This page allows a user to submit a PubMed document file in NXML format or a plain text file to the Reach information extraction system for processing. The text or PubMed file is uploaded from the user's local computer. Reach processing results are returned directly to the browser, or downloaded to the user's computer, in one of several JSON or CSV formats, selected at the time the input file is uploaded.

To upload a text or NXML file for Reach processing, select the file from your local filesystem by clicking the Choose File button.

Note: For PubMed document files, only valid NXML files with extensions .nxml or .xml are permitted.

Next, select the desired output format for the results:

• fries, for FRIES Consortium JSON (the default),
• indexcard, for MITRE IndexCard JSON,
• serial-json, for a JSON serialization of the entire Reach internal document

  The serial-json format can return VERY large result files (10s of megabytes). Use this format with caution.

To initiate processing, click the Upload File button.

After a minute or two of processing, results will be returned to the browser and either displayed or downloaded to a file, depending on the selection of the Download Output checkbox and the browser content settings.