search

OpenPhilology / Iris

The OCR pipeline to succeed Rigaudon
2 stars 0 forks source link

Fix batch input format to support multiple files and synchronization points #3

Closed mittagessen closed 10 years ago

mittagessen commented 10 years ago

The current iris.batch() function accepts multiple input files but the created input chain structure will not be correct. It should be possible to create jobs containing multiple files using synchronization points.

mittagessen commented 10 years ago

Works kinda now. The current implementation writes a subtask file containing all chains associated with a job to the job directory and assembles result values/current task state from this file.