Data is published to the HGL user interface via a Python script. This script is run from PCs (or Macs running Parallel+Windows& or VMWare+WIndows7) with ArcGIS Desktop software installed.
The script processes scanned map data layers as follows:
- Accesses GIS raster data layers stored in ArcSDE (which may be on Alcott (old version) or Mercator (if we succeed in database migration)
- Performs a "Projection" to convert the data to "WebMercator" projection in an ESRI GIS format
- Note - the projection to be applied is different for each layer, depending on the projection stored in the old database to start with
- Writes the data to disk on the Pelham web server
- Creates a 'map service' (Runs an ESRI "Analysis" function, which creates a "Service Definition File")
- Publishes the Service Definition File to the ArcGIS Server running on Pelham
- and generates a map tile cache to speed data display.
The only way to tell if the script really worked is to bring up the newly created layer in ArcGIS Desktop software and look at the map display.
The map data cache, the map service and the map tile cache are all controlled by ArcGIS Server. For raster data (scanned maps), copying the data means that the data set will be stored on the filesystem mounted on Pelham going forward, and not in ArcSDE on Mercator. For vector data sets, layers are copied from ArcSDE 9.3.1 on Alcott into ArcSDE 10.4.1 on Mercator. Then, a separate script creates a map template and then another script publishes the data as a map document.
Vector data sets will still reside in ArcSDE. They cannot be stored in the DRS. The Python scripts to handle vector data have not been prototyped yet but templates exist. We are looking into modifying these templated scripts into the HGL environment. Publishing vector data will not be prototyped until it is determined what approach to take with the ArcSDE migration.
A prototype script for publishing scanned maps stored in ArcSDE was obtained from ESRI and altered for our needs. This script needs enhancements, some cleaning up and error checking added to it. It also needs a second component to support publishing vector map documents once the ArcSDE migration strategy is formalized.
|Meet with Dave and Bonnie Burns to discuss what the script needs to do||Tim, Dave, Bonnie||12/6/16|
|Reviewing code provided by Bonnie, and decide on what needs to be done||Tim||Jan|
|Improve its structure and add error handling and logging||Tim||Jan|
|Test the publish script and benchmark to predict throughput||Tim, Bonnie, Dave||Feb|
|Run the publish script in prod on all 14,000 layers||Tim||Feb|