renv.lock — remove the lock file from version control (dependencies managed elsewhere)

Test plan

• Verify collate_met() returns meta columns (e.g., submitting_author_first_name)
• Verify collate_chamber_wp() returns meta columns
• Verify collate_auto_wp() returns meta columns
• Verify collate_soil() returns meta columns and sensor_location at all three levels
• Verify sensor_location values match expected levels (Individual, Plot, Whole study)
• Verify no duplicate columns in joined output
• Run devtools::check() to confirm no regressions

[jessicaguo]

Oops, pressed too soon

[keatonwilson]

@jessicaguo merging - will deal with meta table rename in subsequent PRs here and on PSINetDB.