Update paths used in pipeline

[sgreenbury]

Closes #53.

The following changes are included in this PR:

• Revise the filepath structure for pipeline outputs, include @property methods within the config providing the paths and refactor scripts to use these
• Add a Population class for reading pipeline outputs
• Fallback download option when pyrosm fails (e.g. for West Yorkshire)
• Configurable time tolerance
• Rewrite of the run pipeline script with Python
• Multiprocessing for script 3.3

---

Remaining tasks:

• Add config to override paths
• Update path for logs
• Run with MSOA with travel_times = false
• Merge 56 write output to matsim format #57
• Check determinism from a given seed. Update: fixed input data produces deterministic outputs for a given seed but note that the osmox output can change regularly and this can lead to different downstream AcBM outputs.
• Run pipeline for:
  • Leeds (5000 households)
  • Leeds (all households)
  • Greater London (5000 households)
  • Greater London (all households): currently script 3.2.2_assign_primary_zone_work.py is not progressing, consider alternative parameters for the optimization updating optimization for only total deviation and adding multiprocessing for 3.3 enables pipeline to completely run.

[sgreenbury]

Fallback for when pyrosm fails

[Hussein-Mahfouz]

This looks great to me! Happy to merge it

Only one minor comment on logging: It seems some src code is using the old logging setup. Would this conflict with your new imlementation? e.g.

acbm/src/acbm/assigning/feasible_zones_primary.py

Line 21 in b3d5581

logger = logging.getLogger("assigning_primary_feasible")

[sgreenbury]

Thanks @Hussein-Mahfouz - merging now with comments addressed or opened as new issues.