Robert Sachunsky
Robert Sachunsky
this would allow specifying `FROM ocrd/core-cuda-tf1` for all modules depending on Tensorflow 1 – so this (huge!) Docker layer can be **shared** same could be worked out for TF2 and...
This starts Conda with `deps-conda` as a replacement for Apt with `deps-ubuntu` to install system dependencies. System dependencies should be encapsulated better than via fixed Linux distributions in OCR-D. [Long...
https://github.com/OCR-D/core/blob/742906e330d5ef1139fd18c86b73c154c0a67eae/ocrd/ocrd/cli/workspace.py#L472 This implementation is strange: 1. if it encounters a file ref without any URL FLocat (i.e. if the file was created by OCR-D itself as output of a processor),...
When you `ocrd workspace clone /some/path/to/mets.xml` (**without the indiscriminate download option**) on a workspace which contains local files, the following happens: 1. a mets:file with remote FLocat will still keep...
For processors consuming MODS metadata, it would help (as in: easier and more efficient code) being able to use the Python object model. For example, querying `language` or `script` by...
We have used `mets:note` to record the workflow provenance now, in accordance with [this proposal](https://github.com/OCR-D/spec/issues/108). Alas, currently, we define the namespace prefix `ocrd="https://ocr-d.de"` ad-hoc on each leeve element. This can...
For running jobs, we agreed there is a need for some kind of timeout anyway, universally. It gets enforced by - Processor Server model: the calling Processing Server (currently [with...
In the Processing Server, we currently add jobs to the queue unconditionally, without checking whether any job is already running on the respective workspace: https://github.com/OCR-D/core/blob/b7130307ce68c5f074c1ceea24b66ae9ee9ef289/ocrd_network/ocrd_network/processing_server.py#L281 Obviously, this will create inconsistent...
Currently, our precaution of only downloading to those `resource-locations` which the processor supports resolving, is quite user-unfriendly: ``` ERROR ocrd.cli.resmgr - The selected --location {location} is not in the {this_executable}'s...
IMO we should strive to support much more validation and repair features in ocrd_validators.page_validator – esp. functionality known from [PRImA Converter and Validator](https://primaresearch.org/tools/PAGEConverterValidator) (PCV) and [HTR United VX](https://github.com/HTR-United/HTRVX) (HTRVX). From...