Skip to contents

Changelog

Source: NEWS.md

creelutils 0.3.0

Released 2026-09-03

Large update to how catch groups are queried in from the database and how model estimates are uploaded back the creel database.

New features

  • ETL now allows estimates of “combined” catch groups to be uploaded to the creel database (e.g., Coho_Adult|Jack_AD|UM_Kept). Previously, CreelEstimates could process them but the database model estimates tables and the ETL were incompatible. See #35 and companion update at CreelEstimates#83.

  • Updated fishery_catchgroups() queries model catch group lookup table and can internally filter to only catch groups with observed catch. The new argument ‘observed_only’ (default FALSE) leverages an internal helper .observed_catch_group_ids() rather than in a separate step (#35).

  • New combine_catch_group() streamlines converting catch group component fields(i.e., species, life stage, fin mark, fate) into the concatenated form separated by underscores (e.g., Coho_Adult_AD_Kept) (#35).

  • Updated transform_estimates() to further standardize naming conventions used across PE and BSS model types for the fields ‘estimate_category’ and ‘estimate_type’. Several ‘estimate_type’ categories were dropped from the stratum and total tables. Those removed from the total table distracted from primary intent of being a high-level catch/effort summary. Those removed from the stratum table included BSS model convergence diagnostics that were only ever evaluated at the total level, leading to thousands of rows per analysis_id that were never looked at. The volume of rows for BSS stratum estimates has subsequently decreased by ~30-40% (#35).

  • New finalize_analysis_lut() consolidates several steps that were occurring within export_estimates() into a single process that extends from the analysis session-specific metadata created by CreelEstimates’ generate_analysis_lut(). Several additional metadata columns were also added to the ‘analysis_lut’ (#35).

Minor improvements and bug fixes

  • export_estimates() now calls write_lut, write_stratum, and write_total with DBI::dbWithTransaction(). This ensures that all rows are successfully appended to the database tables as a batch. If expected row counts differ or the database connection drops during the transaction, all three tables are rolled back as a unit.

  • New internal function .resolve_conn() streamlines how lazy data connections are opened and how stale connections passed into a querying function are handled. Previously fetch_db_table() and its wrappers independently validated connections from connect_creel_db() (#35).

    • A stale or closed connection provided to an argument will now produce an informative error rather than opening a connection to ‘prod’ by default. This prevents a bug where a connection to ‘test’ could silently change to a connection to ‘prod’ that produced confusing results.

Documentation enhancements

  • Add “Git Workflow & Conventions” article (#39). This describes basic coding conventions we have adopted, shows examples of routine usage, and covers version releases and continuous integration.

creelutils 0.2.1

Released 2026-08-20

  • Add instructions to fetch_data() for setting up optional Socrata API token for https://data.wa.gov to prevent occasional throttling (#38).

  • fishery_lut() helper function now has a ‘fishery_name’ argument for filtering to a specific fishery rather than returning the full table (#34)

creelutils 0.2.0

Released 2026-06-02

New features

  • Overhaul database connectivity and streamline user experience

    • Store user credentials in their OS credential manager with the keyring package (#19).

    • Relocate local config.yml from working directory of package creelutils is being called from to a user-level dotfile stored at ~/.config/creelutils (#22).

    • Implement lazy database connections for fetch_db_table() and its wrappers in R/query_helpers.R (#26).

  • New fetch_data() serves as a generalized method of querying raw datasets from either the Postgres database directly or from public views hosted at https://data.wa.gov (#29, #30).

    • Previously fetch_dwg() only allowed downloading from the public data source. The internal option is a more direct approach that does not have the limitation of waiting for the public views to sync on the preset schedule. Both paths now return type-identical output (the internal route includes several additional metadata columns), so database-sourced data feeds cleanly into the downstream CreelEstimates Stan models.

  • New plot_zipcodes() produces heat maps of angler-reported ZIP codes for individual states and full US (#18).

  • New fishery_catchgroups_obs() leverages fishery_catchgroups() and the raw catch table to produce a tibble of reference catch groups of interest for a given fishery filtered to just those with observed catch data (#20).

  • New catchgroups_to_params() helps convert from catch group component fields (i.e., species, life_stage, fin_mark, fate) into the vector format used by wdfw-fp/CreelEstimates/template_scripts/fw_creel.Rmd and its YAML param params$est_catch_group (#28).

Minor improvements and bug fixes

Deprecation

  • Removed experimental function render_progress_report() (#17).

  • fetch_dwg() was folded into fetch_data() and now serves as a wrapper for the new generalized data querying function (#29).

Documentation enhancements

  • Changed pkgdown.yaml to trigger on version tags instead of on every pull request and merge (#16).

creelutils 0.1.1

Released 2026-03-23

  • establish_db_con(): changed default conn_type from "odbc" to "config" to use the faster RPostgres driver by default. ODBC is still available via conn_type = "odbc". Note: password prompt from config will block execution in non-interactive (automated) sessions.

creelutils 0.1.0

Released 2025-05-08

  • Early development version. Putting the package together with basic structure, documentation, GitHub actions, etc.

creelutils initialized on 2025-01-31