refactor: stop mutating inputs in extraction pipeline

[boscorat]

Parent issue

#194 — function purification

Problem

Three functions silently mutate their inputs:

1. get_standard_fields (statement_functions.py:775) — mutates the caller's checks_and_balances DataFrame in-place via hstack(..., in_place=True) (lines 890-891, 900, 902-903, 905). This is the most problematic — the caller's DataFrame is silently modified.

2. extract_fields (statement_functions.py:314) — mutates location.top_left coordinates in-place (line 354-355) and mutates results DataFrame via vstack(in_place=True) (lines 401, 489, 566).

3. get_table_from_region (pdf_functions.py:109) — mutates location.top_left and location.bottom_right in-place (lines 167-168) and mutates vertical_lines list in-place (lines 152-153, 157-159, 175).

Proposal

• get_standard_fields should return a tuple (data, updated_checks_and_balances) instead of mutating in-place.
• extract_fields should work on copies of location and return new coordinates.
• get_table_from_region should accept a copy of vertical_lines and return updated coordinates instead of mutating.

Code relationships

Target functions:

• get_standard_fields() — statement_functions.py:775
• extract_fields() — statement_functions.py:314
• get_table_from_region() — pdf_functions.py:109

Callers (will need updating):

• get_results() — statement_functions.py:688 (calls all three)
• Statement.get_results() — statements.py:631 (calls module-level get_results)

Related (pass mutated objects):

• Location dataclass — data.py (the object being mutated)
• spawn_locations() — statement_functions.py:77 (produces Location objects that get mutated)

Existing tests

None. None of these functions have direct test coverage. Refactoring will not break any existing tests.

Siblings

• refactor: BatchMetadata dataclass to replace 13 duplicated parameters #195 — BatchMetadata dataclass
• refactor: ExtractionContext dataclass for extraction pipeline #196 — ExtractionContext dataclass
• refactor: decompose Statement constructor into pure factory + persistence #197 — Statement constructor decomposition
• refactor: decompose process_pdf_statement into pure extraction + persistence #198 — process_pdf_statement decomposition
• refactor: add optional conn parameter to build_datamart and related functions #200 — DB connection threading
• refactor: move processing out of StatementBatch constructor #201 — StatementBatch constructor
• refactor: remove dead logs and file_path parameters from pipeline functions #202 — Dead parameter cleanup
• refactor: Parquet class accepts pre-loaded data instead of reading from disk #203 — Parquet class purification
• refactor: _handle_parquet_write_error returns error dict instead of mutating #204 — _handle_parquet_write_error purification