Feat/improvements

[dawidlinek]

@copilot write summary and review

[Copilot]

Pull request overview

This PR refactors the StandardScaler and BaseModel classes to work with NumPy arrays instead of Pandas DataFrames, improving performance and simplifying the API. The changes also standardize the column naming convention for one-hot encoded calendar features.

• Migrated StandardScaler from DataFrame-based to NumPy array-based implementation
• Optimized BaseModel data slicing by pre-grouping data by hour and using boolean masks
• Updated calendar feature column naming from specific prefixes (weekday_, hour_, month_) to a unified "is_" prefix
• Added error handling for concurrent task execution

Reviewed changes

Copilot reviewed 6 out of 7 changed files in this pull request and generated 9 comments.

Show a summary per file

File	Description
epftoolbox2/scalers/standard.py	Refactored to use NumPy arrays instead of DataFrames, added static method for scalable column detection
epftoolbox2/models/base.py	Optimized data slicing with pre-computed hour groups, added error handling for concurrent tasks, converted to NumPy arrays
epftoolbox2/models/ols.py	Set fit_intercept=False for LinearRegression (appropriate for standardized data)
epftoolbox2/data/sources/calendar.py	Changed one-hot encoding prefix from feature-specific (weekday_, hour_, month_) to unified "is_" prefix
tests/scalers/test_standard.py	Updated tests to work with NumPy arrays instead of DataFrames
tests/data/sources/test_calendar.py	Updated column detection logic to match new "is_" prefix naming convention
.gitignore	Broadened patterns from specific files to more general patterns (main[1-9]*.py, *.jsonl)

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.