Why to use WebLicht-Batch

WebLicht is a workflow engine that gives users access to a wide number of natural language processing tools. It is often not advisable to feed such tools with large input data as the time to process the data rises faster than the input size. WebLicht-Batch helps you to split your input into smaller chunks. It then feeds the resulting batch of items to various instances of WebLicht processes (your input is, so to speak, processed in parallel), waits for the results to come back, and then merges the individual results into an accummulated result (as if you had submitted your input to WebLicht at once).

How to use WebLicht-Batch

WebLicht-Batch helps you to send batch jobs to WebLicht. You have two options to upload your data:

• Drag your data (a big text file or a zip file of text files) from your file manager to drop area in the "Upload File" tab. Or you can click onto the drop area to select the file to be uploaded.
• Paste any URL that refers to your data into the input box in the "Submit URL" tab. For example, you can input a shared link from your B2DROP account.

Frequently Asked Questions

• Q: What kind of zip files can I submit to WebLicht-Batch?

  A: Only submit a zip file whose contents is flat (ie, which has no hierarchical directory structure) and homogeneous: all files share the same mediatype (at the moment, only text/plain is accepted) and language. Note that all files will be subjected to the same WebLicht processing chain.

• Q: Why do you accept only files with mediatype "text/plain"?

  A: WebLicht-Batch splits large input files into smaller ones. At the moment, our splitter works on plain text files only. In the future, we might support other formats such as TCF or CONLL.

• Q: Why do you accept only zip files whose content files are all in the same language?

  A: With WebLicht-Batch, you select a single easychain for the entire zip file. A chosen easychain assume a language that hence must be shared by all files of a zip archive.

• Q: What happens to the data that is transfered to WebLicht-Batch?

  A: When you submit a zip file, each file of the zip archive is processed individually. If a given file (being part of the archive) is larger than 100kb, it is split into parts that WebLicht-Batch processes individually. The results of processing the parts are then re-combined into file-level results.

• Q: Can I use WebLicht-Batch to select parts of the zip file for processing?

  A: Not yet.

• Q: What is a user key, why would I need it?

  A: Whenever you enter a new browser session with WebLicht-Batch, a user key is generated for you. All tasks that you initiate with WebLicht-Batch are registered with this user key. When you inspect the task status within the same browser session, see "Inspect Task Status", you do not need the user key. But what if processing your files with WebLicht-Batch takes quite some time, and for some reason, you want to close your browser session? In this case, copy the user key to a safe place; you will need to paste it into a text field (see "Inspect Task Status") to retrieve your tasks from an earlier browser session.

For any questions please contact the WebLicht-Batch team.