GSoC Blog 2
Jun 17 — Jul 5
By June 21
The past few days were used to document the progress so far and the testing of the majority of the spatial datasets encountered so far. Most of these installed perfectly and all of these results were put in a google doc and linked to in the PR so far.
Some time was spent on code cleaning and making sure the scripts run properly.
By June 24
Worked on some documentation for the dashboard and added descriptions on the working of the new functions that were added.
Had discussions with the mentor regarding access to the server on which the dashboard was actually going to be run.
These discussions involved 3 major things:
- Whether we need the ability to create diffs for spatial datasets as they are very large.
- The dropping of the dataset once the actual check was performed. This wasn’t needed according to the mentor.
- The need for having to press the enter key for every spatial key. This wasn’t ideal as the script was meant to be run remotely by a server at regular intervals.
By June 28
By this time I had solved the third problem stated above. This was due to a missing special argument in the args object that was passed onto the postgres engine. This was solved successfully.
Testing of the dataset for diff generation was started and it seemed to be giving satisfactory results nad the diffs created were also satisfactory to denote change in the data.
This also solve issue #1 in the above problems listed.
By July 1
On further dealing with the diff generation testing, it was found that the diffs were generated each time irrespective of the change in the md5 hash of the dataset.
On discussing with the mentor and a fellow GSoC participant, it was found to be an error in the existing code. This was supposedly solved by seeing the schema of the current data_details.json file.
By July 5
I undertook the testing of the script under the server. For that I had to go through a bit of the djang docs and the docs of the django plugin that we use that is ‘django-contrab’, after learning about this I tried it. But it gave an error.
The source was found to be a error in the code added previously.
As of the time of writing, this error is being investigated.
