Skip to main content

tilearn.joblist APIs

These modules orchestrate ranking and extraction across multiple CSV-backed lists.

tilearn.joblist.plat.List​

Represents one working CSV list.

Constructor​

List(backup_path, file_path, prec)
  • backup_path: directory containing copied CSV files.
  • file_path: original file path (basename is mapped into backup_path).
  • prec: mode flag (1 precedence, 0 non-precedence).

Methods​

  • info(): read typed rows from backing CSV.
  • check(): return mode flag.
  • run(): return rows with appended ranking factor:
    • sum_factor for prec=1
    • factor_data for prec=0

Other plat helpers​

  • count_file(folder_path): count top-level CSV files.
  • ja_all(lists, path): total job count across active list objects.
  • location(lists, type): locate best candidate list index (type="sub") or row index (type="row").

run helpers​

  • set_prec(...): copy selected prefix block for precedence mode.
  • set(...): copy one selected row for non-precedence mode.
  • set_const(...): dispatch copy strategy by mode.
  • file_seek(...): count list sources still containing jobs.
  • optimal_list(lists, path, backup_path): end-to-end sequence builder.

tilearn.optimal_list(lists, path, backup_path)​

This is the main orchestration API exposed at package level.

Workflow​

  1. Backup source CSV files into backup_path.
  2. Repeatedly select the current globally best candidate row/block.
  3. Append selected jobs to output and update working CSV files.
  4. Clear temporary CSV artifacts in backup_path.

Output​

Returns the final ordered list of rows, typically with one appended factor column.

Example​

import tilearn as tl
from tilearn import plat as pl

lists = [
    pl.List("data/backup", "data/list1.csv", prec=0),
    pl.List("data/backup", "data/list2.csv", prec=1),
]

schedule = tl.optimal_list(lists, "data", "data/backup")
On this page