[HELP]

[martinkilbinger]

Can the exclusive input ID (-e flag) be replaced by the existing NUMBER_LIST?

[cailmdaley]

Short answer: yes, functionally — with a few deltas worth knowing before pulling the trigger.

Where they meet. Both end up in the same place, FileHandler._format_process_list: NUMBER_LIST replaces the scanned number list as the source (file_handler.py:77–83 — inline comma list or a file of one ID per line), while -e filters whatever list is in play down to the one matching ID (file_handler.py:1111–1124). For a single ID the resulting process list is identical, module by module — so a per-job NUMBER_LIST = <id> reproduces -e <id> exactly.

The deltas:

1. Format. -e 277.282 is transformed internally to the numbering string -277-282 (id_to_test = f"-{exclusive.replace('.', '-')}"). NUMBER_LIST takes raw numbering-scheme strings, so the dot→dash transform moves into the calling script.
2. CLI vs config. The canfar chain (curl_canfar_local.sh → init_run_exclusive_canfar.sh → job_sp_canfar.bash) passes one ID per headless job via -e with a shared config. Replacing it means writing the ID into a per-job config — and job_sp_canfar.bash already has exactly that machinery (update_config, currently used for SMP_BATCH_SIZE), so it's one more rewrite line, not new infrastructure. Side benefit: the run copies its config into the output log dir, so the processed ID becomes part of the recorded config — better provenance than a CLI flag that vanishes.
3. Early validation — the one real loss. The input-dir scan runs unconditionally either way (_save_num_patterns is always called; NUMBER_LIST doesn't skip it, so no startup saving). With -e, the ID is intersected with what the scan actually found, and a typo fails immediately with "No input file found matching exclusive ID". NUMBER_LIST is taken on faith — file paths are constructed blindly (_get_file_name) and a wrong ID only surfaces when a module tries to open them. If -e goes, a cheap NUMBER_LIST ⊆ scanned check in _format_process_list would preserve the early failure.
4. Cleanup win. Removing -e deletes plumbing in args.py, run.py, and file_handler.py — plus job_handler.py:78, where exclusive is stored and never used (dead already).

Happy to do this as a follow-up PR (code + the job_sp_canfar.bash/curl-chain migration + the validation check) if you want it — it composes naturally with the #745 cleanup.

— Claude on behalf of Cail.