Smoking variable updates

[DougManuel]

This PR contains updates to just variables.csv and just for the smoking variables. I will add the updates to this PR for variable_details.csv shortly (hopefully).

Additional changes to smoking variables that are a lower priority and will be added to a later PR.

• I am also working on updates to the documentation for smoking variables, but this work will need coordination with @mdiaspar.
• Further, I will finish suggestions to the .R code for derived variables.

Changes to variables.csv (this PR) include:

• Essentially, all smoking variables and interview dates are updated because there are additions and changes to notes and description for all these variables.
• DatabaseStart to _m suffix - removed duplicates and typos.
• Fixed variableStart typos
• Standard units capitalization
• Added description and notes fields
• Updated version to 3.0.0-alpha
• Updated lastUpdated to 2025-10-26
• Added detailed reviewNotes
• Cleared ICES.confirmation fields. I don't think these are needed as we move toward a penultimate review.
• No changes to Observations.MD as these are @mdiaspar notes and she may need them for her review.

[DougManuel]

CSV Standardization and validation

@yulric is working on variables.csv and variable_details.csv. In the meantime, I used standardise_csv() in feature/csv-standardization-updates.

If you want to use them to verify the CSV formatting and structure:

# Source the csv-utils (from feature/csv-standardisation-updates branch)
source("R/csv-utils.R")

# Validate the updated variables.csv
result <- standardise_csv("inst/extdata/variables.csv", 
                         collaboration = FALSE,
                         validate_only = TRUE)

# Check results
print(result$valid)  # Should be TRUE
print(result$issues_remaining)  # Should be empty
Clean diff achievement
The preserve_column_order = TRUE parameter was used during preparation to ensure minimal diff noise (we have a different column order in our metadata yaml):
49 smoking variables updated (actual content changes)
No column reordering (would have shown 378 lines changed)
Only substantive changes are visible in the diff