1  """Shared cache-path helpers for the coaching pipeline.
 2  
 3  Internal to the pipeline package -- consumed by coach_cv and export_coaching.
 4  """
 5  
 6  __all__ = ["CACHE_DIR", "safe_id"]
 7  
 8  import re
 9  from pathlib import Path
10  
11  CACHE_DIR = Path(__file__).resolve().parents[1] / "cache" / "v1"
12  # Anchored to repo root -- safe regardless of the directory from which the CLI is invoked.
13  # Version prefix so schema changes don't corrupt old cache entries.
14  
15  
16  def safe_id(job_id: str) -> str:
17      """Sanitize job_id for use in filesystem paths.
18  
19      Args:
20          job_id: Raw job posting ID string.
21  
22      Returns:
23          String with non-alphanumeric/underscore/hyphen chars replaced by '_'.
24      """
25      return re.sub(r"[^a-zA-Z0-9_-]", "_", job_id)