Changelog • ALprekDB

ALprekDB (development version)

Added db_init() and db_close() for creating/opening DuckDB databases with schema versioning and validation.
Added db_write_panel() and db_write_master() to persist processed panel data and linked master datasets.
Added db_write_year() for incremental year-by-year data addition without full reprocessing; validates no duplicate years.
Added db_read_panel() and db_read_master() with full R type reconstruction (factor levels, Date, integer) via internal column type registry.
Added db_list_tables(), db_table_info(), and db_query() for database inspection and arbitrary SQL queries.
DuckDB and DBI are optional dependencies (Suggests); the package works without them.

Added alprek_synthetic_budget(), alprek_synthetic_classroom(), and alprek_synthetic_student() for generating realistic panel data without confidential ADECE files.
All three generators share classroom codes (via seed) so outputs are linkable with linkage_create_master().

Added “Getting Started with ALprekDB” vignette covering installation, codebooks, and a quick-start pipeline.
Added “Building Multi-Year Panel Data” vignette for budget, classroom, and student panel construction workflows.
Added “Cross-Module Linkage and Analysis” vignette demonstrating master dataset creation, transform enrichment, and DuckDB integration.
Added “Package Architecture and Data Dictionary” vignette with S3 class hierarchy, data dictionary, and all 37 validation checks.

Added linkage_classroom_budget() to join classroom and budget panels by school_year + classroom_code, with automatic column deduplication (authoritative source strategy).
Added linkage_student_classroom() to join student and classroom panels.
Added linkage_aggregate_students() to compute classroom-level summaries from student data (38 aggregate variables including demographics, GOLD gains, chronic absence rates, service density, and eDECA gains).
Added linkage_create_master() to produce a two-level master dataset: $classroom_level (~208 columns) and $student_level (~445 columns), with derived per_child_budget and per_seat_budget.
Added linkage_validate() with 8 diagnostic checks (match rate, orphans, key uniqueness, NA introduction, year coverage, region consistency).
Added linkage_summary_stats() for descriptive statistics on linked data.
Added linkage_export_csv(), linkage_export_excel(), linkage_export_rds(), linkage_export_stata(), and linkage_export_parquet().

Added student_transform() with 27 derived analysis variables:
- 12 GOLD gain scores (6 domains x raw + scale)
- 6 K-readiness transition indicators (Emerging -> Accomplished)
- 2 chronic absence variables (flag + percentage, configurable threshold)
- 2 service density variables (n_services, risk_index)
- 5 eDECA pre-post T-score gains
Transform is opt-in with per-category toggles; preserves the input S3 class.
Extended linkage_aggregate_students() with 14 additional classroom-level aggregates derived from transform variables.

Added student_read() with auto format detection (legacy 202-column vs new 270-column formats) and footer removal.
Added student_clean() with 15 cleaning steps: demographics standardization, gross income parsing (dual-format: simple ranges and FPL-based), delivery type normalization, service indicator binary encoding, attendance correction (abs + extreme cap), IEP2 enhanced indicator derivation, and 4 assessment batteries (GOLD 6-domain, PPVT, eDECA pre/post, ASQ).
Added student_validate() with 12 advisory checks.
Added student_bind_years() for multi-year panel construction (no imputation; student attributes change yearly).
Added student_track() for tracking student presence across years.
Added student_summary_stats() for descriptive statistics.
Added student_process(), student_process_years(), and student_config() convenience functions.
Added student_export_csv(), student_export_excel(), student_export_rds(), student_export_stata(), and student_export_parquet().

Added alprek_student_race_mapping() for student race/ethnicity standardization (7 levels from cross-year variants).
Added alprek_student_delivery_mapping() for student delivery type normalization (handles 2024-25 mixed capitalization).

DOB retained by default (include_pii = FALSE removes names, guardian info, and state/student IDs but preserves DOB for age analysis).
eDECA Post-test scores preserved (unlike prior ad-hoc scripts that dropped them; 2024-25 has 95% Post data).
Attendance: negative values corrected with abs(); values >180 days set to NA with warning.

Added classroom_read() with auto format detection (legacy ~100-column vs new ~125-column formats).
Added classroom_clean() with degree classification (8 credential levels), race/ethnicity normalization, experience calculation, and coordinate standardization.
Added classroom_validate() with 10 advisory checks.
Added classroom_bind_years() with forward-fill imputation for geographic coordinates and year_first_funded within site groups; imputation log tracks every change.
Added classroom_process(), classroom_process_years(), and classroom_config() convenience functions.
Added classroom_export_csv(), classroom_export_excel(), classroom_export_rds(), classroom_export_stata(), and classroom_export_parquet().

Added budget_read() with auto format detection (legacy ~176-column vs new ~28-column formats).
Added budget_clean() to transform raw Excel data to long-format intermediate with payroll tax proportional allocation (legacy) and Additional Funds 1 & 2 aggregation.
Added budget_validate() with 7 checks including reconciliation diagnostics ($1.00 tolerance).
Added budget_transform() for wide master with derived share variables.
Added budget_bind_years() for multi-year panel construction.
Added budget_track_classrooms() for tracking classroom presence across years.
Added budget_summary_stats() for descriptive statistics.
Added budget_process(), budget_process_years(), and budget_config() convenience functions.
Added budget_export_csv(), budget_export_parquet(), budget_export_excel(), and budget_export_rds().

Package skeleton with testthat, MIT license, GitHub-ready structure.
Data-driven codebook system with CSV files in inst/extdata/: alprek_category_groups(), alprek_delivery_types(), alprek_county_codes().
Utility functions: parse_classroom_code(), parse_classroom_codes(), alprek_infer_school_year(), alprek_clean_colnames().
Messaging system: msg_info(), msg_success(), msg_warn().