Downcast
--------
This repository contains tools for processing and converting data from
the DWC system into WFDB and other open formats.
Requirements
------------
Python 3.4 or later is required. A Unix-like platform is required -
Debian and CentOS have been tested; Mac OS might work as well. This
package will not work on Windows.
For processing data in BCP format, the ply package is required.
For processing data directly from SQL Server, the pymssql package is
required. (This package is now mostly abandoned and should probably
be replaced with a different backend.)
Quick start
-----------
If you have access to the demo DWC database, download and unpack these
files (about 30 GB uncompressed.) You will then need to create a
"server.conf" file, which should look like this:
[demo]
type = bcp
bcp-path = /home/user/dwc-demo
(where /home/user/dwc-demo is the directory containing "Alert.dat",
"Alert.fmt", etc.) See server.conf.example for other examples.
The demo database spans the time period from 1:00 AM EDT on October
31, 2004, to midnight EST on November 1. To parse and convert a slice
of the data (say, from 10:00 to 10:05 AM), first we initialize an
output directory and set the starting time:
$ ./downcast.py --init --server demo \
--output-dir /home/user/dwc-test-output \
--start "2004-10-31 10:00:00.000 -05:00"
Then run a batch conversion while specifying the end time:
$ ./downcast.py --batch --server demo \
--output-dir /home/user/dwc-test-output \
--end "2004-10-31 10:05:00.000 -05:00"
If we wanted to keep going, we could run the same --batch command
again, increasing the end timestamp each time. We don't need to
specify the starting timestamp for --batch, since the "current"
timestamp is saved automatically.
To "finalize" the output (and forcibly truncate all patient records at
the specified end time), we use the --terminate option. This wouldn't
be done for a real database conversion, but it's useful for a simple
test:
$ ./downcast.py --batch --server demo \
--output-dir /home/user/dwc-test-output \
--end "2004-10-31 10:05:00.000 -05:00" \
--terminate
This should result in a bunch of patient records in WFDB format,
stored in /home/user/dwc-test-output.